Comprehensive Chiroptical Spectroscopy, Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules (Volume 2)

Citation preview

COMPREHENSIVE CHIROPTICAL SPECTROSCOPY Volume 2

COMPREHENSIVE CHIROPTICAL SPECTROSCOPY Volume 2

Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules

Edited by Nina Berova Prasad L. Polavarapu Koji Nakanishi Robert W. Woody

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright © 2012 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750–8400, fax (978) 750–4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748–6011, fax (201) 748–6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762–2974, outside the United States at (317) 572–3993 or fax (317) 572–4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Advances in chiroptical methods/edited by Nina Berova . . . [et al.]. p. cm. Includes index. ISBN 978-0-470-64135-4 (hardback : set)—ISBN 978-1-118-01293-2 (v. 1)—ISBN 978-1-118-01292-5 (v. 2) 1. Chirality. 2. Spectrum analysis. 3. Circular dichroism. I. Berova, Nina. QP517.C57A384 2012 541.7–dc23 2011021418 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

IN MEMORY OF CARLO ROSINI (1948–2010)

Carlo Rosini obtained his degree in Chemistry (1973) at the University of Pisa, where he completed his thesis on the stereochemistry of Ni(II) complexes. He entered the Italian CNR by joining the group of Professor Piero Salvadori and the research on determination of absolute configuration by Circular Dichroism. Later on, Carlo Rosini spent two years (1977–1979) at the King’s College in London, under the supervision of Professor Stephen F. Mason. During this period he studied polarized-light-based spectroscopy and its application to structural determinations. He was appointed as associate professor (1992) at the University of Pisa and then as a full professor (1997) at the University of Basilicata, Potenza. The field of chirality was fundamental to the scientific activity of Carlo Rosini. His broad scientific interests included many aspects of organic stereochemistry, like asymmetric organic synthesis, chiral discrimination mechanisms, chiral stationary phases for enantioselective chromatography, and structural characterization of organic molecules by Circular Dichroism. The last research projects of Carlo Rosini were oriented toward chemical/computational approaches for the determination of absolute configuration by linking experimental and theoretical studies. We miss his enthusiasm and his charisma, but we will remember his life and his contributions to the science and the chemical community. Carlo Rosini was one of the first scientists who accepted to contribute a chapter to this volume. Although his premature and tragic death prevented his submission, his spirit never died and is now, not only in the chapter contributed by his co-workers and former students, but also in the minds of all of us who had the privilege to know him and collaborate with him.

CONTENTS

PREFACE CONTRIBUTORS

PART I A HISTORICAL OVERVIEW

1

THE FIRST DECADES AFTER THE DISCOVERY OF CD AND ORD BY AIME´ COTTON IN 1895 Peter Laur

PART II ORGANIC STEREOCHEMISTRY

2

SOME INHERENTLY CHIRAL CHROMOPHORES—EMPIRICAL RULES AND QUANTUM CHEMICAL CALCULATIONS

xi xiii

1 3

37 39

Marcin Kwit, Pawel Skowronek, Jacek Gawronski, Jadwiga Frelek, Magdalena Woznica, and Aleksandra Butkiewicz

3

ELECTRONIC CD OF BENZENE AND OTHER AROMATIC CHROMOPHORES FOR DETERMINATION OF ABSOLUTE CONFIGURATION

73

Tibor Kurt´an, S´andor Antus, and Gennaro Pescitelli

4

ELECTRONIC CD EXCITON CHIRALITY METHOD: PRINCIPLES AND APPLICATIONS

115

Nobuyuki Harada, Koji Nakanishi, and Nina Berova

5

CD SPECTRA OF CHIRAL EXTENDED π -ELECTRON COMPOUNDS: THEORETICAL DETERMINATION OF THE ABSOLUTE STEREOCHEMISTRY AND EXPERIMENTAL VERIFICATION

167

Nobuyuki Harada and Shunsuke Kuwahara

vii

viii

CONTENTS

6

7 8

ASSIGNMENT OF THE ABSOLUTE CONFIGURATIONS OF NATURAL PRODUCTS BY MEANS OF SOLID-STATE ELECTRONIC CIRCULAR DICHROISM AND QUANTUM MECHANICAL CALCULATIONS Gennaro Pescitelli, Tibor Kurt´an, and Karsten Krohn DYNAMIC STEREOCHEMISTRY AND CHIROPTICAL SPECTROSCOPY OF METALLO-ORGANIC COMPOUNDS James W. Canary and Zhaohua Dai CIRCULAR DICHROISM OF DYNAMIC SYSTEMS: SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

217

251

289

Angela Mammana, Gregory T. Carroll, and Ben L. Feringa

9 10

11

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS Cheng Yang and Yoshihisa Inoue

317

THE ONLINE STEREOCHEMICAL ANALYSIS OF CHIRAL COMPOUNDS BY HPLC-ECD COUPLING IN COMBINATION WITH QUANTUM-CHEMICAL CALCULATIONS Gerhard Bringmann, Daniel G¨otz, and Torsten Bruhn

355

DETERMINATION OF THE STRUCTURES OF CHIRAL NATURAL PRODUCTS USING VIBRATIONAL CIRCULAR DICHROISM

387

Prasad L. Polavarapu

12

DETERMINATION OF MOLECULAR ABSOLUTE CONFIGURATION: GUIDELINES FOR SELECTING A SUITABLE CHIROPTICAL APPROACH

421

Stefano Superchi, Carlo Rosini, Giuseppe Mazzeo, and Egidio Giorgio

PART III INORGANIC STEREOCHEMISTRY

13

APPLICATIONS OF ELECTRONIC CIRCULAR DICHROISM TO INORGANIC STEREOCHEMISTRY

449 451

Sumio Kaizaki

PART IV BIOMOLECULES

14

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS Robert W. Woody

473 475

ix

CONTENTS

15

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES Claudio Toniolo, Fernando Formaggio, and Robert W. Woody

499

16

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

545

Claudio Toniolo and Fernando Formaggio

17

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS Jaroslav Kypr, Iva Kejnovsk´a, Kl´ara Bedn´arˇ ov´a, and Michaela Vorl´ıcˇ kov´a

18

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

575

587

Roberto Corradini, Tullia Tedeschi, Stefano Sforza, and Rosangela Marchelli

19 20 21 22

23 24

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS Donald M. Gray

615

DRUG AND NATURAL PRODUCT BINDING TO NUCLEIC ACIDS ANALYZED BY ELECTRONIC CIRCULAR DICHROISM George A. Ellestad

635

PROBING HSA AND AGP DRUG-BINDING SITES BY ELECTRONIC CIRCULAR DICHROISM Mikl´os Simonyi

665

CONFORMATIONAL STUDIES OF BIOPOLYMERS, PEPTIDES, PROTEINS, AND NUCLEIC ACIDS. A ROLE FOR VIBRATIONAL CIRCULAR DICHROISM Timothy A. Keiderling and Ahmed Lakhani

707

STRUCTURE AND BEHAVIOR OF BIOMOLECULES FROM RAMAN OPTICAL ACTIVITY Laurence D. Barron and Lutz Hecht

759

OPTICAL ROTATION, ELECTRONIC CIRCULAR DICHROISM, AND VIBRATIONAL CIRCULAR DICHROISM OF CARBOHYDRATES AND GLYCOCONJUGATES

795

Tohru Taniguchi and Kenji Monde

25

ELECTRONIC CIRCULAR DICHROISM IN DRUG DISCOVERY Carlo Bertucci and Marco Pistolozzi

INDEX

819

843

PREFACE

Chirality is a phenomenon that is manifested throughout the natural world, ranging from fundamental particles through the realm of molecules and biological organisms to spiral galaxies. Thus, chirality is of interest to physicists, chemists, biologists, and astronomers. Chiroptical spectroscopy utilizes the differential response of chiral objects to circularly polarized electromagnetic radiation. Applications of chiroptical spectroscopy are widespread in chemistry, biochemistry, biology, and physics. It is indispensable for stereochemical elucidation of organic and inorganic molecules. Nearly all biomolecules and natural products are chiral, as are the majority of drugs. This has led to crucial applications of chiroptical spectroscopy ranging from the study of protein folding to characterization of small molecules, pharmaceuticals, and nucleic acids. The first chiroptical phenomenon to be observed was optical rotation (OR) and its wavelength dependence, namely, optical rotatory dispersion (ORD), in the early nineteenth century. Circular dichroism associated with electronic transitions (ECD), currently the most widely used chiroptical method, was discovered in the mid-nineteenth century, and its relationship to ORD and absorption was elucidated at the end of the nineteenth century. Circularly polarized luminescence (CPL) from chiral crystals was observed in the 1940s. The introduction of commercial instrumentation for measuring ORD in the 1950s and ECD in the 1960s led to a rapid expansion of applications of these forms of chiroptical spectroscopy to various branches of science, and especially to organic and inorganic chemistry and to biochemistry. Until the 1970s, chiroptical spectroscopy was confined to the study of electronic transitions, but vibrational transitions became accessible with the development of vibrational circular dichroism (VCD) and Raman optical activity (ROA). Other major extensions of chiroptical spectroscopy include differential ionization of chiral molecules by circularly polarized light (photoelectron CD), measurement of optical activity in the X-ray region, magnetochiral dichroism, and nonlinear forms of chiroptical spectroscopy. The theory of chiroptical spectroscopy also goes back many years, but has recently made spectacular advances. Classical theories of optical activity were formulated in the early twentieth century, and the quantum mechanical theory of optical rotation was described in 1929. Approximate formulations of the quantum mechanical models were developed in the 1930s and more extensively with the growth of experimental ORD and ECD studies, starting in the late 1950s. The quantum mechanical methods for calculations of chiroptical spectroscopic properties reached a mature stage in the 1980s and 1990s. Ab initio calculations of VCD, ECD, ORD, and ROA have proven highly successful and are now widely used for small and medium-sized molecules. Many books have been published on ORD, ECD, and VCD/ROA. The present two volumes are the first comprehensive treatise covering the whole field of chiroptical spectroscopy. Volume 1 covers the instrumentation, methodologies, and theoretical xi

xii

P R E FA C E

simulations for different chiroptical spectroscopic methods. In addition to an extensive treatment of ECD, VCD, and ROA, this volume includes chapters on ORD, CPL, photoelectron CD, X-ray-detected CD, magnetochiral dichroism, and nonlinear chiroptical spectroscopy. Chapters on the related techniques of linear dichroism, chiroptical imaging of crystals and electro-optic absorption, which sometimes supplement chiroptical interpretations, are also included. The coverage of theoretical methods is also extensive, including simulation of ECD, ORD, VCD, and ROA spectra of molecules ranging from small molecules to macromolecules. Volume 2 describes applications of ECD, VCD, and ROA in the stereochemical analysis of organic and inorganic compounds and to biomolecules such as natural products, proteins, and nucleic acids. The roles of chiroptical methods in the study of drug mechanisms and drug discovery are described. Thus, this work is unique in presenting an extensive coverage of the instrumentation and techniques of chiroptical spectroscopy, theoretical methods and simulation of chiroptical spectra, and applications of chiroptical spectroscopy in inorganic and organic chemistry, biochemistry, and drug discovery. In each of these areas, leading experts have provided the background needed for beginners, such as undergraduates and graduate students, and a state-of-the-art treatment for active researchers in academia and industry. We are grateful to the contributors to these two volumes who kindly accepted our invitations to contribute and who have met the challenges of presenting accessible, up-to-date treatments of their assigned topics in a timely fashion. Nina Berova Prasad L. Polavarapu Koji Nakanishi Robert W. Woody

CONTRIBUTORS

S´andor Antus, University of Debrecen, Research Group for Carbohydrates of the Hungarian Academy of Sciences, Debrecen, Hungary Laurence D. Barron, Department of Chemistry, University of Glasgow, Glasgow, United Kingdom Kl´ara Bedn´arˇ ov´a, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Nina Berova, Department Chemistry, Columbia University, New York, New York, USA Carlo Bertucci, Department of Pharmaceutical Sciences, University of Bologna, Bologna, Italy Gerhard Bringmann, Institute of Organic Chemistry, University of W¨urzburg, W¨urzburg, Germany Torsten Bruhn, Institute of Organic Chemistry, University of W¨urzburg, W¨urzburg, Germany Aleksandra Butkiewicz, Polish Academy of Sciences, Institute of Organic Chemistry Warsaw, Poland James W. Canary, Department of Chemistry, New York University, New York, New York, USA Gregory T. Carroll, Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, USA Roberto Corradini, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Zhaohua Dai, Department of Chemistry and Physical Sciences, Pace University, New York, New York, USA George A. Ellestad, Department of Chemistry, Columbia University, New York, New York, USA Ben L. Feringa, Stratingh Institute for Chemistry, University of Groningen, Groningen, The Netherlands Fernando Formaggio, Department of Chemistry, University of Padova, Padova, Italy Jadwiga Frelek, Polish Academy of Sciences, Institute of Organic Chemistry, Warsaw, Poland Jacek Gawronski, Department of Chemistry, A. Mickiewicz University, Poznan, Poland Egidio Giorgio, Department of Chemistry, University of Basilicata, Potenza, Italy xiii

xiv

CONTRIBUTORS

Daniel G¨otz, Institute of Organic Chemistry, University of W¨urzburg, W¨urzburg, Germany Donald M. Gray, Department of Molecular and Cell Biology, The University of Texas at Dallas, Richardson, Texas, USA Nobuyuki Harada, Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai, Japan Lutz Hecht, Department of Chemistry, University of Glasgow, Glasgow, United Kingdom Yoshihisa Inoue, Department of Applied Chemistry, Osaka University, Suita, Japan Sumio Kaizaki, Department of Chemistry, Graduate School of Science, Osaka University, Osaka, Japan Timothy A. Keiderling, Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois, USA Iva Kejnovsk´a, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Karsten Krohn, Department of Chemistry, University of Paderborn, Paderborn, Germany Tibor Kurt´an, Department of Organic Chemistry, University of Debrecen, Debrecen, Hungary Shunsuke Kuwahara, Department of Chemistry, Toho University, Funabashi, Japan Marcin Kwit, Department of Chemistry, A. Mickiewicz University, Poznan, Poland Jaroslav Kypr, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Ahmed Lakhani, Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois, USA Peter Laur, Institute of Inorganic Chemistry, RWTH Aachen University, Aachen, Germany Angela Mammana, Department of Chemistry, University of Dayton, Dayton, Ohio, USA Rosangela Marchelli, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Giuseppe Mazzeo, Department of Chemistry, University of Basilicata, Potenza, Italy Kenji Monde, Faculty of Advanced Life Science, Frontier Research Center for Postgenome Science and Technology, Hokkaido University, Sapporo, Japan Koji Nakanishi, Department of Chemistry, Columbia University, New York, New York, USA Gennaro Pescitelli, Department of Chemistry and Industrial Chemistry, University of Pisa, Pisa, Italy Marco Pistolozzi, Department of Pharmaceutical Sciences, University of Bologna, Bologna, Italy

CONTRIBUTORS

Prasad L. Polavarapu, Department of Chemistry, Vanderbilt University, Nashville, Tennessee, USA Carlo Rosini, (deceased) Department of Chemistry, University of Basilicata, Potenza, Italy Stefano Sforza, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Mikl´os Simonyi, Chemical Research Center, Department of Molecular Pharmacology, Hungarian Academy of Sciences, Budapest, Hungary Pawel Skowronek, Department of Chemistry, A. Mickiewicz University, Poznan, Poland Stefano Superchi, Department of Chemistry, University of Basilicata, Potenza, Italy Tohru Taniguchi, Faculty of Advanced Life Science, Frontier Research Center for Postgenome Science and Technology, Hokkaido University, Sapporo, Japan Tullia Tedeschi, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Claudio Toniolo, Department of Chemistry, University of Padova, Padova, Italy Michaela Vorl´ıcˇ kov´a, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Robert W. Woody, Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, Colorado, USA Magdalena Woznica, Polish Academy of Sciences, Institute of Organic Chemistry, Warsaw, Poland Cheng Yang, Department of Applied Chemistry, Osaka University, Suita, Japan

xv

a

c

o

Projection along b-axis

Figure 5.38. Absolute stereostructure of the C60 fullerene cis-3 bisadduct (R,R,f,s A)-[CD(+)280]-32 (top) and projection along b-axis (bottom). (Redrawn from reference 54, with permission.)

(a)

(b)

(c)

Figure 8.1. Dynamic chirality at the molecular and supramolecular level detected by CD spectroscopy. (a) A chiral molecule can direct achiral molecules to self-assemble into chiral supramolecular structures. (b) A chiral molecular switch or motor undergoes conformational changes that include inversion of molecular helicity. (c) Chiral molecules can self-assemble into chiral supramolecular structures, the chirality of which is determined by the enantiomer in excess.

S

Rotor Axle Stator O

Legs

O O

O

n

n S

S

Au Surface

2 CD (mdeg)

hν

0 –2

hν

200

240

280

320

λ (nm)

Figure 8.6. Assembly of thiol-terminated light-driven rotary molecular motors on a semitransparent gold film provides a monolayer of chiroptical material that can be analyzed with CD spectroscopy. The CD signals invert between positive and negative bands, corresponding to changes in the helicity of the molecules comprising the monolayer upon the application of photons and thermal energy. The initial spectrum (solid black) inverts (dotted black) after irradiation with UV light (λmax = 365 nm) at room temperature. After heating the surface (70◦ C, 2 h) the spectrum inverts again to restore the original (solid gray). A second dosage of photons inverts the signal (dotted gray). Heating brings the rotors back to the original orientation relative to the substrate [30].

UV O NH N2N

S

S

O

DET-4o

O NH

Vis

HN NH2

H2N

+

S

S

O HN NH2

DET-4c

DET-4o

–

+ –

dG

dC

Figure 8.18. A photoswitchable chiroptical DNA complex. At the top is shown the photoequilibrium between the open (DET-4o) and closed (DET- 4c) forms of a dithienylethene molecular switch that contains pendant ammonium groups to confer water solubility and allow the switches to bind electrostatically to the polyanionic backbone of DNA when the amine is protonated. Photochemical ring closure was accomplished through the use of a UV lamp (λmax = 365 nm) using a 340-nm cutoff filter. Photochemical ring opening was performed with visible light using ® a 520-nm cutoff filter. Molecular models (created using Hyperchem ) show that the distance between the terminal ammonium functionalities closely resembles the distance between the anionic phosphate groups of a guanosine (G)–cytosine (c) base pair [58].

Interface

Interface

Compression

M-Chiral

Compression

Achiral

P-Chiral

Figure 8.24. A supramolecular chiroptical switch composed of achiral amphiphiles. Space-filling structures of achiral amphiphile (TARC18), which forms a Langmuir–Schaefer film at the air–water interface, and chiral supramolecular structures formed upon interface compression. (Reprinted by permission of John Wiley & Sons, Inc. [66].)

OR

complexation N

I

N

N

N

N

N

N

N

NOH

N OH

N mutarotation of glucose

N

N

42a: R = (C2H4O)8CH3 42b: R = n-C4H9

complexation

N

N

n

42

N

N

N

N

N OH

N N N

OH

left-handed helical complex

right-handed helical complex

Figure 9.24. Chiral self-aggregation of achiral polymer induced by a saccharide. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

3 MeOH / water 5:1 6:1 7:1 CD/(mdeg)

CD/(mdeg)

2

1

0

–1 300

8

8:1 9:1 10:1 310

320

330

340

350

360

6 β-glucose +7 +2.4 mdeg 4 (337 nm) 2

time 0h 1h 3h 8h 15 h 24, 48 h 15, 24, 48 h 8h

0

5h 3h α-glucose –2 –3 +2.4 mdeg 1h 0h (337 nm) –4 300 310 320 330 340 350 360

λ (nm)

λ (nm)

(a)

(b)

Figure 9.25. (a) Induced CD spectra of a mixture of 42a (1mM in monomer unit) and D-glucose (0.3M) in 5:1–10:1 MeOH/H2 O at 25◦ C. (b) Time-dependent CD spectra of a mixture of 42a (1mM in monomer unit) and α- or β-D-glucose (0.3M) in 5:1 MeOH/H2 O at 25◦ C. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

(a)

(b)

Figure 12.6. Exciton chirality defined by the allowed B naphthalene and S = O transitions in 1

(a) E-conformer of (S)-8 and (b) Z-conformer of (S)-9.

0.05 0

Δε

−0.05 Nd

−0.15

ν −0.25 14.00

(103 cm−1) 16.00

18.00

20.00

0.3 MI

Δε

0 −0.3 −0.6 −0.9 20.00

Ln Ho -SAPR-8-C4(llll)-M[Ln(+)-(hfbc)4] with an encapsulated alkali metal ion

ν (103 cm−1) 21.00

22.00

23.00

24.00

25.00

0.04

Δε

0 −0.04 −0.08 −0.12

ν (103 cm−1)

Er

−0.16 12.00 14.00 16.00 18.00 20.00 22.00 24.00

Figure 13.6. CD spectra in the hypersensitive 4f –4f transitions of Cs[Ln((+)-hfbc)4] in CHCl3 (left) and the proposed structure in solution (right).

200

Δε

100 0 −100 −200 250

300 λ (nm)

350

Figure 13.9. Exciton CD spectra of M[La((+)-hfbc)4 ] in CHCl3 . M: Cs (red), Rb (green), K (blue), Na (black).

Internuclear (net effect) Intranuclear (Ln end) Intranuclear (Cr end)

Δε (M−1cm−1)

50

CrIII

LnIII 0

N

N

−50

N

N

N

N

L2 275

300

375 325 350 Wavelength/nm

400

425

O N

Figure 13.10. Right: Structure of the ligand L2(below) and -[LnIII CrIII (L2)3 ]6+ (above). Left: Schematic vertical lines summering the dominant coupling effects in the CD spectra of -LnIII CrIII (L2)3 ]6+ . The black line corresponds to the CD spectrum of -[GdCr(III)(L2)3 ]6+ in CH3 CN.

10 8 6 4 Ellipticity (mdeg)

2 0 –2 –4 B-DNA hZαADAR1 yabZαE3L IsZαE3L orfZαE3L spZαE3L vZαE3L

–6 –8 –10 –12 –14 –16 230

240

250

260

270

280

290

300

310

320

2800

3200

3600

Wavelength (nm) (a) 6

Ellipticity (mdeg) at 255 nm

4 2 0 hZαADAR1 yabZαE3L IsZαE3L orfZαE3L spZαE3L vZαE3L

–2 –4 –6 –8 –10 0

400

800

1200

1600

2000

2400

Time (sec) (b)

Figure 19.1. (a) CD spectra of poly[d(G–C)] in the B form and in the presence of Zα domains from human ADAR1 editing enzyme (hZα), yaba-like disease virus (yabZα), lumpy skin disease virus (lsZα), orf virus (orfZα), swinepox virus (spZα), and vaccinia virus (vZα), listed in decreasing order of their abilities to convert B-DNA to Z-DNA. CD spectra of the added proteins contributed the negative signals at wavelengths shorter than about 250 nm. (b) Kinetics of the B to Z conversion in the presence of the same domains. Proteins were added to poly[d(G–C)] at a protein to base-pair ratio of 0.4 (with the final protein concentration being 90 μM), in a buffer of 10 mM HEPES, pH 7.4, 10 mM NaCl, and 0.1 mM EDTA, except for yabZα where the buffer included 100 mM NaCl. Spectra were taken at 25◦ C using a 2-mm-pathlength cell. CD values are in mdeg ellipticity. (Reproduced from Quyen et al. [6] by permission of Oxford University Press, copyright 2007.)

θ (mdeg)

5

0

–5

Sp1ZF6 (ER)4 + [2GC (10)] Sp1ZF6 (KE)4 + [2GC (10)] Sp1ZF6 (G4S)4 + [2GC (10)] Free [2GC (10)]

–10 200

220

240

260

280

300

320

Wavelength (nm)

Figure 19.8. CD spectra of a DNA containing two GC-box sequences separated by a 10-bp spacer, 2GC(10), complexed with each of three peptides containing six zinc fingers but with different linkers between zinc fingers 3 and 4: Sp1ZF6(ER)4 with linker (Glu–Ala–Ala–Ala–Arg)4 , Sp1ZF6(KE)4 with linker (Lys–Ala–Ala–Glu–Ala)4 , and Sp1ZF6(G4 S)4 with linker (Gly–Gly–Gly–Gly–Ser)4 . Spectra were taken at 20◦ C using a 1-mm-pathlength cell. Samples contained 4.5 μM peptide–DNA complex in 10 mM Tris–HCl (pH 8.0), 50 mM NaCl, 0.005% Nonidet P-40, and 1 mM dithiothreitol. CD values are mdeg ellipticity. (Reprinted with permission from Yan et al. [42], ©2005, American Chemical Society)

B-DNA

A-DNA

Z-DNA

Binding modes minor groove binding

major groove minor groove

major groove

intercalation

minor groove

Figure 20.1. Representation of the three principal secondary structures of DNA. The right¨ handed A and B form are obtained from standard parameters within the Schrodinger–Maestro graphical interface. The thinner and more elongated Z form is obtained from X-ray parameters of a hexamer as imported from the protein data bank (PDB). In this representation, three units of the hexamer are stacked in order to display the overall left-handed zig-zag helicity. The structure on the far right depicts drug–DNA double-helix interactions with the drug colored black: minor groove binding (top) and intercalation between base pairs (bottom).

Figure 21.1. Binding sites are indicated by specific ligands in white, warfarin (Site I, right-hand side) and diazepam (Site II, left-hand side). (Reprinted with permission from reference 3, copyright 1996, Elsevier.)

Cleft Thyroxine 5 2°: lodipamide

IIIB FA 5 Thyroxine 2,3 2° : Oxyphenbutazone 2° : Propofol

IIIA: Drug Site 2 FA 3, 4 Thyroxine 4 Diflunisal Diazeapam Halothane Ibuprofen Indoxyl sulphate Propofol 2° : CMPF

IB FA 1 Hemin 2° : Azapropazone 2° : Indomethacin 2° : TIB

FA 2 IIA: Drug Site 1 FA 7 Thyroxine 1 Azapropazone CMPF DIS Indomethacin Iodipamide Oxyphenbutazone IIA-IIB Phenylbutazone FA 6 2° : Diflunisal TIB 2° : Halothane Warfarin 2° : Ibuprofen 2° Indoxyl sulphate 3° Diflunisal

Figure 21.2. Ligand-binding capacity of HSA defined by crystallographic studies. (Reprinted with permission from reference 9, copyright 2005, Elsevier.)

Figure 21.16. Mutual positions of quercetin (molecular modeling 48) and warfarin (X ray [8]) in the cavity of Site I subdomain IIA. (Reprinted with permission from reference 48, copyright 2003, Elsevier.)

IB

IA

IIIB

site II IIA site I

IIIA

IIB

Figure 21.21. X-ray crystallographic structure of HSA [8] with curcumin molecules localized by docking. Subdomains are indicated. (Reprinted with permission from reference 55, copyright 2003, Elsevier.)

1 5

7 2 4 3 6 C16:0

Figure 21.37. Structure of HSA complexed with seven palmitic acid molecules. (Reprinted with permission from reference 105, copyright 2000, Elsevier.)

(a)

(b)

Figure 21.38. Conformational changes in warfarin binding (Site I) as a result of fatty acid binding. (a) Helices h2 and h3 are shown by light shades for defatted and by dark shades for myristate bound HSA. (Reprinted with permission from reference 8, copyright 2001, American Society for Biochemistry and Molecular Biology.) (b) The volume of Site I in defatted HSA is depicted by a light brown semitransparent surface that becomes expanded upon myristate binding (blue semitransparent surface), the red arrows point to structural changes associated with fatty acid binding. (Reprinted with permission from reference 9, copyright 2005, Elsevier.)

Figure 21.40. Two crocetin molecules fitted to FA3 and FA4 sites; the negative exciton dictates the horizontal crocetin molecule to be behind the slanting one. (Reprinted with permission from reference 108, copyright 2001, Elsevier.)

(a)

(b)

Figure 21.41. Electrostatic potentials of genetic variants. (a) lysophospholipid ligand binding at the surface of variant F1-S. (b) variant a. (Reprinted with permission from reference 124, copyright 2006, ACS.)

Figure 21.55. Preferred conformer of diazepam docked into the crystal structure of AGP F1. Hydrogen bonds of the carbonyl oxygen to Glu64 and Gln66, as well as contacts of the ring nitrogens with Arg90 and Tyr127, are indicated. (Reprinted with permission from reference 125, copyright 2008, Elsevier.)

HR1-25

40

Millidegrees

20 10

Millidegrees

30

HR2-25

10

0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

0

0 0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

–10

–10 –20

–20

–30 200

210

220 230 240 Wavelength (nm)

250

200

260

210

220

230

240

1:1 Mixture 10

0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

20 5

0 0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

–10 –20 –30 –40 220

230

240

250

260

Millidegrees

Millidegrees

10

210

260

Difference Spectra

30

200

250

Wavelength (nm)

0

–5

–10 200

210

Wavelength (nm)

220 230 240 Wavelength (nm)

250

260

Figure 25.9. CD spectra of HR1–C25, HR2–C25, and their 1:1 mixture: [peptide] 50 μM, PBS, pH7.4, TFE as the co-solvent (0%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%). Difference CD spectra as a function of TFE concentration (0%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%) is also depicted. The difference CD spectra are calculated subtracting the two individual peptide spectra from those of the mixture. (Reproduced with permission from reference 69.)

PART I A HISTORICAL OVERVIEW

1 THE FIRST DECADES AFTER THE DISCOVERY OF CD AND ORD BY AIME´ COTTON IN 1895 Peter Laur

1.1. SCOPE: SUBJECTS AND TIME FRAME TO BE REVIEWED The story of the Cotton effect begins with its discovery in 1895. Although the news was hailed by leading physicists and chemists, studies to extend, exploit, and apply Cotton’s findings developed at a slower pace than one might have anticipated. One of the reasons for this delay was simply the necessity of the researchers to construct their own optical apparatus. Gradual technical improvements eventually allowed one, in the 1920s, to take chiroptical measurements in the ultraviolet as well as the visible, thus making accessible in principle a great many Cotton effects in colorless (mostly organic) compounds. Despite the paramount importance of such developments, neither the technical details nor the physics involved will be discussed in the following. Rather, a chemist’s view will prevail, paying attention chiefly to experimental results and the application of chiroptics to chemical problems. Since much of the work during the first 20 or so years after Cotton’s discovery was done by physicists and physicochemists, it is not surprising that many investigations were interconnected with or even motivated by the concomitant progress of the theory of optical activity. But also the discussion of this part of (theoretical) physics will be curtailed in the following. The exclusion in this chapter appears justified, because various comprehensive reviews are readily available, as they are for the field of optical instrumentation. By about 1935, Cotton effect measurements were possible with most organic and inorganic chromophores. It is rather surprising that not much use was made of the chiroptical techniques, especially by organic chemists. On the other hand, physical chemists had demonstrated the feasibility of Cotton effect studies in various classes of chemical compounds, but seemed satisfied with this result. Likewise, the advancement of optical Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

3

4

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

instruments for chiroptical measurements slowed down. All this led to a certain climax of chiroptical studies in the early 1930s, to be followed by a near standstill. It is not unreasonable to symbolically connect this phenomenon with the death in 1936 of T. M. Lowry, one of the most active scholars in the field. Arguably, the death of T. M. Lowry ended the first, pioneering period of chiroptical studies. The present chapter will concentrate on reviewing these first “historical” decades. Some work on the experimental study of the Cotton effect continued after 1936 until World War II on a minor scale, on, for example, organic compounds (S. Mitchell) or platinum complexes (I. Lifschitz). But at exactly the same time, new developments took place in the theory of optical activity and its application to chemical problems: Werner Kuhn’s calculation of the absolute configuration of lactic acid in 1935 rang in a new era. The waning interest of the experimentalists contrasts with the increased activity of theoretical chemists like J. G. Kirkwood, E. U. Condon, H. Eyring, or W. Kauzmann, who in the late 1930s advanced different models of optical activity. Still, chemistry had to wait for the period of 1950–1960 for a revitalization of chiroptics. Some reasons for the animation are: (1) the development of X-ray scattering methods for the determination of the absolute configuration, thus anchoring the stereochemistry unambiguously, following J. M. Bijvoet’s seminal publication of 1951; (2) the advent of new, commercially available measuring devices of ORD and CD; and (3) growing interest in natural products chemistry and, generally, optically active systems. But to discuss these topics would need another chapter.

1.2. EARLY CHIROPTICAL STUDIES The discovery of optical activity is credited to the two distinguished French mathematicians, physicists, astronomers, and geodesists (and more) Dominique-Franc¸ois Jean Arago (1786–1853, of Catalan origin) and Jean-Baptiste Biot (1774–1862) [1]. Arago and Biot had been closely associated at least since 1806 in the pursuit of other scientific subjects, and they sometimes published together. Both investigated the optical activity of quartz, and apparently they also shared their equipment to some extent. If, on the one hand, Arago was the first to go into print, Biot, on the other hand, soon became more active in this field and extended the studies. He undoubtedly observed optical activity for the first time in organic compounds such as natural oils and terpenes, or solutions of camphor [2] and cane sugar [3]. Biot continued his research on optical activity throughout his life, later concentrating particularly on tartaric acid. He noticed the wavelength dependence of the optical rotation even at the very beginning of his studies, albeit in a rather qualitative way. Whereas eventually the rotatory dispersion of quartz could be elucidated satisfactorily (which led to Biot’s law, stating that the rotation is inversely proportional to the square of the wavelength), similar solution studies were seriously impeded by experimental deficiencies, particularly the lack of suitable monochromatic light sources. Genuine chiroptical studies were, therefore, rather infrequent until the end of the nineteenth century. One of the most important papers here is a report by the Norwegian physicist Adam Arndtsen, who discussed his studies of aqueous solutions of (+)-tartaric acid [4]. Using sunlight, he was able to visually determine the angle of rotation at some of the principal Fraunhofer lines, that is, C (656), D (589), E (527), b (517), F (486), and e (438 nm). He could confirm and extend Biot’s earlier finding that the rotation exhibits a maximum in the spectral region studied, with its wavelength shifting from the

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

blue to the red on increasing concentration. This unexpected and intriguing result led the Swiss chemist Hans Landolt (an important pioneer of the investigation and application of optical activity, as well as one of the “fathers” of Physical Chemistry) in 1877 to introduce the expression “anomale Rotationsdispersion” (anomalous rotation dispersion) [5], which since has become established for the description of such rotatory dispersion curves that run through a maximum or minimum, or show a reversal of sign. It had thus become apparent that spectropolarimetry promised to develop into an interesting field in the future. In his last and comprehensive paper on optical activity, Biot [6] suggested, therefore (translation from the French by the present author): I should like to draw the attention of experimentalists to a class of phenomena which, hitherto, has been little studied but which, nevertheless, for both theoretical and practical purposes, ranks in importance with that of the optical rotatory power itself of which it is a constituent element. I refer to the specific mode of dispersion that each optically active substance or compound imparts to plane polarized light of different wavelengths [literally: refrangibility].

Despite this exhortation, reports on rotatory dispersion remained scarce until the end of the century. This is also evident from the very first book on optical activity, where all data known at that time are summarized, which was published in Germany in 1879 by Landolt [7]. Here, he also describes in detail the optical equipment used by himself and his predecessors. Therefore, it is not necessary to dwell at this point on the measuring devices and optical methods. Although most of the rotations listed (many of which had been determined or redetermined by Landolt himself) refer to the sodium D line only, his book also has short sections on normal and anomalous rotatory dispersion. It is important to realize that so far all reported optically active liquids or solutions were based on organic compounds without absorption bands in the visible. In fact, Landolt emphasized that there is not a single inorganic substance known which shows optical activity in solution (or in the gas phase), from which he tentatively—but incorrectly—concluded that optical activity might be restricted to carbon compounds, except for the solid phase. Surprisingly, he gave no reference to any optically active transition metal complex, although at least Fehling’s solution (a mixture of several Cu(II) tartrate complexes) had been around since 1848 [8]. One might speculate whether such coordination compounds (of a still unknown nature) were ignored as a result of theoretical considerations. It should also be mentioned that measurements in general were limited to practically colorless samples and to merely certain frequencies of the visual solar spectrum. The only other reasonably monochromatic light sources available were based on lithium, sodium, and thallium salts heated in a Bunsen burner (invented in 1866), giving access to the wavelengths 671 nm (Li), 589 nm (Na), and 535 nm (Tl), respectively. It is worthwhile to briefly turn to the “anomalous” refractive dispersion using unpolarized light—that is, the characteristic sigmoidal variation of the index of refraction in the absorption region, running through a maximum and minimum, instead of steadily increasing as the wavelength decreases, as in normal dispersion. This behavior had been discovered in iodine vapor in 1862 by the French physicist F.-P. Leroux [9], and around 1870 it attracted the attention of several investigators, who published independently on anomalous dispersion in the visible, using solutions of organic dyes like fuchsine [10]. There was some dispute as to priority among the Danish physicist Christian Christiansen, the Swiss chemist and physicist Jacques-Louis Soret, and the German physicist August Kundt. While it is clear that Christiansen was the first to publish, the most extensive studies were carried out by Kundt. The relevance of these findings to the present subject

5

6

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

lies in the fact that here was proven the possibility of successfully studying the index of refraction even near absorption bands in the visible. Consequently, a similar anomalous dispersion could be expected to exist for the optical rotation, keeping in mind the relation between the velocity of light, the index of refraction, and the optical rotation. This anomalous dispersion feature of the rotation should have been accessible by existing techniques, if only suitable colored optically active samples had been available. It took more than two decades, however, before this problem was addressed.

1.3. THE DISCOVERY OF THE COTTON EFFECT In 1895, two short papers (“notes”) appeared in the fortnightly journal of the French Academy of Sciences, entitled “Unequal absorption of right and left circularly polarized light by certain optically active substances” [11] and “Anomalous rotatory dispersion of absorbing substances” [12]. The author was the 26-year-old physicist Aim´e Auguste ´ Cotton (Bourg-en-Bresse 1869–S`evres 1951), a PhD student at the prestigious Ecole Normale Sup´erieure in Paris. The first of these papers describes and names the property of “dichro¨ısme circulaire” (what we now call “CD”) associated with an absorption band of an optically active compound in solution, and the second one introduces the corresponding effect in the dispersion mode (now called “ORD”). The full paper of 85 pages, also incorporating studies on magnetic optical activity, was published in 1896 under the heading “Investigations of the absorption and the dispersion of light by optically active media” [13]. It summarizes A. Cotton’s Th`ese de Doctorat, which he prepared from ´ November 1893 to July 1896 at the Physics Laboratory of the Ecole Normale with Professors Marcel Brillouin and Jules Violle as advisors. Based on his important discoveries, Cotton was accorded the degree of Docteur e` s Sciences in 1896. In his thesis, Cotton for the first time reports data of (a) optical rotations close to both sides of an absorption band in the visible, using solutions of Cu(II) and Cr(III) coordination compounds with tartrate or malate ligands, and (b) the associated circular dichroism. It is quite obvious that Cotton was successful to a large degree owing to both the quality of his optical components and the skillful and precise construction of the measuring devices, especially for the determination of very small values of the ellipticity, but also to his power of observation, and—last but not least—to a fortunate choice of optically active samples. In this chapter, however, his technical equipment and the underlying physical principles shall not be discussed in detail, because Cotton himself gives a full description in his major paper, and there are also comprehensive reviews elsewhere as, for example, in the books by Mitchell and Lowry (see below). While Cotton’s expression “dispersion rotatoire anomale” (anomalous rotatory dispersion) is self-explanatory, a comment concerning his novel term “dichro¨ısme circulaire” (circular dichroism) may be appropriate. Cotton did not always measure directly or indirectly the difference in absorption of left- and right-circularly polarized light by his sample [i.e., (εL –εR )], but rather the ellipticity of the emerging elliptically polarized light. In this case, his measuring device included, apart from a Nicol prism to provide plane-polarized light, a “double circular polarizer” consisting of two quarter-waveplates placed side by side in the plane-polarized light beam in such a way that their principal axes are at 90◦ to one another and at 45◦ to the plane of the incident light. This arrangement allowed the observation of left- and right-circularly polarized light beams next to each other. On the introduction of a sample showing circular dichroism, the beams would be differently absorbed, which could be detected visually or photometrically. The field of vision was thus divided into two halves by these λ/4 plates. When

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

he used white instead of monochromatic light for the examination of his optically active sample solutions, these two halves showed different colors. This reminded him of the dichroism observed with certain doubly refracting crystals, where the ordinary and the extraordinary ray are absorbed unequally, as found by Biot in tourmaline and later by the Austrian mineralogist Haidinger in many other cases [14]. Cotton therefore chose the term “circular dichroism.” In fact, Haidinger had already discovered this phenomenon in amethyst quartz in 1847 [15]. Nowadays, the expression “circular dichroism” probably just awakens vague memories of the original visual observations; this context has been largely forgotten nowadays, with the advent of automated electronic spectropolarimeters. Actually, Cotton himself had already performed some photometric measurements, but had found them inferior to his visual results.

1.4. THE FIRST CD AND ORD CURVES Cotton’s measurements obviously were not only restricted to the visible, but also quite limited as to the wavelengths available. Even under favorable conditions, at most eight spectral lines were at his disposal, namely, 657 (red, near C), 589 (yellow, sodium D line), 581 (orange, near D), 562 (greenish yellow), 522 (green, between E and b), 475 (blue, near F), 459 (blue-violet), and 437 nm (violet, near e) [the letters C, D, E, b, F, and e refer to the Fraunhofer lines so designated]. A comparison with Arndtsen’s paper of 1858 [4] shows that hardly any improvement of the spectral availability had taken place until the end of the nineteenth century. However, on the positive side it can be seen that these lines are spread rather evenly across the whole visual region. Nevertheless, the generation of continuous absorption and rotation curves, as often published, on the basis of observations at some of these individual wavelengths, leaves much to the whim of the draftsman, especially concerning the position and magnitude of any maxima and minima. Such “data” should not be overinterpreted. This situation would prevail in the decades to come. It is not unexpected that, at the onset of his investigations, Cotton chose Fehling’s solution (“liqueur de Fehling”) for his studies. It is, after all, in the direct line of Biot’s research to look at derivatives of active tartaric acid. Secondly, the only area of importance where the application of optical activity had become established was saccharimetry; and thirdly, Fehling’s solution was a proven and powerful reagent in carbohydrate chemistry [16]. It seems that Cotton systematically progressed from the complex and notoriously unstable Fehling’s solution to simpler alkali copper(II) tartrates, the preparation of which he describes in detail. By the way, it is amusing to note that in one case he reports the precipitation of copper tartrate from a copper sulfate solution by adding the aqueous solution of a crystal of Seignette salt (potassium sodium tartrate); this crystal had been prepared by Pasteur himself. Unfortunately, these copper complexes proved to be very unstable; they changed or simply decomposed with time or at elevated temperature and were also light-sensitive. Furthermore, the chemical composition of these aqueous solutions was unknown (and, to some extent, still is), and attempts at isolating any well-defined compound failed. A solution of crystalline copper malate, perhaps more stable, did not show any observable circular dichroism. Despite these drawbacks, Cotton did obtain many ORD and some CD data, but obviously the reproducibility of the experiments remains questionable, and the curves shown in print [13] should be interpreted with caution.

7

8

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The most convincing chiroptical effects, however, were observed with aqueous solutions of potassium chromium(III) tartrate, prepared in situ. They shall be discussed here in some detail. Figure 1.1 is Cotton’s Figure 18 on page 408 of reference 13, and it shows a complete “Cotton effect” in the ORD and the CD near 570 nm. Because of its seminal importance, this figure has been later reprinted by others a number of times. The actually measured data are given as follows: 657 nm, rotation ρ + 1◦ 26 , ellipse [sic] φ + 32 ; similarly: 589, +2◦ 30 , (−3◦ 40 ); 581, +1◦ 46 , −4◦ 54 ; 562, −1◦ 21 , −4◦ 16 ; 522, −2◦ 50 , −1◦ 25 ; and 475, 1◦ 52 [no sign given in the paper; from the curve it is evident that ρ must be negative], +28 . Data were thus collected at six wavelengths only, because the onset of a second strong absorption band made observations at shorter wavelengths impossible. The parentheses around the ellipticity value at the sodium D line are Cotton’s and indicate that this number results from photometric measurements. Despite its beautiful appearance, there are unfortunately some flaws in this figure and the data as printed. A comparison of the figure with the data listed above makes evident two discrepancies at 562 nm: In the figure, the angle φ is given as −4◦ 46 (not −4◦ 16 ), and the corresponding angle ρ is given as −0◦ 21 (not −1◦ 21 ). On reexamination, the true values were verified to be φ − 4◦ 46 and ρ − 1◦ 21 . The figure should be redrawn, therefore, using this value of ρ. Such a modification would necessarily modify the shape of the ORD curve, while not basically changing it. Cotton gives these corrections in

+3°

+2°

ρ) ons (

Rotati

+1°

0

Elli s pse

–1°

(ϕ)

–2°

–3°

–4°

–5°

Figure 1.1. CD and ORD of potassium chromium(III) 657

589 581562

522

475 459 437

tartrate (solvent H2 O). (From A. Cotton, Ann. Chim. Physique 1896, [7] 8, 347; Figure 18, p. 408.)

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

a letter to Professor Ladislas Natanson in Cracow, Poland, quoted on pages 33/34 of reference 17. He explains the first error as a misprint, and he states that the second one is a mistake by the “dessinateur” (draftsman). However, the really important disagreement between the ORD and CD, as we can see immediately from the curves—with hindsight—lies in the incompatibility of their signs. If we accept the rotation values as correct, as seems reasonable, the sign of the CD is in error. And so it is! Cotton himself redressed this flaw two years later [18] in a paper, the first sentence of which runs as follows (translated from the French): ‘It is easy to be mistaken as to the sense of a circular vibration.’ Admitting his mistake in the assignment of the direction of the rays circularly polarized by a Fresnel rhomb, he imputed it to his misinterpretation of some of Billet’s tenets in the latter’s “Trait´e ” [19]. Apparently, Billet had used the expression “principal section” of a mica crystal in an unorthodox way and had also treated this crystal as positive, contrary to the common practice. Therefore, all of Cotton’s CD curves, and the sign of all ellipticities published before 1898, ought to be inverted. But not everyone read or responded to this correction; others did so, but without indicating it. The confusion that might have been generated was fortunately curtailed by the fact that very few scientists, apart from Frenchmen, studied the circular dichroism in the following decades. But as late as in 1923, (Ms.) N. Wedeneewa in Moscow (for example) still used the earlier “wrong” sign of the CD, when she reported the ORD and CD of camphor quinone [20]. Similarly, T. M. Lowry just reprinted Cotton’s Figure 18 in his famous classic of 1935 [21] without any comment, whereas S. Mitchell in his treatise on the Cotton effect [22] of 1933 simply shows an inverted CD curve in ostensibly the same figure (see Figure 1.2), also without any further comment. Another point of criticism could be raised because of the all-too-vague identity of the samples investigated. Although Cotton carefully describes the preparation of his samples, as mentioned earlier, their inherent instability cannot preclude changes with time, perhaps also as the result of shifting equilibria between the several complexes present. Indeed, small changes even in the synthesis of the tartrate complexes can lead to the total inversion of the anomalous rotatory dispersion, as has been observed by Wedeneewa [20]. All this calls for caution with respect to the early ORD and CD publications. However, concerning the key compound discussed at length, potassium chromium(III) tartrate, all doubts were finally set to rest by W. Kuhn [23], who much later very carefully repeated Cotton’s work and found it fully correct (Figure 1.3).

1.5. THE REACTION OF THE LEARNED WORLD TO COTTON’S DISCOVERIES Cotton’s papers raised the immediate attention of Wilhelm Ostwald (Nobel Prize 1909), who, one year after the publication of the original notes in the Comptes Rendues [11, 12], wrote two abstracts thereof himself for his journal Zeitschrift f¨ur Physikalische Chemie [24]. This was followed by his six-page review of Cotton’s full paper [13] in the same year [25], with several CD and ORD curves reprinted, including Cotton’s original Figure 18, discussed above. It should be pointed out that the lack of correspondence of the sign of the ORD and the CD could not have been noticed by Ostwald at that time, since the necessary theoretical background had not yet been provided. With these reviews, Ostwald acquainted the chemical world with Cotton’s results, and his name carried much weight. It is certainly unusual that preliminary notes by a foreign physics student and extracts of his

9

10

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

5°

4

3

2

1

0

–1

–2

–3 6600

6200

5800

5400

5000

Wave-length Å.U. Rotations Ellipticities

4600

Figure 1.2. CD and ORD of potassium chromium(III) tartrate (solvent H2 O). (From S. Mitchell, The Cotton Effect, Bell, London, 1933; Figure 12, p. 23; reproduced with permission.)

thesis should induce an already famous physical chemist to such a presentation. Incidentally, already the “sponsoring” of Cotton’s notes by the renowned physicist Gabriel Lippmann from Luxembourg (Nobel Prize 1908)—such notes had to be presented by an academician—attests to the importance attributed to them. One might well say that chiroptics had a splendid start. The speed with which the news was reported and hailed is altogether breathtaking. For example, the physical chemist Landolt referred to Cotton’s studies already in the second edition of his book, published in 1898 [7]. Mention should also be made of the German physicist Paul Drude, who included a treatment of Cotton’s “(anomalous) rotary dispersion” in his famous Lehrbuch der Optik of 1900 [26]. So, by the beginning of the twentieth century, the international world of physics and physical chemistry was well aware of Cotton’s results. It took only a few additional years before a thorough theoretical treatment was provided by L. Natanson, Professor of Theoretical Physics at the Jagiellonian University Krak´ow (Cracow, Poland). The title of his important paper, “On the elliptic polarization of light transmitted through an absorbing naturally-active medium” [27], with a supplementary note [17], needs no further comment. Here, Natanson treated the interdependence of absorption, optical rotation, and circular dichroism. Probably in order to spread his results further, also an amalgamated and shortened French translation of both papers by the Count of Ballehache was published very shortly thereafter [28]. The relations presented here between the sign of the rotation and the circular dichroism have become known as the “R`egle de Natanson” or “Natanson’s Rule” [29]. This finally allowed the

11

xz–xT

log x

[A]–10–3

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

2,75 7,5 3,5

2,50 5,0 3,0

Z

3

1

2,25 2,5 2,5 + 0 2,0

2,00 −

2

1,75 –2,5 1,5 1 1,50 –5,0 1,0 g –0,10

1,25 –7,5 0,5 4

–0,05

g 1,00 –10,0 0

–0,05 –12,5 –0,5

7000 Å

6000

5000

4000

3000

λ

Figure 1.3. UV, CD, and ORD of potassium chromium(III) tartrate (solvent H2 O). (From W. Kuhn, ¨ A. Szabo, Z. Phys. Chem. 1931, B15, 59; Figure 1, p. 62; Oldenbourg Wissenschaftsverlag Munchen, reproduced with permission.)

prediction of the sign of the circular dichroism associated with a specific absorption band, based on just the anomalous rotation curve, which should not be too difficult to obtain. Natanson’s papers included the following sentences on the first page: “Effects of this kind have been observed and investigated by Monsieur A. Cotton” [27] and “Des ph´enom`enes de ce genre ont e´ t´e observ´e et analys´es par M. Cotton” [28]. Here we find the seed that has developed into the important technical terms “Cotton’s Phenomenon” and “Cotton Effect,” which have been used ever since, with the first one preferred in the early decades of the twentieth century. At this point it may be timely to more formally give a definition of the Cotton effect as we understand it today. It may be interesting to compare the definition given in 1933 by Stotherd Mitchell on page 24 of his book on the Cotton effect (incidentally the first monograph of this kind) [22] with the definition by Werner Kuhn from 1960 [30]. Mitchell wrote: “A maximum ellipticity and zero rotation are found in this region [of the absorption band]. The rotation reaches a maximum value on one side of the band and

12

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

a minimum on the other. This variation of rotation and ellipticity in the neighbourhood of an absorption band has been called the Cotton effect.” [Mitchell’s italics] Kuhn stated: “[Cotton] found that optical rotatory power as a function of the wavelength often shows, in the region where the substances show ordinary absorption, a characteristic anomaly which is associated with a circular dichroism in the absorption region and which after the name of its discoverer is called a Cotton effect.” [Kuhn’s italics] It is satisfactory that both definitions, published some 30 years apart, fully agree with one another; furthermore, we still can subscribe to both of them, even 50 years later. Many similar definitions can be found over the last 80 years, all of them stressing the point that the ensemble of rotatory dispersion and circular dichroism in the absorption region collectively constitute the Cotton effect. Nevertheless, quite commonly the term Cotton effect has loosely been used to characterize merely the “anomalous” rotation features, since in the decades following Cotton’s discoveries the available data were mostly limited to the optical rotation. In fact, in many cases it has been considered sufficient to have reached the first maximum of the rotatory dispersion curve, still outside the absorption band, to apply the term Cotton effect. In recent decades, when ORD effectively disappeared in favor of CD, the term usually means the CD curve only.

1.6. MORE TARTRATES: THE PHYSICIST’S PLAYGROUND Cotton’s discovery of circular dichroism raised so much interest in Brace’s Physics Laboratory at the University of Nebraska that it was decided to construct an improved and more sensitive apparatus for measuring both elliptical polarization and rotation, in order to repeat and extend the French findings. The American physicist DeWitt Bristol Brace was himself active in the field of optical activity and had in 1904 described an elliptical polarizer and compensator that was incorporated not only in the optical system used in Nebraska, but also later in Europe. Brace died in 1905 and had, therefore, no part in the further development. The first results on, for example, complex chromium, copper, cobalt, and nickel tartrates and copper malate were presented by M. F. McDowell in 1905 [31]. The ellipticity had been measured in “all parts of the spectrum,” which means at some 10 different wavelengths of the visual solar spectrum. Unfortunately, the calculation of the ellipticity was found to be incorrect, and some compounds were irreproducible. This was carefully rectified at the same laboratory in 1912 by L. B. Olmstead, who studied tartrates, malates, and lactates of chromium, copper, cobalt, and manganese [32]. Also here, the so-called “monochromatic” light, with a spectral band width of perhaps 20 nm, was obtained from sunlight. Although the optical part of the investigation seems to be impeccable (except that Cotton’s first—incorrect—sign protocol of the circular dichroism was still used), the identity of the compounds studied is uncertain. Olmstead himself points out: “No chemical analyses of the compounds were made; the names assigned being merely for convenience, and not indicating that the chemical formulæ are known.” [Olmstead’s italics]. He observed that Cotton’s results for potassium chromium tartrate could be repeated quantitatively when the sample was prepared from potassium dichromate and potassium tartrate, but an oppositely signed Cotton effect developed when the potassium dichromate was replaced by chromium acetate. Undoubtedly, the samples consisted of a mixture of complexes, as was also indicated by color changes of the solutions, depending on variations of the concentration and with time. As a result, even these carefully collected data are of a qualitative nature only.

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

The same qualifications pertain to a large number of papers of the early twentieth century on the rotatory dispersion of transition metal complexes with optically active ligands. In addition, the rotatory dispersion data were often collected at only four or five rather ill-defined spectral bands in the visible. Examples are found in the report by H. Grossmann and A. Loeb on copper tartrate and malate coordination compounds [33], as well as in the paper by H. Volk on copper, nickel, and cobalt complexes with lactate ligands [34]. Also these investigations were initiated, by the way, in order to verify and extend Cotton’s findings. Cotton himself did not continue his work on the optically active tartrates, but motivated his student Georges Bruhat to address the problem again [35]. Bruhat tried to synthesize and isolate individual, well-defined compounds, but succeeded with most tartrates and malates in part only, because of the easy decomposition of the respective solutions. He did isolate and investigate uranyl tartrate that seemed to be stable and showed a Cotton effect near 500 nm. His optical equipment limited the quality of his measurements rather severely, however. After a disruption of the research by the First World War, he resumed his studies again in 1919 with a much more advanced apparatus. This allowed him to reduce the spectral band width from 30 nm to 10 nm, which was essential to avoid “flattened-out” dispersion curves. In this way, he obtained splendid CD and ORD data for uranyl tartrate and ammoniacal cobalt tartrate [36], for example. But regrettably, not even the high quality of the physical data allows any better analysis of the compounds responsible. The interest in this topic was not yet put to rest in Cotton’s laboratory. In the earlier work, the copper complexes had been particularly unsatisfactory. Therefore, the study of alkaline copper tartrate solutions was taken up again in order to enhance the quality of the samples [37]. Somewhat later, complex chromium [38] and cobalt tartrates [39] were reinvestigated. Good CD data could be collected, but the chemical identity of the species in solution remained uncertain, despite Mathieu’s extensive experiments. Such tartrate studies were not wholly limited to Paris. Also W. Pfleiderer in Basel, Switzerland, had returned to measuring the optical rotation of aqueous alkaline solutions of copper tartrate, and he found his data to qualitatively agree with Cotton’s of 1895 [40]. The chiroptical instability that Nina Wedeneewa in Moscow, Russia, had encountered with alkaline chromium tartrates in the absorption region has been mentioned already [20]. Last but not least, attention is drawn to W. Kuhn’s reevaluation of the same problem, as outlined earlier [23]. The overview presented here is not exhaustive. Because of the similarity of the problems, the preceding discussion pertains also to, for example, optically active lactates, malates, and “sucrates” of transition metals. The respective chiroptical results are not basically different from those with tartrate ligands. It remains to report that even many years later the chemical identity of these complicated coordination compounds has not been fully understood, with several questions still unsettled even today [41]. While some of the variability observed is certainly caused by the gradual replacement of coordinated water by the organic ligands, condensation processes leading to multinuclear species also seem to be involved. It is intriguing that the chiroptical properties of tartrate complexes dominate the study of circular dichroism for three decades. In fact, during these years very few CD measurements have been carried out outside of this area (see later). One might speculate whether this conservatism would perhaps result from the fact that practically all researchers were physicists, who might have had limited awareness of colored optically

13

14

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

active compounds in other fields of chemistry. After all, with, for example, organic xanthates, such compounds did exist, but their investigation was largely confined to only one chemical research group in Russia, as will be seen. The question might well be asked, What motivated these tenacious tartrate investigations? It is obvious that genuine chemical topics were not addressed, such as the stereochemical correlation and the application of optical activity to the study of reaction mechanisms, or as a tool in the elucidation of the chemical constitution. This contrasts with the aims in connection with the Werner complexes, to be discussed in the following section of this report. But as to tartrates, the investigators were primarily interested in the circular dichroism in its own right. They tried to effect improvements in the optical instrumentation in order to enhance the sensitivity and precision of their measuring devices. For testing the various theories of optical activity and to compare calculated and experimental data, the latter should be measured with a maximum of accuracy and reliability. It might well have been felt that the continuation with “well-known” samples like the complex tartrates would be advantageous, with a host of data already existing for comparison.

1.7. WERNER COMPLEXES: INORGANIC CHEMISTS LEARN TO MAKE USE OF THE COTTON EFFECT According to common practice, the tartrate systems discussed in the previous section can be considered to belong to the realm of inorganic chemistry. But they were chosen for chiroptical research without paying much attention to their chemical nature. Chemists have performed hardly any systematic studies of these compounds and have instead tended to neglect them. The situation was quite different with regard to the chemically and structurally welldefined octahedral transition metal complexes, following the Alsatian Alfred Werner’s (1866–1919) introduction in 1893 of his geometric model for centers with the coordination number six [42]. At that time, Werner worked in the laboratory of his doctoral advisor, Professor Arthur Rudolf Hantzsch, at the University of Zurich, Switzerland. He had obtained his doctoral degree only in 1890, but was quickly promoted to a chair of chemistry at this University in 1895. Although his revolutionary concept eventually secured him the Nobel Prize in 1913, it met much resistance among his chemical colleagues. The opposition diminished, however, after he had achieved the resolution of some of his complexes into enantiomers [43], since the occurrence of optical activity was hard to reconcile with other than the octahedral geometry. The optically active compounds, mostly Co(III) complexes, were of greatly varying chemical and optical stability, often racemizing at room temperature within a few hours. It was found that chelating ligands like oxalate ions (O, O  -donor ligands) or 1,2-diamines (N , N  -donor ligands) led to increased stability. In the beginning, the optical rotation at only one wavelength was considered sufficient to characterize a specific complex. But it was soon realized that the stereochemical correlation should not be based on such an individual value, since it varied too much in magnitude and even in sign from one complex to the next, notwithstanding a close chemical relationship. Furthermore, as the enantiomeric purity of the compounds was often uncertain, a particular, selected rotation value could be quite misleading. On the other hand, these complexes were well-suited to measurements of the Cotton effect due to their color, which often brought them within the range of visual observation. The sign of this Cotton effect, associated with electron transitions at the coordinating metal

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

center, was found to be a reliable, characteristic feature. Consequently, measuring the rotatory dispersion including, hopefully, a more or less complete Cotton effect became a common goal. The first Werner complex for which both ORD and CD data through an absorption band were obtained, namely, potassium (trisoxalato)iridate(III) dihydrate, K3 [Ir(C2 O4 )3 ]·2 H2 O, was resolved by M. Del´epine [44], and its chiroptical properties were determined by G. Bruhat [45]. A Cotton effect near 450 nm was found. This is, incidentally, one of the rare instances where the circular dichroism of a Werner complex has been determined in the early part of the twentieth century. In practically all other cases, the term Cotton effect just refers to the rotatory dispersion near or at an absorption band in the visible. Werner himself had already reported similar dissymmetric, optically active complexes with bidentate ligands like [Co(en)3 ]3+ or [Cr(ox)3 ]3 – (en = 1,2-diaminoethane; ox = oxalate ion C2 O4 2− ), but without any further spectropolarimetric data [46]. It is amusing to see that in some other cases the Cotton effect—or, rather, the anomalous rotatory dispersion—rests on measurements at only three different wavelengths as for [Rh(ox)3 ]3− , for example [47]. Werner’s concept achieved its final breakthrough when he published the resolution of the “completely inorganic” complex [Co{Co(NH3 )4 (OH)2 }3 ]Br6 , an octahedral Co(III) complex with bidentate O, O  ligands. This complex, without any carbon atom, is sterically related to the simpler [Co(en)3 ]3+ system. It showed a Cotton effect near 600 nm [48]. This finding finally put to rest the long-lived but obsolete theory that the presence of carbon atoms was essential for the unfolding of optical activity. Werner showed no particular interest in the Cotton effect in its own right, however, but rather made use of it for the stereochemical correlation and the characterization of his compounds. A further example which may be mentioned is [Co(NO2 )2 (en)(pn)], with pn = 1,2-diaminopropane, with either rac-pn or l -pn, that had Cotton effects in the 530- to 540-nm range [49]. In general, Werner’s interest in spectropolarimetry remained limited, and usually he left further chiroptical studies to others. Even when he reported Cotton effects, it is not always clear how, where, and by whom the data were obtained. Meanwhile, a new “center of gravity” for the examination of Werner complexes was developing in Groningen in the Netherlands. Here, the stereochemist Franciscus Mauritius Jaeger (1877–1945), Professor of Inorganic and Physical Chemistry at the Rijksuniversiteit Groningen (RUG) from 1908 to 1945, had by 1915 embarked on a program of the comprehensive investigation of these systems [50]. A great many new, and some already known, Werner complexes were synthesized and studied. Jaeger’s interests lay largely in their crystallographic description, but he also included optical activity in his research program. Most of the spectropolarimetric data generated “plain” ORD curves only, because the anticipated Cotton effect was often beyond the wavelength limit of the optical devices or inaccessible because of too strong an absorption. In Jaeger’s extensive paper of 1919 [51], many such plain curves are reported, but only two cases of a bona fide Cotton effect, namely, in K3 [Cr(ox)3 ] (Cotton effect near 565 nm) and K3 [Co(ox)3 ]·H2 O (near 620 nm). The chromium compound had already been synthesized and resolved by Werner in 1912, but had not been investigated by spectropolarimetry, whereas the (trisoxalato)cobaltate(III) was new. It was very advantageous that Jaeger could induce Israel Lifschitz (1888–1953), Private Docent at the University of Zurich, to join his laboratory in 1921. Lifschitz, whose special area of research had been the absorption spectroscopy and photochemistry of organic compounds, now became Private Docent of Electrochemistry and Photochemistry

15

16

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

and tenured staff member at RUG. He turned to the spectroscopic study of transition metal complexes, with particular attention to spectropolarimetry. Largely owing to his efforts, the laboratory developed into a center of chiroptical spectroscopy, of international repute. Lifschitz had taken his doctorate in 1911 with A. Hantzsch in Germany (Hantzsch had changed from Zurich to the University of Leipzig) and had moved to Switzerland in 1914. It is interesting to note that he there worked close to Werner, who also had been a student of Hantzsch. Perhaps the shift of Lifschitz from organic photochemistry to coordination stereochemistry has thereby been influenced, but presently nothing is known of a personal interaction with Werner. It is difficult to assay in detail Lifschitz’s contribution to the Groningen laboratory, except for his own publications, since Jaeger as laboratory head regularly included the research results of his local colleagues in his own publications, without giving any individual credit. This procedure was rather common in those days. Only sometimes is Lifschitz mentioned in a vague way as a “collaborateur” (coworker). In 1923 Lifschitz started a series of papers called “Investigations of Rotatory Dispersion” [translated from the original German]. In the first paper, he presented and discussed ORD data of complexes of Cr(III), Co(III), Ni(II), and UO2 2− ions with optically active camphor derivatives, including nitrocamphor. These compounds exhibited Cotton effects in the visible. The paper is also noteworthy, because here the technical term “Cotton effect” was introduced into the chemical literature; the earlier term had been “Cotton phenomenon” [52]. In the second paper of this series, Lifschitz reported the ORD Cotton effects of Co and Cu complexes with amino acid ligands (alanine, asparagine), and also of the complex [Cu-(l -pn)2 ]SO4 (effect at 510 nm) [53]. Slightly later, Jaeger extended the chiroptical studies to cobalt complexes with 1,2-diamino ligands, and he reported the Cotton effects in [Co(rac-trans-1,2-diaminocyclopentane)3 ]Cl3 · 4H2 O at 470 nm and in [Co(rac-trans-1,2-diaminocyclopentane)(en)2 ]Br3 · 2H2 O at 500 nm [54]. A few years later, some of these Werner complexes were reinvestigated by Werner Kuhn, making use of advanced instrumentation. Measurements had now become possible down to 280 nm. With potassium (trisoxalato)cobaltate(III), for example, the Cotton effect at ∼600 nm was measured both in rotatory dispersion and in circular dichroism [55]. But now the aim had shifted from using the spectropolarimetric data for chemical and stereochemical correlation, as had been the purpose in Zurich and Groningen, to probing the stereochemistry in depth, with the elucidation of the absolute configuration in mind, and to testing new theoretical models of the optical activity. But it would take a few additional decades before eventually another experimental reinvestigation of the circular dichroism of some of these complexes, in connection with an analysis based on an improved theory, led to the desired knowledge of both the structure in solution and the absolute configuration. To this end, once again the circular dichroism of the (trisethylenediamine)cobalt(III) cation [56] and of the (trisoxalato)cobaltate(III) anion [57] was studied. But to trace this development would far exceed the scope of the present overview.

1.8. THE COTTON EFFECT IN ORGANIC CHEMISTRY, A RUSSIAN DOMAIN It would not be correct to claim that organic chemists neglected optical activity in the early twentieth century, except to characterize compounds by their D-line rotation. As an example to the contrary can be cited a series of papers by the Swiss chemist Hans Rupe (1866–1951) on the influence of the constitution on the rotatory power of optically active

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

compounds, starting in 1903 [58]. However, usually colorless solutions were examined in the visible, as also done by Rupe, and could provide data on “plain” rotatory dispersion only. These investigations are, therefore, outside of the scope of the present overview. However, it must be admitted that, in general, organic chemists seem to have been less interested in chiroptical effects than their inorganic or physicochemical colleagues. Thus the study of organic molecules, as much as there is, had to rest on the “good will” of people from the latter fold. Here, credit is to be given first to the distinguished English physical chemist, Thomas Martin Lowry (1874–1936). Lowry’s interest in optical activity dates back to 1898, when he noted the change of optical rotation on nitrocamphor with time and introduced the term “mutarotation” to characterize this phenomenon [59]. He greatly improved the mathematical treatment and the theoretical understanding of rotatory dispersion and circular dichroism, based in part on the experimental data collected in his own research group in London, and later in Cambridge. It is interesting to see that there was a certain lack of understanding of the theory of optical activity on the part of some organic chemists. Even Rupe himself, for example, maintained as late as in 1921 that there had not been established with certainty any connection between the Cotton effect and the absorption of light [60]. In the earlier period of his career, Lowry dealt mostly with features outside of absorption bands; that is, he did not penetrate by experiment into the Cotton effect region itself. The state of the research on optical activity by the year 1914 has been summarized in the report “Optical Rotatory Power. A general Discussion” [61], and later in Lowry’s classic book [21], and it need not be described further in this paper. Lowry’s work in the early 1930s on the Cotton effect of organic molecules will be discussed later in this section. The most active pioneer in the study of the rotatory dispersion of organic molecules, and the only one who obtained data for the Cotton effect before World War I, is undoubtedly the Russian Leo [Lev] Alexandrovitsch Tschugaev (1873–1922) [62]. Tschugaev was a prolific research worker, who from the beginning of his career engaged in the chemistry of compounds like terpenes and camphor and, secondly, that of transition metal complexes. It was probably the study of optically active natural products that aroused his interest in optical activity generally. Eventually, he turned to a third research topic, after he had become Professor of Inorganic Chemistry at the Imperial University of St. Petersburg, and began in 1909 a series of papers on (anomalous) rotatory dispersion [63]. Tschugaev was fully aware of Cotton’s ground-breaking discoveries, and he was also aware of the problems inherently connected with the samples chosen for this early work. Therefore, he proudly, and correctly, pointed out in his initial papers that he now for the first time employed well-defined compounds for the study of the rotatory dispersion (he himself had no instrument that would allow him to measure the circular dichroism, in addition). But also he was still limited to visual observations at certain spectral lines. He used samples from two different families of sulfur-containing colored derivatives of optically active terpene alcohols like borneol, menthol, or fenchol, namely, xanthates RO–C(S)–SR and related compounds, along with “dithiourethanes” RO–C(S)–NPh–C(S)Ph (with Ph = C6 H5 ) and similar compounds. In all cases he found an anomalous dispersion of the rotation, but for different reasons. The xanthates give colorless or yellowish solutions, because there are no absorption bands in the visible. The anomalous rotatory dispersion detected is, therefore, of the type already observed for tartaric acid by Biot, and it is not caused by a Cotton effect in this spectral region. The red dithiourethanes, however, do show an absorption band at ∼520 nm, and the rotatory dispersion features indeed result from a Cotton effect associated with this absorption.

17

18

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

It must be admitted that Tschugaev was only able to detect the first maximum of the ORD curve, because of the onset of high absorption near 450 nm precluding measurements of the short-wavelength part of the dispersion curve. Therefore, it might be considered stretching a point rather far to speak of his truly having detected Cotton effects, since this depends on quite some interpolation and interpretation. But his contention was much strengthened by G. Bruhat, to whom he had sent samples of both the d - and the l -bornyl dithiourethane mentioned above. Bruhat was able to measure the circular dichroism, both in toluene solution and in the melt [35, 64]. The CD maximum at 520–530 nm coincided with the absorption maximum at 520 nm. Bruhat could confirm, but not extend, Tschugaev’s ORD data, by the way. This then not only proved that Tschugaev’s interpretation of his rotatory dispersion curves had been correct, but also provided the first example of circular dichroism observed in a well-defined (organic) compound. With Tschugaev’s papers at hand, it is interesting to follow the progress of his search for a fitting technical term to describe what he initially calls an anomalous rotatory dispersion “in the sense of Cotton,” until he finally in 1912 arrived at the “Cotton phenomenon.” This then became the internationally accepted term to be used for two decades, until it lost ground against Lifschitz’s “Cotton effect.” T. M. Lowry revisited Tschugaev’s compounds in 1932, confirming the earlier data and extending the wavelength range of the observations, thanks to improvements in the optical instrumentation and the introduction of photography. He also performed calculations on a more advanced basis in order to analyze and simulate the data [65]. Now it had become possible to take photographic readings at many points of the wavelength scale down to 325 nm. Lowry not only supplemented and slightly extended Tschugaev’s earlier ORD results, but now he could additionally make available circular dichroism data for the xanthates. These compounds, with a weak absorption at 360 nm (which Tschugaev had missed) and a strong one at 280 nm, exhibit a CD maximum at ∼355 nm. The steep rise of the absorption toward shorter wavelengths still precluded the precise observation of the second ORD maximum at about 330 nm. It is notable that Lowry found his photographic CD measurements of the dithiourethanes in the year 1932 less exact than Bruhat’s visual measurements of 1911. Colored organic compounds were not unknown apart from the sulfur-containing derivatives discussed above, but were not easily available in optically active form. For chiroptical studies they should advantageously stem from the pool of optically active natural compounds or their derivatives, because the organic chemists of those days seem to have tended to avoid resolutions, contrary to their colleagues in the field of Werner complexes. Therefore, it is not surprising that the yellow camphor quinone attracted the attention of physical chemists and physicists alike. The Russian physicist Nina Wedeneewa detected in this compound a Cotton effect near 490 nm by ellipticity and rotatory dispersion measurements; although the work had been done in 1919, its publication was delayed until 1923 because of the political turmoil in Russia [20]. Her main interest was the analysis of the data in terms of the Drude theory of optical dispersion. Slightly later, in 1925, Israel Lifschitz also had tried to measure the optical rotation of camphor quinone near the absorption region, but could reach the first maximum only [53]. Lowry later repeated, confirmed, and extended Nina Wedeneewa’s findings, thereby tacitly correcting the sign of her circular dichroism data, as he studied camphor quinone both in solution and in the vapor phase [66]. Also this second case, in which the circular dichroism of an organic molecule has been measured successfully, in addition to Tschugaev’s/Bruhat’s dithiourethane,

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

stems from Russia. Unfortunately, the First World War, the Russian Revolution, and Tschugaev’s premature death interrupted and eventually ended a period of intensive research on chiroptics in that country. Intriguing colored organic compounds are the “nitrosites” (1,2-nitroso-nitrites) and “pseudo-nitrosites” (1,2-nitroso-nitro compounds), which can be made by the addition of (formally) N2 O3 to olefins. These compounds exhibit green or blue colors in solution, provided that the nitroso group is monomeric. E. Deussen, working with the blue optically active caryophyllene nitrosite, had noticed an anomalous rotatory dispersion, minimalistically based on readings at just two wavelengths [67]. The suggestion by Tschugaev, that this might be caused by a Cotton effect in the visible [68], led Stotherd Mitchell in Glasgow in 1928 to check the chiroptical properties of this compound [69], but with readings taken at eight wavelengths between 691 and 436 nm. Indeed, a Cotton effect was found in the CD at ∼680 nm. It might be remarked that the absorption curves were measured separately with right- and left-circularly polarized light, whereas in practically all earlier cases data on ellipticities had been collected. In summary, the circular dichroism of only three (colored) organic compounds had become known by 1928. Although in some cases the claim has been put forward to have seen in such compounds a Cotton effect by rotatory dispersion, this should be taken with a grain of salt, because this cannot normally be substantiated by presenting the whole sigmoidal dispersion curve. Usually one wing of the curve is missing (commonly on the high-energy side of the absorption band), for reasons discussed above. It may come as a surprise, therefore, that as early as 1910, Eug`ene Darmois, one of Cotton’s students, published the ORD Cotton effect of even a colorless organic compound, namely, camphor, at ∼300 nm [70]. It is true that he had not been able to obtain rotation data at the absorption maximum itself, but only on either side of it (with a gap between 313 and 265 nm), but the dispersion curve can be easily completed by interpolation. This finding is all the more remarkable, because it demonstrated for the first time the possibility of taking rotation values down to ∼250 nm in the ultraviolet. Regrettably, the response of the chemical community was slow, probably because of the technical difficulties involved in the construction of a suitable spectropolarimeter. It took some 20 years before Darmois’s work could be taken up again; then, W. Kuhn published ORD, CD, and UV data of camphor, taken right through the absorption band [71], and T. M. Lowry likewise reported data of the related camphor-β-sulfonic acid [72]. It seems rather daring that Darmois [70] also tried to measure the chiroptical properties of olefins like α- and β-pinene or limonene, but—not unexpectedly—without much success. This chromophore still resisted the efforts of R. Servant in 1932, but at least this time some indication of a first ORD maximum seemed to be suggested at around 280 nm in the case of the pinenes [73]. From the foregoing it appears that rather suddenly, by around 1930, many more types of compounds were studied by chiroptical techniques. This resulted from the progress in instrumentation, to be related briefly in the next section. Now, a great many additional chromophores in colorless organic compounds, like nitro, azido, nitrito, and particularly carbonyl groups, opened the way for studies of the circular dichroism and the rotatory dispersion. Nevertheless, just a few typical samples were used to be investigated for physical–chemical purposes. It would take many more years before organic chemists made use of the now accessible optical techniques for stereochemical correlations and the determination of the absolute configuration. But this then is far outside the scope of this overview. These newer developments since the early 1930s will not be treated here, because many summaries are already available. The major source of information

19

20

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

for the organic chemist on these later chiroptical investigations is Carl Djerassi’s book [74], which has become a “classic” in the field to which the reader should turn.

1.9. ADVANCES IN INSTRUMENTATION AND THEORY; THE WAY INTO THE ULTRAVIOLET Very briefly, the technical advances shall be sketched here that, around 1930, led to the (short-lived) upsurge in chiroptical activity mentioned in the foregoing section. Details shall be omitted, because comprehensive reviews are readily available. In addition to Lowry’s encyclopedic monograph of 1935 [21] and Mitchell’s book of 1933 [22], one should pay particular attention to Bruhat’s 1930 treatise on polarimetry, since most of the “newer” development of spectropolarimeters and ellipsometers resulted from the efforts at his laboratory in Paris [75]. Another active center was at the Technische Hochschule (Institute of Technology) Karlsruhe, Germany, where the Swiss physical chemist Werner Kuhn not only worked on theoretical and experimental chiroptics, but also developed his own apparatus [76]. Relevant optical instruments have also been reviewed by R. Descamps in Brussels, Belgium, who had himself constructed and perfected a spectropolarimeter for the UV region [77]. Of course, Lowry’s important contributions from his Laboratory of Physical Chemistry in the University of Cambridge, UK, are by no means to be forgotten. Progress in instrumentation for chiroptical studies meant, besides the obvious improvement of sensitivity and reliability, by and large the extension of the wavelength range into the ultraviolet. It was evident that the absorption bands of the vast majority of chemical compounds are located in the UV. It may come as a surprise that polarimetric measurements in the ultraviolet have been known for a considerable time and can be traced back to the nineteenth century. Various instruments for this purpose are described by Lowry [21], but the operation of the apparatus was laborious, the accuracy of the results questionable, and the accessible wavelength range rather limited. It does not seem, moreover, that these techniques have been applied to the detection of Cotton effects before the exceptional pioneering work of E. Darmois in 1911 [70]. One of the earlier attempts at measuring optical rotations in the UV was by Lowry himself, who in 1908 combined a half-shadow polarimeter with a UV spectrograph [78]. Darmois could collect data to the wavelength limit of 250 nm. But in order to make this wavelength range more generally accessible, and even extend it toward higher energy, three different problems had to be solved: first, with regard to the light sources; second, with the transparency of optical components to the UV light; and third, in connection with the efficiency of detectors. 1. The solar spectrum provided light in the laboratory only down to the limit of 300 nm, because of atmospheric absorption. Mercury vapor lamps allowed readings to be taken for another 50 nm, down to 250 nm. Beyond that wavelength, various other light sources have been used, in all cases providing an array of separate spectral lines—for example, the iron arc to 233 nm, or the cadmium spark to 210 nm. 2. The transparency limits of glass preclude its use in UV instruments for lenses, prisms, or similar optical devices, except for the near-UV range. Materials like Iceland spar (cutoff at 250 nm), fluorspar, or quartz had to be introduced to improve the transmittance of UV radiation. Even the Canada balsam used in

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

conventional Nicol prisms had to be replaced by, for example, glycerol, in order to allow their utilization below 340 nm. 3. As to the detectors, the fluorescent screens of the early days were soon replaced by photographic implements, but the evaluation of the developed plates posed new problems in its own right. The improvements in all these areas led to the independent construction of two similar recording spectropolarimeters in the year 1926, giving access to UV measurements up to the high energy limit of 250 nm [79]. An important objective in the construction of these instruments was the desire to hold to a minimum the number of optical parts that had to be transparent to UV light. In Cotton’s photographic spectropolarimeter, the optical density of the photographs had to be evaluated by means of a photometer or, better, by a microphotometer. Although the recording was automatic, it was not continuous, but consisted of discrete exposures. Bruhat, on his part, used a photoelectric device. The data analysis could still be cumbersome, but the measurement of rotations in the UV had nevertheless become relatively easy. Bruhat’s polarimeter was perfected still further [80], to the extent that the photoelectric measurements became superior in precision to those obtained by the photographic method applied by Servant [73]. All the instruments mentioned so far are polarimeters. Similar advances had taken place in the construction of ellipsometers. Again, Bruhat was in the forefront with the development of a polarimeter–ellipsometer [81]. This widely used visual instrument consisted of an ordinary polarimeter, fitted with a mica λ/4 plate. Werner Kuhn developed an even more advanced photographic device for use in the ultraviolet [82]. It contained optical parts of quartz and fluorite only, and it allowed measurements to be taken all the way to 190 nm. This apparatus was the preferred instrument for many years to come and was marketed by the well-known makers of optical instruments in Berlin, Schmidt & Haensch. A similar ellipsometer was described by Mathieu in Paris [83]; it was designed particularly for the wavelength range of 450–280 nm. In conclusion, in the beginning of the 1930s, the development of instrumentation to measure the optical rotatory dispersion and the circular dichroism (in terms of ellipticities) had progressed to a state of perfection that was hard to improve upon in the decades to come. But, although it had been convincingly shown now that the Cotton effects of a wealth of optically active (organic) compounds were accessible, the chemical community at large was slow to make use of the chiroptical techniques. This resulted probably in part from the fact that the importance of stereochemistry had not yet been realized widely, especially among organic chemists. Also, the necessary apparatus still had to be built individually, and the measurement of ORD and CD was still far from being routine. Only with the advent of commercial recording instruments for ORD [84] and CD [85] many years later was the field of chiroptical investigations opened to the general chemist. Very few sentences must suffice on the contemporary development of the theory of optical activity, because this topic lies far away from the present overview. Paul Drude’s theory of optical activity in isotropic media, as expanded in his famous book of 1900 [26], has been the standard with which most physical(–chemical) research had to contend for the first three decades of the twentieth century. Many experimental investigations, including Lowry’s, have been motivated by the search for an improved version of the “Drude equation” of optical activity. An important step forward in this line was taken by L. Natanson, who in the year 1909 succeeded in deriving an equation that could approximate the optical dispersion within the absorption band, a problem that Drude’s original equation could not handle [17, 27, 28]. An advanced theory of optical activity

21

22

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

was published by Max Born (1882–1970, Nobel Prize 1954) in 1915 [86] and at about the same time by C. W. Oseen [87] and F. Gray [88]. Although the importance of these papers, especially Born’s, was immediately recognized, practically no direct influence on the study of chiroptical properties was engendered. Likewise, the important paper by the Belgian physicist L´eon Rosenfeld (1904–1974) on the quantum-mechanical theory of optical activity did not affect the chemical community at the time [89]. It took Werner Kuhn’s (1899–1963) simplification of Born’s theory into his coupled-oscillator model to attract the chemists’ attention [76, 90]. Now the application of theoretical methods to actual problems of stereochemistry seemed to become realistic. Even the determination of the absolute configuration of molecules might come into reach. Indeed, Kuhn’s conclusion that (−)-butan-2-ol had the (R) configuration was the culminating point of his theoretical work [91]. A comprehensive overview of these developments in theory is found in Mathieu’s monograph on the molecular theories of natural optical activity [92].

1.10. SOME WORDS ON NOMENCLATURE: COTTON EFFECT, OPTICAL ROTATORY DISPERSION, CD, ORD The expression “Cotton effect” (originally in German) was introduced by Israel Lifschitz in 1922. The earlier technical term, first used by Tschugaev in 1912 (also originally in German), is “Cotton’s phenomenon.” This latter expression developed gradually by the contraction of phrases like “anomale Rotationsdispersion im Sinne A. Cottons” (anomalous rotation dispersion in the sense of A. Cotton) [93] or “la mani`ere de voir de M. Cotton” (literally: Mr. Cotton’s way of viewing) and “ph´enom`ene de la dispersion anormale” (phenomenon of anomalous dispersion) [94]. Tschugaev never used the word “Cotton effect” in whatever language. It is slightly confusing, therefore, to find this term in his Russian collected works [95], but on closer inspection this turns out to appear in a posthumous translation of his German papers into Russian. He himself always wrote ” (i.e., Cotton’s phenomenon). in Russian “ As can be seen from Table 1.1, both expressions were used side by side for a period of 10 years, but eventually “Cotton effect” was victorious. The situation is slightly simpler with “rotatory dispersion.” From the beginning in 1877, this term had become established, with minor variations as shown in Table 1.2. The only point meriting some attention is the question, Why has rotatory dispersion nowadays become “optical rotatory dispersion”? The “optical” was originally added in order to differentiate between the two effects of magnetic and optical rotatory dispersion, both of which were the subject of a series of papers by Lowry, with the first one appearing in 1913. Usually, however, it was considered unnecessary to point to this difference, because magnetic rotatory dispersion rarely plays a role in chemistry. The modern habit of always referring to “optical rotatory dispersion”, which also led to the common abbreviation “ORD,” seems to originate with Carl Djerassi, who used it since 1955 in a great many papers and who publicized it further by his textbook on ORD. The counterpart expression with respect to circular dichroism should be “optical circular dichroism,” in order to likewise differentiate between optical and magnetic circular dichroism. Indeed, the French “fathers” of modern CD instruments and their application in chemistry have used the term “dichro¨ısme circulaire optique” [96], but for reasons unknown, in this case the chemical community has continued to ignore the “optical.” For the first half of the twentieth century, abbreviations were not much used in (physical) chemistry. With fashions changing, this became a craze, however, in the second half, especially in the United States and the Soviet Union. Typical is the general use

23

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

TAB L E 1.1. The Change from Cotton’s Phenomenon to the Cotton Effect The Cotton Phenomenon 1912 1913 1914

1915 1922 1922 [1925]

1926

1930, 1933

L. Tschugaeff: Cottonsches Ph¨anomena L. Tchougaeff: ph´enom`ene Cottonb L. Tschugaeff: Cotton’s phenomenonc M. Del´epine: les ph´enom`enes d´ecouverts par M. Cottond H. Rupe: Cottonsches Ph¨anomene H. Grossmann: Cotton-Ph¨anomenf F. M. Jaeger: ph´enom`ene de Cottonh

The Cotton Effect

1922

J. Lifschitz: Cottoneffektg

1922 [1925] 1923

F. M. Jaeger: l’effet de M. Cottonh F. M. Jaeger: l’effet Cottoni

1928 1929 1930

S. Mitchell: Cotton Effectk W. Kuhn: COTTON-Effektl G. Bruhat: l’effet Cottonn

1935

T. M. Lowry: Cotton Effecto

W. Pfleiderer: Cottonsches Ph¨anomenj

T. M. Lowry: Cotton phenomenonm

Cotton effect (Engl.); Cottoneffekt (German, Danish); effet Cotton (French); Effeto Cotton (Italian); Efekt Cottona (Polish); (Russian) a L. Tschugaeff, J. Prakt. Chem. 1912, [2] 86 , 545–550; L. Tschugaeff, G. Glinin, Ber. Dtsch. Chem. Ges. 1912, 45 , 2759–2764. b L. Tchougaeff, A. Kirpitcheff, Bull. Soc. Chim. Fr. 1913, [4] 13 , 796–803. c L. Tschugaeff, Trans. Faraday Soc. 1914, 10 , 70–79. d M. Del´ epine, C. R. H. Acad. Sci . 1914, 159 , 239–241. e H. Rupe, Liebigs Ann. Chem. 1915, 409 , 327–357. f H. Grossmann, M. Wreschner, Die anomale Rotationsdispersion, Sammlung chem. u. chem.-techn. Vortr¨ age, W. Herz, ed., Enke, Stuttgart, 1922, 26 , 259–314. g J. Lifschitz, Rec. Trav. Chim. Pays-Bas 1922, 41 , 627–636. h F. M. Jaeger, Rapp. Disc. Inst. Int. Chimie Solvay (Conseil Chim. 1922, Bruxelles), Gauthier-Villars, Paris, 1925, 199–202. i F. M. Jaeger, Bull. Soc. Chim. Fr. 1923, [4] 33 , 853–889. j W. Pfleiderer, Z. Phys. 1926, 39 , 663–685. k S. Mitchell, J. Chem. Soc. London 1928, 3258–3260. l W. Kuhn, Z. Phys. Chem. 1929, B4 , 14–36. m T. M. Lowry, Trans. Faraday Soc. 1930, 26 , 266–271; T. M. Lowry, H. Hudson, Philos. Trans. 1933, A232, 117–154. n G. Bruhat, Trait´e de Polarim´etrie, Editions de la Revue d’Optique, Paris, 1930. o T. M. Lowry, Optical Rotatory Power, Longmans, Green and Co., London, 1935; reprint: Dover Publications, New York, 1964.

of “RD” in C. Djerassi’s book published in 1960, incidentally one of the first major applications of this abbreviation. It is amusing to note that here “RD” still refers to “rotatory dispersion,” of course, while the same author had already adopted “optical rotatory dispersion” for some years. It was consistent with adding “optical” to “rotatory dispersion,” to also add the letter “O” to the abbreviation “RD”: The birth of “ORD” took place in Djerassi’s environment in the early 1960s.

24

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 1.2. From Rotatory Dispersion to Optical Rotatory Dispersion Rotatory Dispersion, Dispersion Rotatoire, Rotationsdispersion

1913 1914 1915 1917 1935

H. Landolta: (anomale) Rotationsdispersion [German] A. Cottonb: dispersion rotatoire (anomale) [French] H. Grossmannc: (anormale) Rotationsdispersion [German] L. Tschugaeffd: anomale Rotationsdispersion im Sinne A. Cottons [German] L. Tschugaeffe: dispersion rotatoire ano(r)male [French] T. M. Lowryf : rotatory dispersion [English] T. S. Pattersong: (normal/abnormal) rotation-dispersion [English] T. M. Lowryh: an exact definition of normal and anomalous rotatory dispersion F. M. Jaegeri : rotatie-dispersie [Dutch] T. M. Lowryj : (normal and anomalous) rotatory dispersion [English]

1913 1926 1955 1959 1960

T. M. Lowryk: magnetic and optical rotatory dispersion [English] W. Pfleidererl : optische/magnetische Rotationsdispersion [German] C. Djerassim: Optical Rotatory Dispersion Studies [1st paper] [English] F. Woldbyen: optical rotatory dispersion [English] C. Djerassi: Optical Rotatory Dispersion, McGraw-Hill, New York [English]

1877 1895 1908 1911

Optical Rotatory Dispersion

a

H. Landolt, Liebigs Ann. Chem. 1877, 189 , 241–337 (p. 274). Cotton, C. R. H. Acad. Sci . 1895, 120 , 989–991. c H. Grossmann, H. Loeb, Z. Ver. Deutsch. Zuckerind., Allg. Teil 1908, 58 , 994–1009. d L. Tschugaeff, Z. Phys. Chem. 1911, 76 , 469–483. e L. Tschugaeff, A. Ogorodnikoff, Ann. Chim. Phys. 1911, [8] 22 , 137–144. f T. M. Lowry, J. Chem. Soc. London 1913, 103 , 1062–1067. g T. S. Patterson, Trans. Faraday Soc. 1914, 10 , 111–117. h T. M. Lowry, J. Chem. Soc. London 1915, 107 , 1195–1202. i F. M. Jaeger, Chem. Weekbl . 1917, 14 , 706–732. j T. M. Lowry, Optical Rotatory Power, Longmans, Green and Co., London, 1935. k T. M. Lowry, T. W. Dickson, J. Chem. Soc. London 1913, 103 , 1067–1075. l W. Pfleiderer, Z. Phys. 1926, 39 , 663–685. m C. Djerassi, E. W. Foltz, A. E. Lippman, J. Am. Chem. Soc. 1955, 77 , 4354–4359. n F. Woldbye, Acta Chem. Scand . 1959, 13 , 2137–2139. b A.

The abbreviation “CD” for circular dichroism was introduced at about the same time. Whereas “OCD” has never become popular, “MCD” for magnetic circular dichroism is a logical and accepted abridgment. In the same vein, a reasonable abbreviation for magnetic rotatory dispersion should be “MRD”. The misnomer “MORD” (which in German means “murder,” by the way) should fall into disuse. Details concerning these various abbreviations are collected in Table 1.3. It remains to remind the reader that the terms “rotatory dispersion” and “circular dichroism” as well as their abbreviations might change, if languages other than English are used. Some examples are collated in Table 1.4.

1.11. BIOGRAPHICAL NOTICES: G. BRUHAT, A. COTTON, W. KUHN, I. LIFSCHITZ, T. M. LOWRY, L. NATANSON, AND L. TSCHUGAEV It is a moot question how far a treatise on the history of natural sciences ought to be supplemented by a personalized account. It is the present author’s contention that such an approach is helpful to provide a balanced background for the scientific results

25

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

TAB L E 1.3. From RD to ORD (but not from CD to OCD) 1960

C. Djerassi,a Optical Rotatory Dispersion: “abbreviated RD (curves)”

1960

W. Klyne, Adv. Org. Chem. 1 , 239–348: “abbreviated R.D. curves”

1965

P. Crabb´e,b Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry: “circular dichroism, abbreviated CD; rotatory dispersion curves, abbreviated RD”

1965

L. Velluz, M. Legrand, M. Grosjean,c Optical Circular Dichroism: “CD-curves”

1963

D. Lightner,d PhD Thesis: ORD

1966

K. Mislow,e Introduction to Stereochemistry: ORD, CD [preface September 1964, © 1965, published 1966]

1967

G. Snatzke,f ed., Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry [Summer School, Bonn, 24 September–1 Oct., 1965]: ORD, CD; “MORD”, MCD

1972

P. Crabb´eg, ORD and CD in Chemistry and Biochemistry

a

McGraw-Hill, New York. Holden-Day, San Francisco. c Verlag Chemie, Weinheim/Academic Press, New York. d Stanford, CA. e W. A. Benjamin, New York. f Heyden & Son, London. g Academic Press, New York. b

TAB L E 1.4. ‘CD’’ and ‘‘ORD’’ [and ‘‘CE’’] in English, French, German, and Russian CD RD ORD

DC DR DRO

КД

circular dichroism, Circulardichroismus rotatory dispersion, Rotationsdispersion optical rotatory dispersion, optische Rotationsdispersion Круговой дихроизм

dichro¨ısme circulaire dispersion rotatoire dispersion rotatoire optique

ДОВ

Дисперсия оптического вращения

[CE

Cotton effect, Cotton-Effekt]

[EC

effet Cotton]

discussed in the foregoing sections. After all, the common expression “it was found that” is deceptive insofar as it tends to obscure the fact that these results were not “found” by anonymous agencies, but earned by individual scientists, working at a specific time and under specific circumstances that pertain to both their professional and their private lives. Among the many actors in the first decades after the discovery of the Cotton effect, who could be considered for some biographical notices, here seven scientists are selected, as listed in the section title in alphabetical order. These names may reflect some personal predilection, but it is not too difficult to defend these choices impartially: Cotton is, of course, of outstanding importance. Natanson opened the route to an understanding of the interconnection of CD, ORD, and absorption generally. Tschugaev pioneered the investigation of chiroptical properties of organic molecules, while Lifschitz did the same with coordination compounds and has, moreover, coined the technical term “Cotton effect.” Bruhat was of particular importance for the development of instruments for chiroptical studies, together with Lowry and Kuhn, while the two latter scientists (both of them chemists, by the way) were instrumental also for the progress in the application

26

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the theory of optical activity to chemical problems. All of these persons earned universal recognition, and even fame, also in other fields of chemistry or physics. All of them would merit attention, even if their part in the progress of chiroptics were to be neglected. Actually, for some among them, their activity in this field was rather incidental, if seen against their lifetime achievements. In the following, summary biographies of these scholars will be related in sequence according to their date of birth, with Cotton’s biography at the end, however. The theoretical physicist Ladisla(u)s [Władysław] Natanson (1864–1937) was born in Warszawa (Warsaw), Poland, the son of a medical doctor. Warsaw was under Russian rule at the time, and Natanson was a Russian subject. After graduation from a classical school in 1882, he enrolled as a student at the Faculty of Sciences of the University of St. Petersburg, where he became a “Candidate” (Licentiate) in 1886. After a few months at the Cavendish Laboratory at Cambridge, U.K., he returned to Imperial Russia in order to fulfil the requirements necessary for Russian subjects who wanted to embark on an academic career. Therefore, in 1887 he moved to Dorpat, Livonia (now Tartu, Estonia), to work for his doctorate with the physicist Professor A. von Oettingen, which he obtained in 1888. Incidentally, the official language at the University of Dorpat was German. After some postdoctoral studies with L. Boltzmann at the University of Graz, Austria, he returned to Warsaw to write an “Introduction to Theoretical Physics” [97]. This book was met with much acclaim, which helped him to be granted a position at the Jagiellonian University Krak´ow (Cracow), Poland (then under Austrian rule). There he moved up through the academic ranks: 1894 Titular Professor, 1899 Extraordinary Professor, and finally 1902 Professor of Theoretical Physics. Later he was also appointed Dean of the Faculty, and in 1922/23 he became Rector of the University. Elected to the Academy of Sciences in 1893, he became President of the Section of Mathematical and Natural Sciences in 1926, until he resigned from both the Academy and the University in 1935 for health reasons. His professional achievements are expanded upon in the obituary notice by L´eon Klecki, which includes a bibliography [98]. The chemist Lev Aleksandrovi(ts)ch Tschugaev (Chugaev) (1873–1922) was born in Moscow, Russia. After completing his studies at the University of Moscow in 1894, he became Assistant at the Bacteriological Institute of the University, where he started his research on the optical activity of organic compounds. From the beginning, Tschugaev strove to be competent in organic as well as inorganic chemistry. His master’s thesis in 1903 dealt with studies in the terpene and camphor series, while in his doctoral thesis (habilitation) of 1906 he presented results from coordination chemistry. Throughout his life, he successfully followed this dichotomy in his research, with some excursions into physical chemistry. In 1906 he was appointed Professor at the Technical University in Moscow, and in 1908 he was called to the Chair of Inorganic Chemistry at the Imperial University of St. Petersburg (later: University of Petrograd). He held this position at the time of his death. During World War I and the Russian Revolution he was mainly active in the field of technical and applied chemistry. He was one of the founders of the Institute of Applied Chemistry in Petrograd and became its director. In the aftermath of the revolution, he died of typhoid fever at Wologda, Russia, at the age of not yet 50. Obituary notices have appeared in England [62] and in Germany [99], acknowledging his abundant contributions to chemistry. Attention should also be paid to his “Selected Works” [95]. The physical chemist Thomas Martin Lowry (1874–1936) was born at Low Moor, Bradford, Yorks., U.K., the son of a Wesleyan Chaplain. He studied at the Central Technical College, South Kensington, London. From 1896 to 1913 he was assistant to Professor

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

H. E. Armstrong and from 1904–1913 he was Lecturer in Chemistry at the Westminster Training College. In 1913 he became Head of the Chemical Department in Guy’s Hospital Medical School and Professor of the University of London. In 1920, finally, he was appointed to the newly created Chair of Physical Chemistry at Cambridge University. He continued at Cambridge for the rest of his life. In his extensive obituary notice, W. J. Pope wrote [100]: “[The book] on “Optical Rotatory Power” was issued in 1935 and will long remain a standard work on the subject. The immense amount of accurate experimental work which Lowry has left on record secures him a permanent place in the history of the science to which he was devoted.” But, although chiroptical methods are central to this overview, Lowry’s important contributions to other areas should not be forgotten. Mention may be made of his studies of the polymorphism of inorganic salts, his “Studies of Valency” and of the nitrogen oxide/water system, the nature of the sulfur halides, and the stereochemistry of tellurium compounds. In all cases he tried to apply modern physical concepts to chemical problems. Lowry was a member of the Faraday Society from its beginnings in 1903 and acted as its president in 1928–1930. He became a Fellow of the Royal Society in 1914. The physicist Georges Bruhat (1887–1945) was born in Besanc¸on, France, the ´ son of a civil servant. He was admitted with honors into the renowned Ecole Normale Sup´erieure (ENS) in Paris in 1906. After having obtained his B.Sc. (“licence e` s sciences physiques”) at the University of Paris, he acquired the qualification to teach at secondary schools (as “Professeur agr´eg´e ”) and taught at a high school in Paris for one year. Perhaps this interval in his scientific career awakened his interest in teaching and in writing textbooks for this purpose. He then got a position as laboratory assistant (“pr´eparateur”) at the ENS, which enabled him to work under the guidance of A. Cotton on his doctoral dissertation: “La dispersion anormale du pouvoir rotatoire mol´eculaire” (the anomalous dispersion of the molecular rotatory power). After an interruption by his military service during the First World War, he could continue his academic career in 1919 in Lille, France, where he was Professor of General Physics from 1921 to 1927. His successor was Marcel Pauthenier, by the way, who had been his partner in the construction of UV spectropolarimeters [79]. Bruhat returned to the University of Paris in 1929 as Lecturer and Professor Extraordinary. In 1938 he was promoted to the Chair of Theoretical and Celestial Physics. During this time he published four compendious textbooks on general physics: Electricity (1924), Thermodynamics (1926), Optics (1930), and Mechanics (1934). These books have become standard texts in French universities, with many editions; Optics, for example, has been reedited as recently as 2004. Bruhat also continued his association with the ENS, serving as “Sub-Director” from 1935, and as acting director during World War II. In the beginning of August 1944, he was arrested by the political police (“Gestapo”) of the German occupation powers and held prisoner in lieu of a student accused of activities in the French Resistance Movement. He was taken to Germany into a concentration camp and died there on January 1, 1945, of pneumonia and exhaustion. The chemist Israel Lifschitz (1888–1951) was born in Shklov, Russia (now Belarus) (see Figure 1.4). His family was German and lived in Leipzig, Germany. His mother had moved to Shklov just for her confinement, motivated by family regards. He studied chemistry at the University of Leipzig and worked there for his doctoral degree under the guidance of A. Hantzsch. His dissertation of 1911 dealt with the spectroscopic properties of various organic nitrogen compounds. Faithful to the ideas already developed in this dissertation, his main interest became the correlation of chemical constitution and bonding with electronic absorption. Although by training he was an organic chemist, his research

27

28

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 1.4. Israel Lifschitz. (Photograph provided by Ms. E. H. Lifschitz, Haifa, and reproduced with her permission.)

led him far into physical chemistry and, later, also inorganic chemistry. Leaving Leipzig in 1911, he moved to the University of Zurich, Switzerland, for his habilitation. To this end, he submitted a second dissertation on the changes of light absorption by organic acids on salt formation. As a result, the title of “Private Docent of Chemistry” was conferred on him in 1914, allowing him to do independent research, but not connected with any paid position. The life of a Private Docent was difficult, unless one was privately affluent. The economic troubles after World War I also forced Lifschitz to interrupt his work at the university in 1920, in order to look for a source of income outside of Academia, to support his growing family. Lifschitz extended his investigations to include the optical rotatory dispersion of transition metal coordination compounds (see, e.g., paper V of a series on the function of chromophores [101]); after 1920 this became his central area of research. From these papers it can be deduced that he had established a cooperative arrangement with the Dutch inorganic stereochemist F. M. Jaeger at the Rijksuniversiteit Groningen (RUG) in the Netherlands, by 1919 at the latest. Jaeger invited Lifschitz to join his laboratory and offered him a tenured staff position as “Conservator”. This induced Lifschitz to move to Groningen, although he had become a Swiss citizen and probably had intended to stay in that country. As a result, in the summer of 1921, Lifschitz became Private Docent

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

of Electrochemistry and Photochemistry, as well as Conservator at the Laboratory of Inorganic and Physical Chemistry at RUG. He helped to develop the department into a center of chiroptical spectroscopy. RUG became the birthplace of the “Cotton effect,” since Lifschitz coined this technical term there. The relevance of his research was readily acknowledged by stereochemists and physical chemists at other institutions. Already in 1935, J.-P. Mathieu in Paris had pointed out that Lifschitz was the first to systematically investigate the relations between the experimental data on the Cotton effect and the chemical bonds in optically active compounds [83]. Lifschitz held his position at RUG until he was dismissed in November 1940 for political reasons. Under the German occupation of the Netherlands in World War II, the situation of Jewish people deteriorated continually. Since Lifschitz was unable to continue his research at RUG, he turned again to his private studies of the mystical movements in Judaism, of which he was a dedicated and competent scholar. He was a deeply religious person, who had also published on related subjects. Now he became absorbed again in his studies of Chassidism, and he gave even private lessons on the Zohar. All this ended, when he, his wife, and their five children were detained in February 1943. He was eventually deported to the Theresienstadt Concentration Camp in Bohemia in September 1944, separated from his three sons, two of whom did not survive the ordeal. Liberated at the end of the war, he, his wife, and the three children left returned to Groningen, where he was reinstalled at RUG. He died in the Netherlands in 1953 and was buried in Haifa, Israel. The dire story of the family’s fate has been reported by his elder daughter Esther Hadassa Lifschitz [102]. It had proved to be fatal that the Dutch authorities had insisted, shortly before the war, that the family accept Dutch citizenship while relinquishing their Swiss citizenship. An attempt to regain Swiss nationality during World War II failed. It is likely that they could have otherwise left for Switzerland. The present paper may perhaps draw some attention to Lifschitz, who has been undeservedly forgotten. The physical chemist Werner Kuhn (1899–1963) was born at Maur am Greifensee, Switzerland, the son of a pastor. He studied chemistry at the ETH Z¨urich and obtained his doctorate at the University of Zurich in 1923, with a dissertation on a photochemical topic. After working with Niels Bohr in Copenhagen, Denmark, for two years, he returned to Zurich for his habilitation in physical chemistry in 1927. He moved to Germany thereafter. For three years he worked at the University of Heidelberg, where he started his research on optical activity; in 1930–1936 he worked at the Technische Hochschule Karlsruhe, and in 1936 he occupied the Chair of Physical Chemistry at the University of Kiel. In the year 1939 he returned to his home country, when he was called to the Chair of Physical Chemistry at the University of Basel. Later he was also elected rector of the University. He remained there until the end of his life. A report on “Leben und Werk von Werner Kuhn” (life and work of W. K.), including a bibliography, appeared in 1984 in connection with a “Werner Kuhn Symposium” of the Swiss Chemical Society [103]. The physicist Aim´e Auguste Cotton (1869–1951) was born in the French provincial town of Bourg-en-Bresse, where his father taught mathematics (see Figure 1.5). He was ´ a student of physics at the Ecole Normale Sup´erieure (ENS) in Paris from 1890 to 1893. He completed his doctoral studies, in the course of which he discovered the “Cotton effect” at the Laboratory of Physics with M. Brillouin and J. Violle and earned the title “Docteur e` s-sciences physiques” in 1896. It is noteworthy that he included in his thesis his first attempts at measuring magnetic optical activity, since his future scientific career was to be centered on the physics of magnetism and magneto-optics. It would be in these fields, rather than in research on natural optical activity, that he would rise to eminence.

29

30

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 1.5. Aime´ Auguste Cotton. (From L. de Broglie, Notice sur la vie et l’œvre de Aime´ ´ Cotton, Academie des sciences—Institut de France, Paris, 1953; reproduced with permission.)

After a few years at the Faculty of Sciences of the University of Toulouse as Assistant/Associate Professor (Maˆıtre de Conf´erences, Professeur adjoint), he returned to the ENS in Paris, entrusted with the substitution of the Academician J. Violle, his teacher, for the period of 1900–1904. From 1904 to 1922 he served as Lecturer (Charg´e de cours) at the Faculty of Sciences of the University of Paris, delegated to the ENS. He was promoted to Professeur-adjoint in 1910 and became Professor of Theoretical and Celestial Physics at the Sorbonne in 1920. Finally, he was called to the Chair of General Physics there in 1922, succeeding G. Lippmann (Nobel Prize 1908). In 1923 he was elected a member of the illustrious Academy of Sciences as successor to J. Violle. He even became President of the Academy in 1938. Cotton retired from the Sorbonne in 1941, but continued until his death as Director of the Laboratory for Magneto-Optical Studies that he had founded in 1927. In France, A. Cotton is considered to be one of the eminent physicists of the twentieth century. He himself described his scientific aims and accomplishments in a 1923 pamphlet on the occasion of his election to the Academy [104]. As he pointed out, he had worked extensively on the Zeeman effect, but even more on the magneto-optical properties of colloids and molecular solutions. Many of the latter investigations were performed in

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

cooperation with the biologist Henri Mouton (1869–1935) of the Institute Pasteur (later Professor of Physical Chemistry at the University of Paris). Together they discovered in 1907 the important “Cotton-Mouton-Effect” (magnetic field induced linear birefringence) [105]. Cotton’s fame in physics rests, arguably, more on this discovery than on the “Cotton effect” in natural optical activity. Detailed information on his life and work is easily available thanks to various obituaries [106] and to a notice published on the occasion of the centenary of his birth [107]. Cotton received many honors and was awarded many prizes. He was nominated for the Physics Nobel Prize in the years 1915, 1916, 1920, 1922, 1925, 1927, 1928, 1929, 1930, 1931, 1932, 1933, 1934, 1944, and 1949, but without success [108]. His memory is kept alive in France with the “Laboratoire Aim´e Cotton,” as his former Laboratory for Magneto-Optical Studies has been renamed. It is now the Atomic and Molecular Physics Laboratory of CNRS, associated with the University Paris XI and situated on its campus in Orsay. Also the “Prix Aim´e Cotton” should be mentioned, which was established by the French Physical Society in 1953 in memory of A. Cotton and is awarded annually. In what esteem he is held by the French physicists is also apparent by the fact that he is one of the twelve “most eminent” physicists chosen by the Academy on the occasion of the World Year of Physics 2005 (WYP2005/UNESCO). Among chemists of countries other than France, and especially among the younger fold, the name “Cotton” rarely brings to their minds memories of a specific person, unfortunately, unless they mistake it for the name of the distinguished American inorganic chemist A. Albert Cotton (1930–2007) of textbook fame. However, whether they know anything about the person or not, for the chemists working in chiroptics or stereochemistry, T. M. Lowry’s words of 1935 still hold true [21]: Cotton’s discovery in absorbing optically-active media of the twin phenomena of circular dichroism and of anomalous rotatory dispersion, which are indissolubly associated with his name, is [. . .] amongst the “classics.”

ACKNOWLEDGMENTS It is a pleasure to acknowledge the kind cooperation of Ms. E. H. Lifschitz in Haifa, Israel, who has provided important information on her father, Israel Lifschitz, and has permitted the publication of his photograph. It is a privilege to acknowledge also the untiring help of Dr. Henry Joshua of New York City, who has overcome many difficulties in his endeavors to establish contacts with Ms. Lifschitz, whose whereabouts had been unknown. The author is also grateful to Professor Jerome Gurst of Pensacola, Florida, for language counseling and for his helpful critique of the manuscript.

REFERENCES 1. D. F. Arago, M´em. Cl. Sci. Math´em. Phys. Inst. France 1811, 12 , I, 1–16, 113–134 (published 1812); [J. B.] Biot, Phys. Inst. France 1812, 13 , I, 1–372 (May 1813). 2. J. B. Biot, Bull. Sci. Soc. Philomath. 1815, [3] 2 , 190–192. 3. J. B. Biot, M´em. Acad. Roy. Sci. Inst. France 1817, [2] 2 , 41–136 (September 1818). 4. A. Arndtsen, Ann. Chim. Phys. 1858, [3] 54 , 403–421; Pogg. Ann. 1858, [2] 105 , 312–317.

31

32

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

5. H. Landolt, Liebigs Ann. Chem. 1877, 189 , 241–337 (p. 274). 6. J. B. Biot, Ann. Chim. Phys. 1860, [3] 59 , 206–326; appendix 326–345 (page 266/7, Section III, 8). In this appendix, by the way, Biot refutes claims that the Baltic-German physicist Thomas Johann Seebeck (1710–1831) might have been the first to observe optical activity in solutions. 7. H. Landolt, Das optische Drehungsverm¨ogen organischer Substanzen und die praktischen Anwendungen desselben, Vieweg und Sohn, Braunschweig, 1879; English translation: Handbook of the Polariscope, D. C. Robb, V. H. Veley, translators, Cambridge, 1882 (reprint: BiblioLife, Charleston, SC, 2009). 2nd ed. Vieweg und Sohn, Braunschweig, 1898; English translation: The Optical Rotating Power of Organic Substances, J. H. Long, translator, Chemical Publishing Co., Easton, PA, 1902. 8. H. Fehling, Arch. Physiol. Heilkunde 1848, 7 , 64–73; Ann. Chem. Pharm. 1849, 72 , 106–113. 9. F. (-)P. Leroux, C. R. H. Acad. Sci . 1862, 55 , 126–128; Pogg. Ann. 1862, [2] 117 , 659–660. 10. C. Christiansen, Pogg. Ann. 1870, [2] 141 , 479–480; Pogg. Ann. 1871, [2] 143 , 250–259. A. Kundt, Pogg. Ann. 1870, [2] 142 , 163–171; Pogg. Ann. 1871, [2] 143 , 149–152; Pogg. Ann. 1871, [2] 143 , 259–269. J.-L. Soret, Arch. Sci. Phys. Natur. 1871, [2] 40 , 280–283. 11. A. Cotton, Absorption in´egale des rayons circulaires droit et gauche dans certain corps actifs, C. R. H. Acad. Sci . 1895, 120 , 989–991. 12. A. Cotton, Dispersion rotatoire anomale des corps absorbants, C. R. H. Acad. Sci . 1895, 120 , 1044–1046. 13. A. Cotton, Recherches sur l’absorption et la dispersion de la lumi`ere par les milieux dou´es du pouvoir rotatoire, Th`ese de Doctorat, Paris, 1896; Ann. Chim. Phys. 1896, [7] 8 , 347–432; summary: J. Phys. Th´eor. Appl . 1896, [3] 5 , 237–244. 14. [J. B.] Biot, Bull. Sci. Soc. Philomath. 1815, [3] 2 , 26–27; W. Haidinger, Pogg. Ann. 1845, [2] 65 , 1–30. 15. W. Haidinger, Pogg. Ann. 1847, [2] 70 , 531–544. 16. It would be interesting to know if Fehling’s solution was indeed Cotton’s very first sample. Apparently his papers, including his laboratory notebooks 1895–1920, are at the Niels Bohr Library, American Center for Physics, College Park, MD. 17. L. Natanson, Bull. Int. Acad. Sci. Cracovie, Cl. Sci. Math. Nat., Jan. 1909, 25–37. 18. A. Cotton, J. Phys. Th´eor. Appl . 1898, [3] 7 , 81–85. 19. F. Billet, Trait´e d’optique physique, Mallet-Bachelier, Paris, 1858/59 (2 vols.). 20. N. Wedeneewa, Ann. Phys. 1923, [4] 72 , 122–140. 21. T. M. Lowry, Optical Rotatory Power, Longmans, Green and Co., London, 1935; reprint: Dover Publications, New York, 1964. 22. S. Mitchell, The Cotton Effect and Related Phenomena, G. Bell & Sons, London, 1933. 23. W. Kuhn, A. Szabo, Z. Phys. Chem. 1931, B15 , 59–73. 24. W. O. [i.e., W. Ostwald], Z. Phys. Chem. 1896, 19 , 383, no. 87 and 88. 25. W. O. [i.e., W. Ostwald], Z. Phys. Chem. 1896, 21 , 158–163. 26. P. Drude, Lehrbuch der Optik , Hirzel, Leipzig, 1900; English translation: The Theory of Optics, C. R. Mann, R. A. Millikan [Nobel Prize 1923], translators, Longmans, Green & Co., London, 1902; reprint: Dover Publications, New York, 1959. 27. L. Natanson, Bull. Int. Acad. Sci. Cracovie, Cl. Sci. Math. Nat., O eliptycznej polaryzacyi s´wiatła, przepuszczonego przez ciało naturalnie , Oct. 1908, 764–783. 28. L. Natanson, J. Phys. Th´eor. Appl . 1909, 8 , 321–347. 29. A. Cotton, C. R. H. Acad. Sci . 1911, 153 , 245–247.

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

30. W. Kuhn, Proc. [2 nd] Conf. Rotatory Dispersion, Santa Fe, CA, 20–22 Jan., 1960; Tetrah. 1961, 13 , 1–12. 31. M. F. McDowell, Phys. Rev . 1905, 20 , 163–171. 32. L. B. Olmstead, Phys. Rev . 1912, 35 , 31–46. 33. H. Grossmann, A. Loeb, Z. Ver. Deutsch. Zuckerind., Allg. Teil 1908, 58 , 994–1009. 34. H. Volk, Ber. Dtsch. Chem. Ges. 1912, 45 , 3744–3748. 35. G. Bruhat, Th`ese de Doctorat, Paris, June 1914; Ann. Physique 1915, [9] 3 , 232–282, 417–489. 36. G. Bruhat, Ann. Physique 1920, [9] 13 , 25–48. 37. R. de Mallemann, P. Gabiano, C. R. H. Acad. Sci . 1927, 185 , 350–352. 38. J.-P. Mathieu, C. R. H. Acad. Sci . 1931, 193 , 1079–1081; Ann. Physique 1935, [11] 3 , 371–460. 39. J.-P. Mathieu, C. R. H. Acad. Sci . 1932, 194 , 1727–1729. 40. W. Pfleiderer, Z. Phys. 1926, 39 , 663–685. 41. R. D. Gillard, Progr. Inorg. Chem. 1966, 7 , 215–276. Oddly, this author maintains that “Aim´e Cotton (1872–1944) discovered the [Cotton] effect while Professor of Physics at the Sorbonne in Paris.” Both the dates (recte: 1869–1951) and the professorship are erroneous (Cotton became a Professor at the Sorbonne in 1910, but had discovered the effect in 1895). 42. A. Werner, Z. Anorg. Allg. Chem. 1893, 3 , 267–330; see also: Vierteljahresschrift der Z¨urch. Naturf. Ges. 1891, 36 , 129–169. 43. A. Werner, Ber. Dtsch. Chem. Ges. 1911, 44 , 1887–1898. 44. M. Del´epine, C. R. H. Acad. Sci . 1914, 159 , 239–241; Bull. Soc. Chim. Fr. 1917, [4] 21 , 157–172. 45. G. Bruhat, Bull. Soc. Chim. Fr. 1915, [4] 17 , 223–227. 46. A. Werner, Ber. Dtsch. Chem. Ges. 1912, 45 , 121–130; 3061–3070. 47. A. Werner, J. Poupardin, Ber. Dtsch. Chem. Ges. 1914, 47 , 1954–1960. 48. A. Werner, Ber. Dtsch. Chem. Ges. 1914, 47 , 3087–3094. 49. A. Werner, Helv. Chim. Acta 1918, 1 , 5–32. 50. Important reviews of stereochemistry and optical activity at Groningen: F. M. Jaeger, Proc. Akad. Wet. Amsterdam 1915, 17 , 1217–1236; Lectures on the Principle of Symmetry, Elsevier, Amsterdam, 1917 (2nd enlarged edition 1920); Bull. Soc. Chim. Fr. 1923, [4] 33 , 853–889; Spatial Arrangements of Atomic Systems and Optical Activity, McGraw-Hill, New York, 1930. 51. F. M. Jaeger, Rec. Trav. Chim. Pays-Bas 1919, 38 , 171–314. 52. J. Lifschitz, Z. Phys. Chem. 1923, 105 , 27–54. It may be pointed out that the author abbreviated his first name, Israel, in print usually by the letter “J”; this was not uncommon in those days. 53. J. Lifschitz, Z. Phys. Chem. 1925, 114 , 485–499. 54. F. M. Jaeger, H. B. Blumendal, Z. Anorg. Allg. Chem. 1928, 175 , 161–230. 55. W. Kuhn, K. Bein, Z. Anorg. Allg. Chem. 1933/34, 216 , 321–348 (ORD); Z. Phys. Chem. 1934, B24 , 335–369 (CD). 56. R. E. Ballard, A. J. McCaffery, S. F. Mason, Proc. Chem. Soc. London 1962, 331–332. 57. A. J. McCaffery, S. F. Mason, Proc. Chem. Soc. London 1962, 388–389. 58. H. Rupe, Zeltner, W. Lotz, M. Silberberg, Liebigs Ann. Chem. 1903, 327 , 157–200 (first paper of the series); H. Rupe, L. Silberstrom, Liebigs Ann. Chem. 1918, 414 , 99–111 (ninth paper). 59. T. M. Lowry, J. Chem. Soc. London 1899, 75 , 211–244.

33

34

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60. H. Rupe, A. Krethlow, K. Langbein, Liebigs Ann. Chem. 1921, 423 , 324–342 (thirteenth paper of the series [58]). 61. T. M. Lowry, ed., Trans. Faraday Soc. 1914, 10 , 44–138. 62. So transliterated by T. M. Lowry in the obituary note (J. Chem. Soc. London 1923, 123 , himself, who often published in Germany, usually spells his name 956–958). . A. there Tschugaeff, and sometimes Tch´ugaeff or Tschugajew, in France also Tchougaeff. Chemical Abstracts always list him as L[ev] A[leksandrovich] Chugaev. In the references to this overview, the original spelling as given in the papers has been chosen. 63. L. Tschugaeff, Ber. Dtsch. Chem. Ges. 1909, 42 , 2244–2247 (paper I of the series); L. Tschugaeff, A. Ogorodnikoff, Z. Phys. Chem. 1910, 74 , 503–512 (II); L. Tschugaeff, Z. Phys. Chem. 1911, 76 , 469–483 (III); L. Tschugaeff, A. Ogorodnikoff, Z. Phys. Chem. 1912, 79 , 471–480 (IV); L. Tschugaeff, A. Ogorodnikoff, Z. Phys. Chem. 1913, 85 , 481–510 (V); L. Tschugaeff, W. Pastanogoff, Z. Phys. Chem. 1913, 85 , 553–572 (VI). See also: L. Tschugaeff, A. Ogorodnikoff, Ann. Chim. Phys. 1911, [8] 22 , 137–144; L. Tschugaeff, J. Prakt. Chem. 1912, [2] 86 , 545–550; L. Tchougaeff, Bull. Soc. Chim. Fr. 1912, [4] 11 , 718–722; L. Tschugaeff, G. Glinin, Ber. Dtsch. Chem. Ges. 1912, 45 , 2759–2764; L. Tchugaeff, A. Kirpitcheff, Bull. Soc. Chim. Fr. 1913, [4] 13 , 796–803; L. Tschugaeff, , A. A. , , B. [L. Trans. Faraday Soc. 1914, 10 , 70–79; . A. 1915, A. Tschugaev, A. A. Glebko, G. V. Pigulevskii], J. Russ. Phys. Chem. Soc. 47 , 774–775. 64. G. Bruhat, C. R. H. Acad. Sci . 1911, 153 , 248–250. 65. T. M. Lowry, H. Hudson, Philos. Trans. 1933, A232 , 117–154. 66. T. M. Lowry, H. K. Gore, Proc. Roy. Soc. 1932, A135 , 13–22. 67. E. Deussen, J. Prakt. Chem. 1912, [2] 85 , 484–488. 68. L. Tschugaeff, J. Prakt. Chem. 1912, [2] 86 , 545–550. 69. S. Mitchell, J. Chem. Soc. London 1928, 3258–3260. 70. E. Darmois, Ann. Chim. Phys. 1911, [8] 22 , 247–281, 485–590; Th`ese de Doctorat, Paris, 1910. 71. W. Kuhn, H. K. Gore, Z. Phys. Chem. 1931, B12 , 389–397. 72. T. M. Lowry, H. S. French, J. Chem. Soc. London 1932, 2654–2658. 73. R. Servant, C. R. H. Acad. Sci . 1932, 194 , 368–369. 74. C. Djerassi, Optical Rotatory Dispersion. Applications to Organic Chemistry, McGraw-Hill, New York, 1960. ´ 75. G. Bruhat, Trait´e de Polarim´etrie [with a preface by A. Cotton], Editions de la Revue d’Optique th´eorique et instrumentale, Paris, 1930. 76. Review: W. Kuhn, Theorie und Grundgesetze der optischen Aktivit¨at [theory and fundamental principles of optical activity], in Stereochemie, K. Freudenberg, ed., Deuticke, Leipzig, 1933, pp. 317–434. 77. R. Descamps, Trans. Faraday Soc. 1930, 26 , 357–371. 78. T. M. Lowry, Proc. Roy. Soc. 1908, A81 , 472–474; Philos. Trans. 1912, A212 , 261–297; T. M. Lowry, W. H. C. Coode-Adams, Philos. Trans. 1927, A226 , 391–466. 79. A. Cotton, R. Descamps, C. R. H. Acad. Sci . 1926, 182 , 22–26; R. Descamps, Rev. Opt. Th´eor. Instr. 1926, 5 , 481–501; G. Bruhat, M. Pauthenier, C. R. H. Acad. Sci . 1926, 182 , 888–890 (with comments by A. Cotton, C. R. H. Acad. Sci . 1926, 182 , 890–891); G. Bruhat, M. Pauthenier, Rev. Opt. Th´eor. Instr. 1927, 6 , 163–184. 80. G. Bruhat, P. Chatelain, C. R. H. Acad. Sci . 1932, 195 , 462–465; J. Physique 1932, [7] 3 , 501–511; G. Bruhat, A. Guinier, C. R. H. Acad. Sci . 1933, 196 , 762–764; Rev. Opt. Th´eor. Instr. 1933, 12 , 396–416. 81. G. Bruhat, Bull. Soc. Chim. Fr. 1930, [4] 47 , 251–261.

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

82. W. Kuhn, Ber. Dtsch. Chem. Ges. 1929, 62 , 1727–1731; W. Kuhn, E. Braun, Z. Phys. Chem. 1930, B8 , 445–454. 83. J.-P. Mathieu, Ann. Physique 1935, [11] 3 , 371–460. 84. H. Rudolph, J. Opt. Soc. Am. 1955, 45 , 50–59; Rudolph Photoelectric Spectropolarimeter®, O. C. Rudolph & Sons, Caldwell, NJ. 85. M. Grosjean, M. Legrand, C. R. H. Acad. Sci . 1960, 251 , 2150–2152 (Centre de Recherche Roussel-Uclaf); Dichrograph®, Societ´e Jouan, Paris. 86. M. Born, Phys. Z . 1915, 16 , 251–258; Ann. Phys. 1918, [4] 55 , 177–240. 87. C. W. Oseen, Ann. Phys. 1915, [4] 48 , 1–56. 88. F. Gray, Phys. Rev . 1916, [2] 7 , 472–488. 89. L. Rosenfeld, Z. Phys. 1928/29, 52 , 161–174; Rosenfeld was at that time a postdoctoral associate with M. Born in G¨ottingen, Germany. 90. W. Kuhn, Z. Phys. Chem. 1929, B4 , 14–36. 91. W. Kuhn, Z. Phys. Chem. 1935/36, B31 , 23–57. 92. J.-P. Mathieu, Les Th´eories Mol´eculaires du Pouvoir Rotatoire Naturel , Gauthier-Villars, Paris, 1946. 93. L. Tschugaeff, Z. Phys. Chem. 1911, 76 , 469–483. 94. L. Tschugaeff, A. Ogorodnikoff, Ann. Chim. Phys. 1911, [8] 22 , 137–144. , , Publishing House of the Academy of Sciences of the SSSR, 95. . A. Moscow, 1955. 96. L. Velluz, M. Legrand, M. Grosjean, C. R. H. Acad. Sci . 1963, 256 , 1878–1881. do fizyki teoretycznej , Redakcye “Prac Matematyczno-Fizycznych”, War97. L. Natanson, saw, 1890. 98. L. Klecki, Prace Matematyczno-Fizyczne 1939, 46 , 1–18 (in French). 99. J. Salkind, Ber. Dtsch. Chem. Ges. 1922, 55 , 141A–142A. 100. W. J. Pope: Thomas Martin Lowry, Obituary Notices of Fellows of the Royal Society 1938, 2 , 287–293. 101. J. Lifschitz, E. Rosenbohm, Z. Wiss. Photogr., Photophys. Photochem. 1920, 19 , 198–214. 102. E. H. Lifschitz, Er was weinig begrip voor de joden [there was not much sympathy with the Jews], in Terug van weggeweest, J. van Gelder, ed., Stichting Geldersboek, Groningen, 1993, pp. 141–148 (Chapter 15); also: private communication to the author. 103. H. Kuhn, Chimia 1984, 38 , 191–211. 104. A. Cotton, Notice sur les Travaux Scientifiques, Presses Universitaires de France, Paris, 1923. 105. A. Cotton, H. Mouton, Ann. Chim. Phys. 1907, [8] 11 , 145–203, 289–339. 106. M. Javillier, C. R. H. Acad. Sci . 1951, 232 , 1521–1527; J. Cabannes [Cotton’s successor at the Sorbonne], Ann. Physique 1951, [12] 6 , 895–898; Louis [Prince] de Broglie [Nobel Prize 1929], Notice sur la vie et l’œvre de Aim´e Cotton, Institut de France, Acad´emie des Sciences, Paris, 1953 [30 pages]. 107. A. Kastler, C. R. H. Acad. Sci . 1969, 269 , 70–74. 108. E. Crawford, The Nobel Population 1901–1950: A Census of the Nominators and Nominees for the Prizes in Physics and Chemistry, Universal Academy Press, Tokyo, 2002. (Nomination 1915, page 62 (C. Fabry, nominator). Similarly: 1916, p.64 (M. Brillouin); 1920, p.80 (J. Bordet); 1922, p.84 (J. Bordet); 1925, p.94 (J. Bordet); 1927, p.102 (J. Bordet, C.E. Guillaume, C.E. Guye, A.Schidlof); 1928, p.106 (C.E. Guillaume); 1929 , p.110 (C.E. Guillaume); 1930, p.116 (C.E. Guillaume, H. Villat); 1931, p.120 (C.E. Guillaume, R. de Mallemann); 1932, p.124 (C.E. Guillaume, G. Reboul); 1933, p.128 (C.E. Guillaume); 1934, p.132 (H. Buisson, C.E. Guillaume, C.E. Guye, J. Perrin, V. Posejpal, P. S`eve); 1944, p.170 (H. Beghin); 1949, p.186 (M. Pauthenier)).

35

PART II ORGANIC STEREOCHEMISTRY

2 SOME INHERENTLY CHIRAL CHROMOPHORES—EMPIRICAL RULES AND QUANTUM CHEMICAL CALCULATIONS Marcin Kwit, Pawel Skowronek, Jacek Gawronski, Jadwiga Frelek, Magdalena Woznica, and Aleksandra Butkiewicz

Molecules that do not contain center of chirality can still be chiral if their structure does not contain a plane of symmetry. Typical examples are hexahelicene and biphenyl (Figure 2.1) in which the chromophoric system (in these cases π electron system) is helical due to steric reasons preventing enantiomerization of the molecule at ambient temperature. Examples discussed in this chapter include (a) conjugated chromophoric systems of dienes, (b) enones, (c) helical chalcogenides, and (d) nonplanar amide chromophores. Many molecules of natural origin contain chromophores of either of this type (e.g. steroids, terpenes, amino acids, antibiotics, and metabolites); hence interpretation of their ECD may be useful for their absolute structure elucidation.

2.1. DIENES AND TRANS-ENONES Dienes and enones form a class of chromophores that can be helical due to nonplanarity of the π -bond system. Their chiroptical properties have been a subject of numerous studies over the past 50 years [1], aimed at clarifying the role of twist of the π -bond system and chirality of the molecule on the ECD spectra and optical rotation. There are other types of helical chromophoric systems; among these are dichalogenides (disulfides, diselenides, ditellurides), which are distinctly different due to their non-π -electron chromophore, and these will be discussed in detail in the following section (Figure 2.2). The conformations of these molecules may be conveniently described by considering their helicities. 1,3-Dienes can exist in two planar conformations defined as s-trans and s-cis and an infinite number of nonplanar, skewed forms, traditionally called cisoid or Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

39

40

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 2.1. Hexahelicene and biphenyl—examples (M)-Hexahelicene

of inherently chiral (helical) chromophores.

(M)-Biphenyl

(a)

(b) C C

C

C

C

C

O

w=C=C−C=C

C

Ch Ch

C

C

C

C

w=O=C−C=C

R

C

w = C − Ch − Ch − C (Ch = S, Se, Te)

t=C−C=C−R

Figure 2.2. Simple helical chromophores and definitions of torsion angles ω (a) and definition of angle τ that describes distortion of the C=C bond from planarity (b).

transoid , depending on the conformation of the nearer planar form. The conformation can otherwise be described by torsion angle ω, which can be either positive (P ) or negative (M ) (see Scheme 2.1). The same is true for enones and dichalcogenides; however, in the last case both planar s-trans and s-cis conformations represent the transition states, not any stable structures. Additional classification of chromophore structures may be accomplished on the basis of molecular symmetry [1b]. Both butadiene and symmetrically disubstituted dichalcogenides in planar forms belong to either C2h or C2v symmetry point group, respectively, for s-trans and s-cis conformers. Rotation around the C–C or Ch–Ch bonds reduces the

(a) X

X S-cis (sp)

X Cisoid

Planar

P-helical

Cisoid

X Transoid

X S-trans (ap)

X Transoid

(b)

M-helical

P-helical

Planar

M-helical

X = CH2 (1,3-dienes) X = O (enones)

Scheme 2.1. (a) Possible planar and nonplanar conformations of 1,3-dienes and enones and (b) definition of M- and P-helicities in the cases of cisoid and transoid 1,3-dienes.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

symmetry to C2 in skewed, chiral conformations. Unlike butadiene and dichalcogenides, the simplest enone (acrolein) exhibits only trivial C1 symmetry when skewed and Cs symmetry for both planar forms. The electronic transitions of 1,3-dienes and enones are reflections of their structures. The degree of electron delocalization determines intensity and energy of the low-energy π –π * electronic transition in the diene chromophore. For the most populated s-trans conformation of butadiene, the most intense band is found at 210 nm with εmax > 20,000 [2], whereas for s-cis planar conformation, both the intensity and energy of the π –π * electronic transition is significantly lower. In the case of cyclopentadiene the long-wavelength absorption band appears at 240 nm (εmax = 3500) [2]. The extinction coefficients and energies of the low-energy π –π * electronic transitions for skewed dienes are higher for transoid compared to cisoid structures and reach a minimum (λmax = 190 nm) for perpendicular diene confomation (ω = ±90◦ ) [2, 3]. In the case of 1,3-cyclohexadiene the observed UV maximum appears at 256 nm (εmax = 9000) [4]. The lowest-energy π –π* electronic transition has been used as a diagnostic band for solving stereochemical problems with the use electronic circular dichroism (ECD) spectroscopy. Contrary to dienes, α, β-unsaturated ketones typically exhibit two absorption bands: one very weak near 330 nm (R-band) and the second at around 230 nm (K-band). The lowest-energy band originates from a forbidden transition from nonbonding 2py orbital into antybonding π * orbital; thus this electronic transition is defined as n –π * [5, 6]. In the case of 2-cyclohexenone the lowest-energy n –π * transition is found at 320 nm (ε = 36) [7]. For chiral enones the dissymmetry factor (g) for the lowest-energy n –π * electronic transition is in the range 10−1 to 10−2 . The second UV absorption band in enone chromophore has the character of an intramolecular charge transfer transition from the vinyl to the carbonyl group and exhibits much higher intensity compared to the lowest-energy n –π* electronic transition. Oscillator strengths for transoid enones are usually higher than for cisoid . The dissymmetry factor for π –π * electronic transition is in the range 10−3 to 10−4 . In the case of 2-cyclohexenone, this band appears at 225 nm (εmax = 13800) [7]. Although only two bands are observed in the UV spectra of enones, ECD spectra of α, β-unsaturated ketones exhibit three (sometimes four) Cotton effects between 360 and 185 nm. The weak UV absorption at about 300 nm is responsible for the long-wavelength Cotton effect of moderate intensity. This band is followed by a π –π * transition Cotton effect corresponding to the UV band placed between 220 and 250 nm. Another Cotton effect that does not have the corresponding UV maximum appears at around 200 nm and usually exhibits a high rotatory strength, sometimes even higher than that for the π –π * electronic transition. The origin of this Cotton effect is not clear. Earlier calculations by Liljefors and Allinger suggested that this is another π –π * transition of low intensity in nearly planar enones [8]. Snatzke suggested that this is the second forbidden n –π * electronic transition [1i, 9]. The direction of polarization of this transition has been determined by linear dichroism measurements [10]. The fourth electronic transition is observed between 195 and 185 nm in ECD spectra of some enones having an axial substituent in α  or β  positions. This transition was considered as n –σ * excitation [1d–1f]. During the last few decades, various empirical rules have been proposed to correlate the signs of the Cotton effects of 1,3-dienes and α, β-unsaturated ketones with their stereochemistry (Table 2.1). Whereas some of the correlation rules are of historical value, two basic stereochemical concepts underlying the development of such correlations are briefly considered below.

41

42

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 2.1. Empirical Rules Correlating Cotton Effect for a Given Electronic Transition with the Chirality of Compounds Containing 1,3-Diene and Enone Chromophores Electronic Transition

Rule

π –π * (dienes) π –π * (dienes) π –π * (dienes) π –π * (dienes) π –π * (planar s-trans dienes) π –π * (enones) n –π * (enones) n –π * and π –π * (enones) n –π * (planar s-trans enones) n –π * and π –π * (enones) 210-nm transition (cyclic enones)

References

Diene Helicity Rule (Moscowitz et al., 1961) Allylic Axial Chirality Rule (Burgstahler and Barkhurst, 1970; Burgstahler et al., 1976) Quadrant Rule (Moriarty et al., 1979) Sector Rule (Weigang, 1979) Planar Diene Rule (Duraisamy and Walborsky, 1983)

11 12, 13

Enone Helicity Rule (Djerasi et al., 1962, Whalley, 1962) Enone Helicity Rule (Snatzke, 1965) Orbital Enone Helicity Rule (Kirk, 1986)

17, 18 19 1e

Sector Rule (Snatzke, 1979; Snatzke, 1965)

1h, 19a

Allylic Axial Chirality Rule (enones) (Burgstahler and Barkhurst, 1970) Carbon–Carbon Bond Chain in Cyclic Enones Rule (Gawronski, 1982)

14 15 16

12 1d

For correlating the chiroptical phenomena with the diene or enone structure, two fundamental effects had to be taken into account: (a) the contribution from the helicity of the chromophore and (b) the effect of extrachromophoric perturbation. Whereas the first of these effects correlates conformation of the chromophore with its spectroscopic properties, the second is related to the configuration at the stereogenic center(s). Both effects may act in the same or in opposite direction, so determining the dominant contribution is an important yet not always easy task. In the case of rigid 1,7,7-trimethyl2,3-dimethylene-bicyclo[2.2.1]heptane (1), chirality originates from the presence of the methyl group connected to C1 carbon atom [20]. As long as the chromophore is planar, the effect of extrachromophoric perturbation dominates. On the other hand, in the case of highly flexible (3S,8S,E,E )-dimethyl-deca-4,6-diene (2), both effects contribute to the π –π * transition Cotton effect [21], whereas in the case of the rigid structure of ergosterol (3) a negative twist of diene moiety is considered as the origin of the long-wavelength Cotton effect [22]. C9H17

H HO 1

2

De = −0.6 (250 nm)

De = +3.0 (230 nm)

3 w = −11° De = −18.0 (269 nm)

According to the Diene Helicity Rule (DHR) [11a], the sign of the long-wavelength Cotton effect reflects directly the sense of helicity of the chromophore; that is, for positive twist of the 1,3-diene system a positive long-wavelength π –π * transition Cotton effect is expected (Figure 2.3a).

43

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

(a)

(b)

Figure 2.3. Empirical correlating P +C.e. w > 0°

M −C.e. w < 0°

+C.e. for right-handed angle Caxial −Callyl −C = C

−C.e. for left-handed angle Caxial − Callyl −C = C

rules for 1,3-dienes: (a) Diene Helicity Rule and (b) Allylic Axial Chirality Rule.

The validity of DHR has been challenged by calculations at different levels of sophistication [3, 23–31]. Most of them were performed for simple models like 1,3butadiene and generally confirmed the validity of DHR, but it should be noted that some of the calculations gave opposite results [28, 31]. To overcome the inadequacy of the Diene Helicity Rule for certain types of diene structures, the concept of allylic axial substituent contributions to optical activity of conjugated dienes was proposed by Burgstahler and co-workers (Allylic Axial Chirality Rule, hereafter referred to as AACR) [12, 13]. According to this concept, the sign of the long-wavelength Cotton effect of conjugated dienes is primarily due to the contribution of allylic axial substituents, such as alkyl groups. The sign of the contribution is determined by the helicity (+ or −) of the Caxial –Callyl –C=C bond system (Figure 2.3b). Thus, antipodal CD curves of 15-methylene-5α-cholest-8(14)-en-3β-ol acetate (4) and 3-isopropylidene-A-norcholest-5-ene (5) were explained as due to contribution from the allylic axial substituents; these dienes exhibit Cotton effects of opposite signs to those required by the DHR. + AcO

H + + – H H H w = +20° 4 De = +4 (250 nm)

H – H

–

w = +20°

– H

5 De = −5 (250 nm)

A corollary to this rule is a low contribution of the diene chromophore to the rotational strength of the π –π * transition. Certain substituents attached to one of the diene carbon atoms (e.g., the CN group) can cause sign reversal of the long-wavelength Cotton effect, in the absence of any obvious structural change [32], and this is another example of inconsistency with the Diene Helicity Rule. An experimental and theoretical study of α-phellandrene (6) and other 5-alkyl-1,3cyclohexadienes (7, 8) by Lightner et al. [30] for the first time provided a dissection of the contributions of various structural elements to the cyclohexadiene 260 nm Cotton effect. In the case of (5R)-5-methyl-1,3-diene (7), variable-temperature ECD spectra were invariant, due to a low energy difference between equatorial and axial conformers, estimated as 0.05 kcal mol−1 (Figure 2.4a,b), and exhibited a positive long-wavelength Cotton effect ( ε = +5 at 260 nm). A dramatic change of ECD spectra with temperature was observed when a tert-butyl group was attached to the 1,3-cyclohexadiene skeleton. At slightly elevated temperature, the sign of the long-wavelength Cotton effect was positive ( ε = +2.5 at +31◦ C) and stepwise decreased upon lowering the temperature. At −180◦ C the value of the long-wavelength Cotton effect was ε = −3.0, which corresponds to energy difference Gax-eq = 0.4 kcal mol−1 .

44

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The rotatory strengths estimated from ECD spectra for 8 are positive for axial P helical conformer (R = +55.6 × 10−40 cgs units) and negative for equatorial M -helical conformer (R = −15.0 × 10−40 cgs units). The sign of the contribution due to the s-cisdiene moiety alone is calculated opposite to that predicted by the Diene Helicity Rule for 5-alkyl-1,3-cyclohexadienes, and the sign and magnitude of the Cotton effect are apparently dominated by the contributions of the axial allylic bonds (groups) (Figure 2.4c). Recently, a more advanced ab initio calculation by Hansen and Bak [33] in the random phase approximation (RPA) using Aug-cc-pVTZ atomic basis set provided important confirmation of the earlier findings on the role of allylic substituents. The effects of the allylic methyl groups were found to follow a quadrant rule being almost additive, and the contributions from axial substituents were calculated significantly larger than those from the equatorial groups. Analysis of chiroptical properties of compounds having enone chromophore is usually much more laborious. The first difficulty is due to small energy differences between conformers, and the second is the origin of the electronic transitions. While the n –π * electronic transition has origin similar to that of saturated ketones, higher energy transitions in enones are not pure and involve various types of orbitals. Estimated inversion barrier for 2-cyclohexenone is lower than the inversion barrier experimentally determined for cyclohexene [34]; and full inversion cycle involves a number of structures, each characterized by out-of-plane position of at least one saturated carbon atom. The diversity of possible conformation does not vanish even if a 2-cyclohexenone ring is a part of polycyclic systems, as in steroids and terpenoids. For example, for testosterone (9) molecule in the crystal lattice the dihedral angle ω that characterizes chromophore helicity takes different values, both positive and negative,

(a)

(c)

(b)

Figure 2.4.

Conformational drawings of (R)-α-phellandrene (6) (a) and (5R)-5-alkyl-1,3-

cyclohexadienes 7 and 8 (b) showing the axial or equatorial position of the alkyl group in relation to diene helicity and estimated group contributions to the rotatory strength (R (×10−40 cgs units) for the lowest-energy π −π ∗ electronic transition of (P)-1,3-cyclohexadiene (c).

45

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

depending on the space group and on the presence or not of other molecules forming hydrogen bonds or complexes [35–38].

Almost 50 years ago, Djerassi et al. [17], Whalley [18], and Snatzke [19] proposed rules for the first two Cotton effects: positive enone helicity was correlated with a negative n –π * and a positive π –π * Cotton effect. This rule was thereafter applied by Kirk to s-cis and s-trans enones unperturbed by polar substituents, and (P ) π orbital helicity at C1 and C2 carbon atoms was correlated with a positive n –π * and a negative π –π * Cotton effect (Figure 2.5a,b) [1e]. These rules worked well for cyclohexenones, but for cyclopentenones inverse rules were proposed [1d, 1f, 19b]. Snatzke proposed a modification of the octant rule that correlates the sign of the n –π * Cotton effects with the stereostructure of planar enones [1h,1i,19c]. The n –π * Cotton effects of steroidal enones in oriented (anisotropic) systems were later studied by Kuball et al. [39]. Their studies demonstrated that sector or helicity rules can be applied, provided that vibronic progressions originating from various conformers are taken into account. On the other hand, Burgstahler and Barkhurst [12] has shown for the first time the importance of allylic axial substituents that led to a breakdown of the enone helicity rule for the π –π * transition of some s-cis steroid enones. The electronic transition observed between 220 and 200 nm in ECD spectra of some enones [9, 40] is characterized by a low oscillator strength, making it difficult to detect in (a)

(b)

O

O

s-cis (P)

O

s-trans (P)

s-cis (P)

−C.e. (n−p*), +C.e. (p−p*) (c)

O

O

s-trans (M )

+C.e. (n−p*), −C.e. (p−p*) (d)

O

R b′ O

O

+C.e. (~220 nm)

−C.e. (~220 nm)

a′ R

b′

b

a′

a O

Figure 2.5. (a) The first Enone Helicity Rule. (b) Kirk’s Enone Orbital Helicity Rule. (c) correlation rule for the third Cotton effect. (d) Correlation rule for the sign of the fourth Cotton effect in 2-cyclohexenones.

46

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the UV spectra, and a high rotatory strength, resulting in some cases in overlap with the Cotton effect originating from the π –π * electronic transition. Attempts to correlate the sign of the third Cotton effect with structural parameters were met with many difficulties. Burgstahler correlated the sign of the third Cotton effect with the chirality of α  carbon atom [12]. In the case of 4 -3-ketosteroids, a positive value of the third Cotton effect is in agreement with the chirality due to an axial 2β-substituent. This relationship can only be treated as tentative due to numerous exceptions. Gawronski suggested that the sign of the third Cotton effect is connected with the presence of saturated carbon–carbon bond chain, especially in polycyclic enones. It shows a clear relation to the absolute configuration at the allylic position (Figure 2.5c) [1d, 1f]. The fourth electronic transition, thought to involve n and σ ∗ orbitals, has been found in enones having α  or β  substituents in an axial position. The sign of this Cotton effect is dominated by the orientation of substituent, with assumption of a sofa(5) conformation of 2-cyclohexenone skeleton (Figure 2.5d).[1d–1f]. In the light of the above-mentioned facts, proper determination of the nature of electronic transition(s) involved is mandatory for rational stereochemical analysis of compounds containing the enone chromophore. This is of special interest in the case of molecules containing polar groups, which may influence strongly both the sign and the magnitude of the Cotton effects [41]. On the basis of DFT calculation at the B2LYP/6-311 + +(2d,2p) level and with the use of NBO method [42], it was possible to determine the origin of the first three electronic transition for a group of model compounds. For s-trans-acrolein, 2-cyclohexenone, (4S )-4-hydroxy-2-cyclohexenone (10), and (5R)-5-hydroxy-2-cyclohexenone (11), characterized by planar conformation of the chromophore (ω = 180◦ , τ = 0◦ ) and sofa(5) conformation of the cyclohexenone skeleton, the origin of the first two electronic transitions did not raise doubts (Figure 2.6). The long-wavelength electronic transition with a small oscillator strength involves HOMO(−1)–LUMO orbitals (nC=O –πC=C * type), whereas the second electronic transition with the highest oscillator strength is of CT character, involving HOMO and LUMO orbitals (πC=C –πC=O * type). The third electronic transitions, responsible for the third strong Cotton effect observed in the ECD spectra of enones, is characterized by a small oscillator strength. For acrolein the third electronic transition involves both πC=O –πC=O * and nC=O –σ * transitions, while in the case of 2-cyclohexenone it is a mixture of electronic transitions of the πC=C –πC=C * type (HOMO–LUMO(+1), the main contribution) and πC=C –σ * [HOMO–LUMO(+2)]. This supports the earlier suggestion [1d] that the configuration of saturated C–C bond chain in cyclic enones may be, in the absence of polar substituents, the controlling factor of the sign of the short-wavelength Cotton effect. The presence of a hydroxy group in enones 10 and 11 causes a change of the character of the third electronic transition, compared to 2-cyclohexenone. This electronic transition appearing at ∼190 nm involves the lone pairs of the hydroxy group [HOMO(−2)] and LUMO orbitals and thus may be referred to as nOH –πC=O * type. This may suggest that the third, short-wavelength Cotton effect in 2-cyclohexenones having polar substituents at C4 and/or C5 position depends on the helicity of the (H)O–C · · · C=O bond system [41]. In the analysis of the origin of optical activity of cyclic 1,3-dienes and enones the effect of nonplanarity of the C=C bond(s) is usually neglected. Recently, Diedrich and Grimme [43] performed an advanced quantum chemical calculation of the rotatory strength for the electronic transitions of twisted (−10 deg) ethylene. The calculated rotatory strength for the π –π * transition was of the order 75–198 × 10−40 cgs units, depending on the method used. Since nonplanar ethylene generated a high rotatory

47

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

(a)

(b)

O 10

IIIb

OH 11

IIId

IIIc

LUMO(+1)

LUMO

(d)

O

O

O

LUMO(+2)

OH

(c)

I II IIIa

I II

I II IIIe

I II IIIe

HOMO

HOMO(−1)

HOMO(−2)

Figure 2.6. Molecular orbitals obtained with the use of the NBO method for representative planar enones: (a) Acrolein and (b) I nC=O –πC=O ∗, II πC=C –πC=O *, IIIa πC=O –πC=O *, IIIb nC=O –σC(O) – H *, IIIc πC=C –πC=C ∗, IIId πC=C –σC – C ∗ /σC – H *, IIIe nOH –πC=O *.

strength, neglecting such an effect in the case of π –π * electronic transitions of nonplanar enone and diene seems unjustified. Nonplanar conformations of simple 1,3-butadiene and acrolein may be chiral due to nonplanarity of the C=C bonds, defined as the sign and value of torsion angle τ (Figure 2.2). Thus, nonplanar molecular conformations may be due to nonzero values of either angle ω or angle(s) τ , or both. If both torsion angles ω and τ are considered, chromophore conformation may be defined as homohelical (if the signs of angles ω and τ are the same) or as heterohelical , if the signs of angles ω and τ are opposite [44]. In the case of s-trans acrolein and 2-cyclohexenone, deformation of the C=C bond from planarity results in the appearance of nonzero rotatory strength. Enone Helicity Rule is obeyed as long as the twist of the C=C bond and enone helicity are of opposite sense (heterohelical [41]). We will now discuss the effect of substitution of 1,3-cyclohexadiene and 2-cyclohexenone, which sometimes leads to contrasting effects. (S , S )-trans-1,2-Dihydroxy-3,5-cyclohexadiene (12) in the crystal phase forms an M -helical conformer (ω = −11.5◦ ), with the hydroxy groups occupying equatorial positions. Variable-temperature ECD spectra measured in methanol reveal that the rotatory strengths for both P- and M-helical conformers are positive, in accordance with the helicity of O–Callyl –C=C system, and the rotatory strength for diaxial P -conformer is one order of magnitude higher than that for the diequatorial one (Scheme 2.2) [45]. Significant solvent effect can be observed in the ECD spectra of 12. ECD spectrum in methanol solution at room temperature shows a Cotton effect ( ε = +11.2 at 258 nm),

48

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

w = −11.5° (X-ray)

OH

H

OH O

OH

H

12

O

methanol

H

cyclohexane

H

H

Δep−p* +6 (cyclohexane) +11 (methanol)

H M (more abundant) R p−p* > O

OH P R p−p* >> O

Scheme 2.2. Conformational equilibrium of dihydroxydiene 12.

due to increased population of P -helical conformer, whereas in nonpolar solvent the presence of intramolecular OH· · ·O bond favors the diequatorial M-conformer, with concomitant decrease of the long-wavelength Cotton effect ( ε = +6 at 259 nm). According to recent calculation performed for 12 at the DFT/B3LYP/6-311++G(d,p) level, diaxial P conformer is of 0.7 kcal mol−1 higher energy relative to the diequatorial one. Rotatory strength of the π –π * electronic transition, calculated at the mPW1PW91/6311++G(2d,p) level for P -helical 12 is almost four times higher than that for the lowest-energy diequatorial M conformer [46], consistent with the experimental ECD data. More complex systems are represented by arene metabolites 13–17 [41, 46–48]. These compounds, being valuable chiral building blocks and ligands in organic synthesis [49], are characterized by the presence of a vicinal cis-diol system and a 1,3-diene or enone chromophore. All of them can exist in solution in an equilibrium of diastereoisomeric diene (enone) conformers of P or M helicity with one of the OH group in an axial and the other in an equatorial position.

The main problem with stereochemical analyses of compounds of this type is reliable determination of conformer population. In general, the number of available conformers is not limited to M and P diastereoisomeric structures. Intramolecular hydrogen bond patterns of the hydroxy groups further increase the number of distinct stereoisomeric structures (Scheme 2.3). Thus, either C1(or C4)–OH or C2(or C5)–OH can be a hydrogen bond donor, and the orientation of the O–H bond against the vicinal C–H bond

49

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

4

R

3

H

R

H O R2

O

3

R4

1

R

R1 2

H

R

4

R

2 1

R

R

4

H

3 R O

H

2

R H

H

R

1

R

4

3 R O

R2

R1

O

H

O

H

H P3

H P1

H M3

H

H a OH

3

R

O H

O

H H M1

R w

R

H

enones, R3

b OH

==O

4

R

R

3

H

O H R2

O

1

R H

H M2

4

R

3

H

R

O

2

R

R H

O H H M4

1

4

R

H 3 O R

R2 H

H H P2

O

1

R

R

4

3 R O

H

H

R2 O

1

R

HH P4

Scheme 2.3. Diastereomeric Pand M conformers of arene metabolites 13–17 and the definition of torsion angles α, β, and ω.

can be either syn or anti . This makes the number of available conformers up to eight (M 1–M 4, P 1–P 4), and the number can still be higher if one includes the rotamers due to the presence of nonspherical substituents in the ring. Contrary to 1,3-cyclohexadiene derivatives, in the case of cis-ketodiols 15–17 the conformational equilibrium is strongly affected by solvent polarity. Calculations performed at MP2/Aug-cc-pVTZ//B3LYP/6-311++G(2d,2p) level and with the use of a PCM model led to a conclusion that in polar surroundings a conformer of type P 4 dominates, in contrast to 1,3-cyclohexadiene derivatives, where conformer P 4 does not participate in conformational equilibrium even to a small extent [46–48]. It should be noted that X-ray diffraction analysis may provide quite different results, since in the crystal lattice the most important structure-determining factor is the possibility of formation of intermolecular hydrogen bonds. Thus, in the case of fluorinated ent-14e the molecular structure found in the crystal corresponds to conformer M 4. In contrast to M 1, M 2, and P 1 are calculated at the B3LYP/6-311++G(d,p) level as the lowest-energy

Figure 2.7. Potential energy surface calculated at the B3LYP/6-311++G(d,p) level for P- and M-helical conformers of fluorinated derivative 14e (a) and X-ray diffraction determined structure of ent-14e (b).

50

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

13b 13d

2 OH OH

Dioxygenase TDO-O2

OH

Dioxygenase TDO-O2

OH

0 Δe −2 13c

−4 minor

major

−6

13e 220

13a 13b 13c 13d 13e

13a 240

260

280

300

320 nm

Figure 2.8. (a) Empirical rule correlating the stereochemical course of arene cis-dihydroxylation with the relative size of substituents. (b) ECD spectra of cis-diols 13a-13e measured in cyclohexane solution.

conformers (Figure 2.7) for isolated molecule 14e [47]. The calculated preferred conformations of dihydrodiols substituted at the positions C3 or C6 by trifluoromethyl group are strongly affected by the presence of sequence of hydrogen bonds Oeq H · · · Oax H · · · FCF2 , which may shift the conformer equilibrium into such a conformer. The CD spectra measured in nonpolar solvent for a number of 3-substituted cisdihydrodiols 13a–13e differ markedly in the sign of the diagnostic long-wavelength spectral region (Figure 2.8b). This suggests that either the absolute configuration assigned according to Figure 2.8a is wrong, or the absolute configuration at C1 carbon atom is the same for the whole series and the cis-dihydrodiols 13a–13e differ in chiroptical properties [46]. Extensive computational study on the structure and chiroptical properties of arene metabolites led to a conclusion that the presence of substituents X and/or Y, as well as the hydroxy groups, is the decisive factor in shifting the P ↔ M conformer equilibrium in one or another direction, and this determines the sign and magnitude of the longwavelength diene Cotton effect. The substituent effect on the ECD spectra of cis-diols 13–14 may be ordered as follows: CN > Br > CH3 > CF3 > F = H. This study indicates that no simple empirical model, including the DHR and the AACR, can fully account for the experimental CD data of all cis-dihydrodiols. In relation to the Diene Helicity Rule, rotatory strength contribution from helical cyclohexadiene chromophore is in any case weak and, in certain cases (Br, CN substituents at C3), does not correlate with the sign of diene torsion angle. The failure of the AACR results from mutually canceling contributions due to allylic hydroxy groups, both axial and equatorial, in conformers of P and M helicity. Deceptively simple cis-ketodiols 15–17 are in fact complex systems, due to their conformational equilibria. Circular dichroism spectra measured in acetonitrile solutions and calculated at the PCM/B2LYP/Aug-cc-pVTZ level are quite similar, regardless the enone structure and in the case of cis-ketodiols 15 exhibit a positive/negative/positive sequence of Cotton effects. However, in the case of 16 and 17 the sign of the longwavelength nC=O –πC=O * Cotton effect appears affected by the substituent in C6 position, but the patterns of the second πC=C –πC=O * and the third nOH –πC=O * Cotton effects are similar to that measured and calculated for 15. Surprisingly, detailed inspection of the ECD spectra calculated for individual conformers of 15a shows the same pattern, but not the magnitudes, of Cotton effects [41]. Rotatory strengths calculated for M -helical conformers remain in agreement with DHR and also with the alternative AACR, which

51

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

assigns a dominant contribution to the axial hydroxy group at the allylic position. On the other hand, ECD spectra calculated for the P conformers, which dominate in the conformational equilibrium, are not in agreement with the empirical rules and show much higher magnitudes than those calculated for their M -helical counterparts. In the case of P -helical conformers a dominant contribution is assigned, rather unexpectedly, to the equatorial hydroxy group [41]. With the use of modern computational approach, Kwit et al. estimated the effect of structural factors on the rotatory strengths of electronic transitions by replacing systematically the hydrogen atoms at C3–C6 in (P )-2-cyclohexenone with either a polar hydroxy group or a nonpolar methyl group. As expected, the dominant role of substituent at C4 over substituents in either C5 or C6 positions is clearly visible, regardless their electronic nature and orientation. For the πC=C –πC=O * transition of monosubstituted (P )2-cyclohexenones the substituent contributions, including signs, are as follows: 4eq-OH (−), 4ax-OH (+), 4ax-Me (+) > 4eq-Me (−), 5ax-OH (−), 6eq-OH (+) > 5ax-Me (+), 6ax-Me (−) > 5eq-OH (+), 5eq-Me (−), 6eq-Me (+), 6ax-OH (+). Another striking result revealed for the first time is that the sign of the principal πC=C –πC=O * transition Cotton effect is less dependent on the enone nonplanarity (angle ω) and more dependent on nonplanarity of the C=C bond [41, 46b]. The models of ECD contributions for arene metabolites of mono-substituted 1,3cyclohexadiene and cis-ketodiol type are proposed shown in Figure 2.9. These models redefine the importance of structural factors previously considered as responsible for chiroptical properties of dienes and enones. Nonplanarity of the C=C bonds, usually neglected, is of equal or even higher importance compared to the distortion of the whole conjugated chromophore from planarity.

2.2. DICHALCOGENIDES: MOLECULES WITH INHERENTLY CHIRAL CHROMOPHORES Dichalcogenides are unique examples of molecules with an inherently chiral chromophore. The dichalcogenide Ch–Ch moiety (Ch = S, Se, or Te) exists in the form of two helical enantiomeric P /M conformations. Both enantiomeric conformers give opposite Cotton effects that cancel in the experiments. When the dichalcogen moiety is placed in a chiral surrounding, the conformers are no longer of equal energy and this (a)

(b)

Figure 2.9. Estimated bond contributions to the rotatory strength R (×10−40 cgs units) of the lowest-energy π − π * electronic transition of arene metabolites of (a) the type (P)-1,3cyclohexadiene derivative (a) and (b) the type (P)-2-cyclohexenone derivative.

52

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Scheme 2.4. Full interconversion scheme for dichalcogenide R–Ch–Ch–R conformers.

gives rise to the Cotton effects in the CD spectra. The conformer equilibrium involves transition states (TS) of either s-cis or s-trans structure in which the Ch–Ch moiety is planar (Scheme 2.4). C–S–S–C torsion angle (ω) for dimethyl disulfide in the gas phase was reported in the review as 90.44◦ (±0.45◦ ) [50]. Values 85◦ , 82◦ , and 87◦ were found for dimethyl diselenide based on the microwave [51] and vibrational spectra [52] as well as on the electron diffraction data [53]. The data from low temperature X-ray diffraction of dimethyl dichalcogenides are shown in Table 2.2. For benzene solution the values of torsion angle ω calculated from the dipole moment measurements (68.5 ± 2.0◦ for Me2 S2 , 65.4 ± 4.0◦ for Me2 Se2 , 35.9 ± 10.0◦ for Me2 Te2 ) [55] are much smaller than those found in the gas phase. This difference was explained by interactions between the solvent and the solute molecules. Data found for the diphenyl dichalcogenides in the crystal state and for benzene solution are more coherent— that is, 96.2◦ vs. 81.3◦ for Ph2 S2 , 97.1◦ vs. 62.6◦ for Ph2 Se2 , 88.5◦ vs. 84.0◦ for Ph2 Te2 [55]. Skewed structure of dichalcogenides is explained by the gauche effect, which was introduced for disulfides [50] but which can be extended to diselenides and ditellurides. In the dichalcogenide molecule the nonbonding electron pairs that reside on the p orbitals perpendicular to the Ch–Ch bond are partially overlapping. The destabilization due to the lone-pair–lone-pair repulsion in orthogonal position is reduced in nonplanar cisoid and transoid conformations of the dichalcogenide molecule. Hyperconjugation by which the p lone pairs of the chalcogen atoms are overlapping with the σ * molecular orbitals of the R–Ch bonds located in the same plane provides additional stabilizing effect [56]. Rotation barrier for the interconversion between enantiomeric conformers of dichalcogenides is low enough to allow free rotation of the Ch–Ch bond in acyclic molecules at room temperature (Scheme 2.4). Due to the repulsion between substituents on the chalcogen atoms, s-trans transition state is of lower energy. For S–S bond the rotation barrier is estimated in the range 6.8–13.2 kcal mol−1 [57–61]. Calculations (including MP2 method) give the value ca. 5 kcal mol−1 through the s-trans transition TAB L E 2.2. X-ray Data for Dimethyl Dichalcogenides Me2 Ch2

˚ Ch–Ch [A]

˚ C–Ch (A)

ω (degrees)

Me2 S2 Me2 Se2 Me2 Te2

2.03 2.31 2.71

1.80 1.94 2.15

86 85 90

Source: Data taken from reference 54.

53

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

state and 9 kcal mol−1 through the s-cis transition state. [50, 62, 63]. Calculated rotation barrier for H2 Se2 is lower (4.47 and 6.21 kcal mol−1 , respectively). Substitution with the phenyl group does not change significantly the magnitude of the rotation barrier which was calculated for diphenyl disulfide, diselenide, and ditelluride as 8.6, 8.2, and 5.3 kcal mol−1 through a s-cis transition state and 5.4, 5.2, and 3.7 kcal mol−1 through a s-trans transition state [64]. It is seen that the rotation barrier is lowered with the increase of the length of the Ch–Ch bond (see Table 2.2). A model for interpretation of the CD spectra of disulfide chromophore was proposed by Bergson [65, 66] and later verified by Linderberg and Michl [67] and by Woody [68]. Woody has shown the dependence of the rotation strength on the torsion angle of disulfide chromophore. In this model the sign of the long-wavelength Cotton effect obeys a quadrant rule (Figure 2.10). Bergson’s model due to its simplicity has been very successful. Recent study of the disulfide transitions using advanced computational methods led to similar conclusions [69–71]. As molecular structures of dichalcogenides are similar, this rule can be easily extended to diselenides and ditellurides [70–73]. It has to be noted that the longwavelength Cotton effect is of differing nature at different values of the torsion angle. The two lowest-energy electronic transitions are described as nA –σ * and nB –σ *. In the range 0◦ –90◦ , nA is a HOMO whereas nB is HOMO(−1). The highest energy difference between nA and nB orbitals is calculated for values of the dichalcogenide torsion angle. This difference diminishes at around 90◦ where nA and nB orbitals are almost degenerate. For the torsion angle greater than 90◦ , nB becomes HOMO and this leads to a change of the rotational strength. Similar switch of orbital positions takes place at −90◦ (Figure 2.11) [71, 72]. The first attempt to experimental study of chiroptical properties of dichalcogenide chromophore was made by Djerassi et al. [74], who recorded the ORD spectra for cystine and selenocystine. They observed negative Cotton effects at 250 and 290 nm, respectively. Ringdahl et al. [75] have found a long-wavelength Cotton effect at around 320 nm in the ECD spectra of selenocystine Other examples of CD spectra are provided by 2,2-dithioand 2,2-diseleno derivatives of propionic acid [76, 77]. It was expected that the CD spectroscopy based on the disulfide chromophore will find application in the analysis of protein structure. However disulfide CD bands in proteins are difficult to identify since the average number of disulfide bonds in proteins is low and the CD band originating from

0° −C.e.

+C.e.

180°

90°

+C.e.

Δe 0

−C.e. −90° −180

−90

0

90

180

Torsion angle R-S-S-R

Figure 2.10. Relation between the torsion angle in R–S–S–R disulfides and observed sign of the long-wavelength Cotton effect.

54

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

l [nm] 480

Rotational strength (R)

20

460 10

440 420

0

400

−10

360 340

−20

Figure 2.11. Rotatory strength (R ×10−40 cgs units,

320

solid line) and wavelength (dashed line) of the long-wavelength transition for dimethyldiselenide. (Data from reference 72.)

380

0

30

60

90

120

150

Dihedral angle C-Se-Se-C

180

the disulfide chromophore is often overlapped by the CD bands originating from aromatic residues, as well as from the amide chromophore. Laur [78] presented the CD spectra of rigid cyclic (9S , 10S )-trans-2,3-dithiadecalin (18a) and (9S ,10S )-trans-2,3-diselenadecalin (18b) and open-chain disulfide (19a), diselenide (19b), and ditelluride (19c) substituted by (S )-2-methylbutyl groups (Figure 2.12). For rigid trans-decaline derivatives (18a, 18b) in a chair–chair conformation, C–Ch–Ch–C bond forms a left-handed (M ) helix. The long-wavelength Cotton effects for 18a and 18b are negative, in direct correlation with the twist direction of the helix. Due to distortion of the dichalcogenide chromophore from the optimum 90◦ value to ∼60◦ , the position of the UV absorption bands and consequently of the Cotton effects is red-shifted by around 40 nm. The torsion angle of the open-chain derivatives 19 is not confined to just one value, so it is expected that these compounds exist as a mixture of freely interconverting diastereoisomers with the torsion angle C–Ch–Ch–C around 90◦ or −90◦ . A small energy difference between distereoismers due to steric repulsion between the alkyl chains leads to a small difference in their population and to the rise of small, but nonzero, Cotton effects. Indeed, the observed Cotton effects are about 30–50 times lower than those observed for trans-decalin 18.

0.4

4

0.2

2

H Ch Ch H 18 a, Ch = S b, Ch = Se

Ch Ch

19 a, Ch = S b, Ch = Se c, Ch = Te

0 Δe

Δe 0.0 −0.2

18b 19b

−0.4 250

300

350

400

450

−2 −4

500 l[nm]

Figure 2.12. ECD spectra of the diselenadecalin (18b) (right-hand scale) and diselenide (19b) (left-hand scale) in acetonitrile. (Redrawn using the data from reference 78.)

55

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

Further examples are the CD spectra of sugar-substituted diselenides and ditellurides [79]. The authors characterize the observed bands but do not consider the effect of chiral substituents on the dichalcogenide conformation. Recently, the CD spectra of diglycosyldisulfides and diselenides were reported, and a good agreement between the experimental and the calculated CD spectra of di(teraacetylated glycosyl)diselenide using the TDDFT method was obtained [71]. The CD spectra of several symmetrical diselenides with chiral alkyl substituents allowed us to determine the effect of chirality of the alkyl groups on the chirality of the diselenide moiety with the aid of TDDFT calculation. Thus, the S configuration of the carbon atom adjacent to the Se atom correlates with positive long-wavelength Cotton effect. This effect was explained by a model based on steric repulsion between bulky substituents [72]. The CD spectra measured for the dichalcogenides substituted with the same chiral alkyl groups show a shift of the position of the long-wavelength n –σ * Cotton effects toward a longer wavelength and a significant reduction of the Cotton effect intensity on going from S to Te. This trend is observed in the CD spectra of di[(S )2-methylbutyl] derivatives of dichalcogenides [78] and also for dineomenthyl derivatives 20 (Figure 2.13). The decrease of the Cotton effect intensity in this series is a result of elongation of the chalcogen–chalcogen and chalcogen–carbon bonds, as shown in Table 2.2, which causes significant reduction of steric repulsion between the substituents and hence lowering of free energy difference between the diastereomeric conformers of dichalcogenides of M and P helicity. Contemporary procedure for the analysis of ECD spectra of chiral dichalcogenides requires computation of conformer distribution and then the average CD spectra [71, 72, 80]. However, from a practical point of view the information about helicity of the dichalcogenide moiety can be obtained directly from the sign of the Cotton effect due to the long-wavelength nA –σ * transition. Higher-energy nB –σ * transition usually overlaps with other transitions and may be not distinguishable in the observed CD spectra. Orbitals nA and σ * involved in the long-wavelength transition are localized on the dichalcogenide moiety and are not affected by alkyl substituents. A further simplification of the analysis of rotational strength of complex chiral dichalcogenides can be based on the data obtained for dimethyl dichalcogenide. Calculation of the whole system seems justified only for structures with the C–Ch–Ch–C torsion angle very close to 90◦ (or −90◦ ) where nA and nB are almost degenerate and their exact energy may depend on the substituents in the molecule.

1.6 Ch Ch

1.2 Δe

20 a, Ch = S b, Ch = Se c, Ch = Te

20a 20b 20c

0.8 0.4 0.0 250

300

350

400

450

500

550 l[nm]

Figure 2.13. ECD spectra of the dineomenthyl disulfide 20a, diselenide 20b, and ditelluride 20c in cyclohexane solution.

56

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

2.3. THE AMIDE CHROMOPHORE IN BI- AND POLICYCLIC β-LACTAM RING SYSTEMS The amide chromophore has been recognized as a basic unit in a variety of bioactive compounds. Among them the β-lactam family of antibiotics represents some of the most clinically relevant antibiotics known [81–84]. This is attributable to their broad spectrum of antibacterial activity and a relatively low level of toxicity [83, 85]. To connect the biological activity with a defined stereochemistry of β-lactams, many research groups were engaged in the analysis of their chiroptical properties [86–99]. In the case of monocyclic β-lactams the planarity of the amide chromophore was demonstrated [90, 96, 99, 100]. The assumption of planarity of this chromophore has allowed to explain many features of protein structure. The (3R, 4S )-3-phenoxy-4-vinylazetidin-2-one presented below is an example of monocyclic β-lactam with a planar β-lactam moiety (Figure 2.14). The O=C2–N1–C4 torsion angle equal to 177.28◦ , derived from its X-ray structure, clearly demonstrates the planarity of amide chromophoric system. According to Moscowitz [101–103], such a planar amide chromophore has to be categorized as inherently achiral but chirally perturbed by its neighborhood. Therefore, one has to take into consideration the chiral perturbation (mainly through space), and this is usually achieved by the help of different sector rules. However, based on both the X-ray and computational studies, it has been demonstrated that in small cyclic peptides and medium-sized lactams the amide chromophore can be slightly nonplanar [94, 97, 104–108]. The skewness of the amide unit causes its inherent chirality, and thus such a nonplanar chromophore now belongs to the Moscowitz inherently chiral class of chromophores [101–103]. These chromophores are characterized by very strong CD effects mainly governed by their chirality. It means that the contributions from all the other atoms and groups may be neglected, so that the rules correlating the stereostructure with CD data can be classified as “chirality rules” or “helicity rules.” As a result of the extensive studies of lactam chromophores, several sector and helicity rules for the correlation between the structure and Cotton effect (CE) signs of n –π* transition have been established. Among them the β-lactam octant rule [90, 97, 109], Weigang’s sector rule [100], a modification of Weigang’s lactam rule [99], and Ogura’s [91] and Wolf’s [96] helicity rules can be mentioned. These empirical rules correlate the sign of the CE designated as the n –π * transition of the β-lactam chromophore with the absolute configuration of monocyclic azetidinones. The 4π electrons and two free electron pairs of the amide chromophore are located on the carbonyl oxygen atom. The molecular orbital (MO) occupied by the highest energy electron pair is largely (80–90%) located on the carbonyl oxygen and is a 2p orbital. Its axis is located in the plane of the amide group and is perpendicular to the direction of the

Figure 2.14. Crystal structure of (3R, 4S)-3-phenoxy-4-vinylazetidin-2-one. Thermal ellipsoids are shown at 50% probability level.

57

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

C=O bond. The second pair of free electrons, with a much lower energy, occupies orbitals extensively overlapping with d orbitals. This orbital is of both 2s and 2p character, and its axis is directed along the C=O bond (Figure 2.15). Four electronic transitions can theoretically occur within the amide chromophore, namely π0 –π * (around 190 nm), π+ –π * (around 140 nm), n –π * (around 220 nm), and n  –π *. The last transition is predicted theoretically but has not been observed experimentally. Amide transition π0 –π * is electrically allowed (εmax ∼ 104 M−1 cm−1 ), and the direction of the electric transition moment μ approximately defines a line connecting the N and O atoms (Figure 2.15). The n –π* transition is electrically forbidden (εmax ∼ 102 M−1 cm−1 ); however, it has a large magnetic transition moment m directed along the line passing through the C=O bond (Figure 2.15). In penicillins and cephalosporins the β-lactam unit is nonplanar and its nitrogen atom is pyramidal. The relationship between an (R) AC at the ring junction carbon atom and the positive sign of the lowest energy CE attributed to the amide n –π * transition which occurs in penicillins and cephalosporins at around 230 nm and 260 nm, respectively, was well-documented [86, 89, 93, 104, 110–112]. There is a nontrivial question whether the same regularity is valid for the oxa- and carbaanalogues of penicillins and cephalosporins. To clarify this, a study was undertaken to establish a correlation of the absolute stereostructure of a variety of β-lactam derivatives with the sign of the amide n –π * transition in their CD spectra. In addition, to rationalize the experimental results and to find out the scope and limitations of observed regularities, the calculation of chiroptical properties of β-lactam antibiotic analogues at B3LYP/TZVP level of theory were performed.

2.3.1. General Structures of Penicillins and Cephalosporins The chiroptical properties of oxacephams have been studied on variety of bi-, tri-, and tetracyclic oxacephams [113, 114]. Some of them, namely oxacephams 21–29, are presented in Chart 2.1 while the CD spectra of representative members of this group are shown in Figure 2.16. As can be seen in Figure 2.16, the investigated compounds exhibit, generally, two CD bands at around 220 and 190 nm. The 220-nm CD band can be assigned to the n –π * electronic transition of β-lactam unit, whereas the band at around 190 nm corresponds (a)

(b) π* O C μπ→π*

mn→π*

π*

π0

n→π* π0→π* n π0 n′

N

C N

n

C N

n′

π+ π+

Figure 2.15. (a) Molecular orbitals and electronic transitions within the amide chromophore. (b) Directions of the magnetic m and electric μ transition moments defining the electronic transitions within the amide system.

58

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Chart 2.1.

(a)

(b)

Figure 2.16. (a) CD spectra of selected oxacephams 21 (——), 26 (· · ·), and 27 (– – –) recorded in acetonitrile. (b) Crystal structures of compounds 26 and 28 with the crystallographic numbering scheme. Thermal ellipsoids are shown at 50% probability level.

to the π –π * excitation of the same unit. With respect to the sign of band at 220 nm, the oxacephams fall into two different groups. In the first group, consisting of compounds with (6R) absolute configuration, the sign of this CD band is positive, whereas in the second group, represented by compounds with (6S ) absolute configuration, this band is negative [113]. Based on that data, additionally corroborated by the specifics of their synthetic pathway and the X-ray analysis obtained for compounds 26 and 28 [113], it can be unambiguously established that the (6R) AC corresponds to a positive CE at around 220 nm whereas (6S ) AC corresponds to a negative sign of the same CE.[113, 114] The data indicate the nonplanarity of the amide chromophore and the pyramidal configuration of the amide nitrogen in the studied oxacephams. Thus, the β-lactam chromophore is inherently chiral, and the sense of its chirality is expressed as a right or left helicity that correlates well with the AC at C6 carbon atom. The helicity appears to be independent of the kind and the position of other substituents present in the oxacepham moiety [113, 115, 116].

59

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

The nonplanarity of the chromophore excluded application of the known sector rules for the prediction of the n –π * CE sign, since these rules were developed for the planar amide chromophore only [90, 97, 100]. In such case, the helicity rule should, in principle, be able to correlate the chiroptical properties and structure. However, for the studied oxacephams a breakdown of Ogura’s [91] and Wolf’s [96] helicity rules was found [113]. The spiral rule [107, 108] based mostly on the CD results obtained for the nonplanar α-lactams and monocyclic β-lactams that correlates the positive/negative torsional angle O=C–N–C with a negative/positive n –π * CE, respectively, in general, was valid for oxacephams studied. However, it was found that the absolute configuration of the bridgehead carbon atom determines the sign of the O=C–N–C torsional angle. Therefore, to connect directly the AC of this carbon atom with the sign of the CE due to n –π * transition observed around 220 nm in oxacephams, a simple helicity rule has been proposed [113]. According to this rule, a positive sign of the 220 nm CE corresponds to an (R) AC at the bridgehead carbon atom, whereas a negative sign of the same CE is related to a (6S ) AC. The rule was experimentally demonstrated to be correct for a variety of oxacephams [114–119]. The DFT conformational analysis of oxacepham 21 at B3LYP/TZVP level of theory indicates the skewness of the β-lactam unit by negative O9–C8–N1–C2 and O9–C8–N1–C6 torsion angles of −24.1◦ and −175.7◦ , respectively. The six-membered ring in the computed lowest energy conformer of oxacepham 21 is in a chair conformation with β-lactam ring in energetically favorable equatorial position. Furthermore, it has been found that the positive long-wavelength CE around 220 nm (Figure 2.16) is in an excellent agreement with the TD-DFT simulated positive CD band of 21 (Figure 2.17a) and also follows the helicity rule. The band at 220 nm has mainly the character of an amide n(O)–π * transition. In summary, the agreement between simulated and experimental CD spectra confirms not only the absolute configuration and conformation of 21 but also the validity of helicity rule for this oxacepham. The same helicity rule works very well also for cephams [118], as demonstrated in Figure 2.17b with cepham 30 as representative example of this group of compounds. As expected for the (R) AC at the ring junction, cepham 30 exhibits, similarly to oxacephams, a positive CE at around 220 nm. The six-membered ring of 30 is in a chair conformation, equally as in 21. Similarly to previously described cephams [118], the conformational flexibility in 30 is limited to the side-chain substituent at C7. The simulated CD spectrum (a)

(b)

Figure 2.17. (a) Simulated at B3LYP/TZVP theory level CD spectrum of 21 (– – –) compared with its experimental CD spectrum (——) and calculated structure of the lowest-energy conformer of 21. (b) Simulated at B3LYP/TZVP theory level CD spectrum of 30 (– – –) compared with its experimental CD spectrum (——) and calculated structure of the lowest energy conformer of 30. The vertical bars represent calculated at B3LYP/TZVP level rotatory strengths. Protons are omitted for clarity.

60

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of lowest energy conformer is in a good agreement with experiment (Figure 2.17b). The positive band at 212 nm is caused by transitions out of the amide n(N) into the amide n(O)π* orbital, as well as by transitions out of the sulfur lone pair into the same acceptor orbitals. The negative excitations at around 235 and 207 nm result mainly from the sulfur lone pair transitions, with some admixture of the amide n(O) as donor orbital. With a torsion angle O=C–N–C6 equal to −175◦ for 30, the helicity rule is satisfied. The applicability of the helicity rule was also tested for carbacephams [118]. Similarly to the earlier discussed cephams and oxacephams, two bands are present in their CD spectra in the 280- to 190-nm spectral range. The long-wavelength band assigned to the n –π * excitation occurs at around 220 nm while the second band, of the π –π * origin, appears at around 200 nm. As shown in Figure 2.18, carbacephams with (6R) configuration possess a positive CE at around 220 nm while carbacephams with an (S ) AC at C6 have a negative CE in the same spectral range. The signs are exactly what one would expect to see applying the helicity rule. Thus, the CD spectra of discussed compounds conform to the rule developed for oxacephams. The presence of isolated double bond in compounds 32 and 33 does not influence their CD spectra, which appear similar to their saturated counterpart 31 (Figure 2.18a). Completely different situation occurs in the case of compounds 34 and 35 where a conjugated double bond is present. The presence of such an α,β-unsaturated amid chromophore, defined also as cephem, causes a red shift (by about 30–40 nm) and the crucial CD band arises, depending on actual substitution, at 250 nm or 260 nm (Figure 2.18b). Except for other differences, compounds with substituents at C2 absorb at longer wavelength when compared to those unsubstituted at this carbon atom. This regularity is independent of the type of heteroatom present at position 5 in a six-membered ring fused with an azetidinone ring. Additional examples, including the 7-aminocephalosporanic acid, the active nucleus for the synthesis of cephalosporins and intermediates, and its derivatives, can be found in the recently published review [119]. The model carbacepham 31 demonstrates the impact of conformational factors on the CD spectra. As predicted by DFT calculations, the six-membered ring of the lowestenergy conformer of this carbacepham exists in a chair conformation [118, 119]. Note also that the only chromophore of 31 that absorbs above 180 nm is the amide group of the azetidin-2-one ring. It is thus very unexpected that despite this apparent structural simplicity, the computed CD spectrum of the lowest-energy conformer of 31 displays little, if any, resemblance with the experimental CD spectrum, shown in Figure 2.19. The computed amide n(O)–π * CD band is considerably blue-shifted, while the amide n(N)–π ∗ band is not only notably shifted to a higher-energy region but even has the wrong sign.

Figure 2.18. (a) CD spectra of carbacephams 31 (——), 32 (– – –), and 33 (· · ·) recorded in acetonitrile. (b) CD spectra of carbacephems 34 (——), 35 (·– · – · –·) and oxacephem 36 (– – –) recorded in acetonitrile.

61

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

Figure 2.19. Simulated CD spectrum of carbacepham 31 at 0 K (– – –) and

5

OH H H

20

N

10

O

31

0

0 −5

−10

−10

−20 190

Rvel⋅10−40 cgs

Δε (M−1cm−1)

10

350 K (· · ·) compared to experiment (——). The 0 K curve corresponds to the optimized PBE0/SV(P) conformer of 31 shown on the right. (From J. Frelek, P. Kowalska, M. Masnyk, A. ´ Kazimierski, A. Korda, M. Woznica, M. Chmielewski, F. Furche, Circular dichroism and conformational dynamics of cephams and their carba and oxa analogues, Chem. Eur. J. 2007, 13, 6732–6744. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.)

230 210 250 Wavelength (nm)

In order to investigate the impact of thermal effects on the CD spectrum, molecular dynamics (MD) simulations of 31 at 350 K were carried out and demonstrated that the saturated six-membered ring system shows a considerable flexibility. The results of the MD simulations revealed that the CD spectrum is highly dependent on the conformation of both the four- and the six-membered rings, with sign inversions occurring for both bands at various points of the MD simulation [118]. The simulated CD spectrum at 350 K compares much better with the experiment than the 0 K spectrum (Figure 2.19). Both bands are broadened and red-shifted compared to the 0 K spectrum and have the correct sign. This result is not surprising because the experimental CD spectra were measured at room temperature and thus represent a thermal average over the CD spectra of many different conformations. This effect more clearly reflects the spectrum at 350 K, taking into consideration all possible conformations. On the basis of the aforementioned discussion, it can be stated that regardless of the presence of carbon, oxygen, or sulfur atom at 5 position of the six-membered ring, all cepham analogues absorb in the same absorption range. The sign of the decisive CE at around 220 nm depends on the AC at C6 only. Thus, oxacephams, cephams, and carbacephams with (6R) AC display a positive CD band in this spectral region, whereas

(a)

(b)

Figure 2.20. (a) CD spectra of penicillin V (38) (– – –) and penam 39 (——) recorded in water and acetonitrile, respectively. (b) Simulated at B3LYP/TZVP theory level CD spectrum of lowestenergy conformer of 39 (– – –) compared with its experimental CD spectrum (——) and calculated structure of lowest energy conformer. The vertical bars represent calculated at B3LYP/TZVP level rotatory strengths. Protons are omitted for clarity.

62

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

their counterparts with (6S ) AC exhibit a negative sign of the n –π * CE in the same spectral range. Therefore, it can be concluded that all those groups of compounds are subject to the helicity rule. In the case of conformationally flexible compounds, however, the additional theoretical support for the experimental data is recommended. The penams also conform perfectly to the helicity rule. As can be seen in Figure 2.20, the penam representatives 38 (penicillin V) and 39 possess a positive CE in the longwavelength part of the spectrum in accordance with the (R) stereochemistry at the ring junction. The shift of the CD maxima into the lower-energy spectral region by ≈10 nm is related to the penam-constrained backbone. The bicyclic system of penams is relatively rigid, and its conformational lability is largely restricted to the side-chain substituent at C6 carbon atom. Its relatively unrestricted mobility is evident, considering the results of conformational search for penam 39 for which seven conformers in the energy range of 2.8 kcal mol−1 are found. In respect to the five-membered ring conformation space, however, only two conformers exist. In the first one, populated at the conformational equilibrium over 88%, the five-membered ring is in an envelope conformation with C2 carbon atom below the average plane passing through the ring (Figure 2.20). In the second conformer, the five-membered ring adopts again an envelope conformation but with the sulfur atom above the ring plane. Beyond these two, the other individual conformers show differences only in the conformation of substituent at C6 carbon atom which demonstrates its relatively substantial flexibility. The computed O=C–C7–N1–C2 and O=C–C7–N1–C5 torsion angles for all conformers of penam 39 are calculated to be negative, thus providing corroborating evidence for the nonplanarity of amide chromophore. In simulated at B3LYP/TZVP theory level ECD spectra the sign of the lowest-energy excitations is positive for all conformers, as predicted by the helicity rule for (5R) AC. The positive CE at around 240 nm is in accordance with both the helicity rule and the calculations. Based on this we can conclude that for penams the requirements of the helicity rule are met, and therefore the rule can be successfully applied to this group of compounds. Oxaanalogues of penicillins, commonly referred to as clavams, exhibit the same shape of the CD spectra as oxacephams and penams (Figure 2.21). The positive sign of the decisive CD band arising at around 240 nm in clavam 40 corresponds to its (R)configuration at C5, and the negative one corresponds to the (S )-configuration of the ring junction in its local enantiomer 41. This finding validates the proposed rule for clavams too. As can be seen in Figure 2.21a, the shape of CD spectra of clavams corresponds very well with the shape of spectra of natural penicillins. (a)

(b)

Figure 2.21. (a) CD spectra of clavams 40 (——) and 41 (· · ·) recorded in acetonitrile compared with the CD spectrum of penicillin V (38) (– – –). (b) Simulated at B3LYP/TZVP theory level CD spectrum of lowest energy conformer of 42 (– – –) compared with experimental CD spectrum (——) and calculated structures of its two lowest-energy conformers. Protons are omitted for clarity.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

A more complex situation is encountered when considering the spectra of clavams with an additional, interfering chromophore in the molecule such as phenyl, which substantially perturbs the electronic structure of the system and therefore significantly influences the CD spectra. However, the recent study on this subject demonstrated that even in the case like that, the requirements of helicity rule have been met and the rule can be successfully applied to correlate the structure and the respective chiroptical properties for unsubstituted and substituted clavams as well [120]. In the ECD spectra of the representative members of carbapenams 43–45, the 240nm ECD band is negative for compounds 43 and 45 with (5S ) AC and positive for 44 with (5R) AC, as predicted by helicity rule (Figure 2.22). Even the presence of interfering chromophores in compound 45 (i.e., phenoxy group and double bond) does not alter the relationship between the sign of 240-nm CD band and absolute configuration of the C5 carbon atom [121]. Similarly to clavams and penams, the bicyclic system of carbapenams is relatively rigid and its conformational lability is largely restricted to substituents at the C6 carbon atom and in the five-membered ring. Although for carbapenam 43 nine conformers were found in the energy range of 2.4 kcal mol−1 , in respect to the five-membered ring conformation space, however, there were only two distinct conformers present. In the first one, populated in the conformational equilibrium over 75%, the five-membered ring is in an envelope conformation with C4 carbon atom below the average plane passing through the ring. In the second conformer the five-membered ring adopts a halfchair conformation with C2 and C3 carbon atoms located above and below the ring plane, respectively (Figure 2.22). Beyond these two, the other individual conformers show differences in the substituent at C6 carbon atom which demonstrates its relatively substantial flexibility. The average ECD spectrum shows very close agreement between experiment and theory, thus providing evidence that these conformers are present in solution under given conditions (Figure 2.22). The negative CE at around 240 nm is in accordance with both the helicity rule and the calculations [121]. In order to show that the nonempirical correlation between chiroptical properties and stereochemistry is not restricted specifically only to very simple cases, more complex

Figure 2.22. CD spectra of carbapenams 43 (——), 44 (·–·–·–), and 45 (– – –) recorded in acetonitrile and simulated at B3LYP/TZVP theory level CD averaged spectrum of 43 (· · ·) as well as lowest-energy envelope (E) and half-chair (H-C) conformers of 43. Protons are omitted for clarity.

63

64

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

compounds such as carbapenam 46 which incorporates a conjugated diene chromophore at C4 in addition to the amide chromophore have been studied. Although the sign of observed at 240-nm CD band in 46 apparently is consistent with the helicity rule, CD contributions to this band of other electronic transitions can play a decisive role (Figure 2.23). In fact, the TDDFT calculations brought some more insight about the complex UV and CD relationship in carbapenam 46. The five-membered ring in this compound is nearly planar with a nitrogen atom slightly deviated from the plane formed by remaining four carbon atoms. The conformational differences in 46 are limited mostly to the conformation of the side chain. In all four conformers calculated for 46, the O8–C7–N1–C2 and O8–C7–N1–C5 torsion angles are negative, and thus helicity rule requirements are met. Consequently, the simulated CD spectrum for these geometries showed an expected positive band at ∼240 nm (Figure 2.23) [121]. The analysis show that the electronic transitions from the amide and diene chromophores are mixed and appear in approximately the same energy range. Thus, the longwavelength CD band is an admixture of the amide n –π * and diene π –π * excitations occurring at 240 nm and 252 nm, respectively (MO49 → MO∗ 52 and MO49 → MO∗ 51, respectively). Regardless of this complexity within the band at 240 nm, the positive sign of its component related to the amide n –π* excitation and its decisiveness in terms of the helicity rule are in accord with the rule. In addition, the agreement between the experimental and Boltzmann-averaged ECD spectra is very accurate and confirms both absolute configuration and conformation of carbapenam 46 [121]. It should be added that very recently the helicity rule was reformulated. The reason for this was the fact that presence of substituents in the vicinity of the ring junction may cause the change of AC descriptors of the bridgehead carbon atom from R to S , and vice versa. According to the CIP priority rules, such a change causes an allyl group at C4 in 46 whereas propionic acid methyl ester substituent attached to the same carbon atom in 45 does not. Therefore, to avoid misunderstandings, after reformulation the rule connects the sign of the n –π * amide transition with the d or l configuration, as defined by Wo´znica et al. [121].

(a)

(b)

Figure 2.23. (a) Simulated ECD and UV spectra of lowest energy conformer of carbapenam 46 (——) and experimental ECD spectrum of 46 (– – –). (b) Dominant contributions of MOs of particular excitations of carbapenam 46 lowest-energy conformer.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

As was mentioned before, the molecular dynamics (MD) simulations carried out for some carba- and oxacephams revealed a considerable flexibility of the saturated sixmembered ring and high dependency of the ECD spectra on the conformation of both the four- and the six-membered rings [113, 118]. In some cases this dependency resulted in a breakdown of the helicity rule caused by a change of the conformation of the pyranose ring from a chair to a boat. Therefore, the question arose about scope and limitations of the helicity rule. To answer this question, chiroptical properties of several model compounds with β-lactam ring fused to a seven-member ring were studied. These ringexpanded cephalosporin analogues 47–50 (Figures 2.24 and 2.25) were chosen due to their increased conformational flexibility in comparison with investigated earlier bicyclic β-lactams with six- and four-membered rings condensed together. In addition, a decrease of the skeletal strain energy expected for compounds in question may result in overall flattening of the system. Therefore, a subsequent breakdown of the helicity rule cannot be excluded. Up to four absorption bands are present in the ECD spectra of investigated compounds 47–49 in the spectral range of 185–300 nm (Figure 2.24). The band occurring around 220 nm attributed to the amide n(O)–π ∗ transition is of a particular interest because this band is the subject of the helicity rule. According to the helicity rule for compounds 47–49 belonging to the (7R)configurational series, a positive CD is expected, whereas for β-lactam ent 48 with (7S )-configuration a negative one is expected. In fact, in both cases opposite bands at around 215 nm were observed in disagreement with the helicity rule (Figure 2.24). The question arises whether compounds 47–49 constitute an exception or the rule itself is imperfect. A finding the reasons for this inconsistency appears undoubtedly very important considering the future applicability of the helicity rule. Among the factors that may play a role are the higher conformational flexibility of compounds 47–49 and/or a significant change in the geometry of β-lactam chromophore by adopting a planar conformation that does not obey the rule [113]. In order to throw light into the origin for observed deviations from the rule, TDDFT calculations were carried out for β-lactams 47 and 48 (Figure 2.24). In these cases, depending on the compound, the main ECD band is a composite of excitations out of the amide n –π * transition and transitions out of the sulfur or oxygen lone pairs into the same acceptor orbitals. In addition, the transitions out of the double-bond orbitals mix strongly with the amide n –π * transition. Some of the structures of conformers of (a)

(b)

Figure 2.24. (a) ECD spectra of β-lactams: 47 (· · ·), ent 48 (– – –); and 49 (——) recorded in acetonitrile. (b) Simulated at B3LYP/TZVP theory level average CD spectrum of 47 and 48 compared with experimental spectra (ε is the molar decadic absorption coefficient. Protons are omitted for clarity.

65

66

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 2.25. (a–g) Computed structures of conformers of β-lactam 50 and their simulated ECD spectra. (h) The Boltzmann-averaged spectrum compared to experiment (ε is the molar decadic absorption coefficient). Protons are omitted for clarity.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

β-lactams 47 and 48, calculated within the range of 3 kcal mol−1 , demonstrate presence of a small deviation from planarity of the amide chromophore manifested by a slight pyramidality of the amide nitrogen. Nevertheless, the negative sign of the decisive ECD band in both experimental and Boltzmann-averaged ECD spectra clearly demonstrates the breakdown of the helicity rule for these compounds. For saturated carbaanalogue 50, seven conformers within the energy range of 2.4 kcal mol−1 were found. To simplify and speed up the calculations, the large substituent at C8 carbon atom (OSit BuMe2 ) was substituted by a smaller OSiMe3 group. Both groups have very similar electronic properties, and the change should not significantly influence the electronic spectra. As can be seen in Figure 2.25, two of seven conformers obtained for 50, namely 3 and 4, have slightly skewed amide chromophore, whereas in the remaining five conformers the chromophoric system is planar. The population ratio of conformers with a planar and with a nonplanar chromophore in β-lactam 50 is approximately 4:1. Therefore, because the average ECD spectrum is, by definition, the sum of weighted contributions of all conformers, the negative sign of the decisive band should predominate, which is indeed the case (Figure 2.25h). However, a small positive ECD band at around 240 nm, originating from the twisted conformers 3 and 4, is present in both experimental and simulated spectra. The band at around 220 nm has primarily the character of an amide n –π * transition. The average ECD spectrum shows a very close agreement between experiment and theory, thus providing strong evidence that these conformers are present in solution under given conditions. The seven-membered ring in conformers of 50 is in a chair or a twist–chair conformation, and the azetin-2-one ring is in equatorial position at C7 (Figure 2.25). Beyond that, the individual conformers show conformational differences mostly within the substituent at C8 carbon atom. Independent evidence comes from the X-ray diffraction data for a thioanalogue of β-lactam 50, namely thiolactam 50a, the only one forming crystals suitable for such an analysis. The solid-state structural data clearly demonstrates the planarity of the amide chromophore and the chair form of seven-membered ring. Additionally, the dihedral angle C9–C7–C2–N1 equal to −2.0◦ points to the sp 2 hybridization of the amide nitrogen atom.

50

50a

In conclusion, the combined experimental and theoretical studies have shown that the ring-expanded β-lactam analogues do not follow the helicity rule since they do not belong to β-lactam type with nonplanar amide chromophore, for which the rule is valid. Therefore, these β-lactam analogues belong to the second class of Moskowitz’s chromophores [102, 103]—that is, to locally achiral but chirally perturbed. Thus, depending on the type of chiral perturbation originating from either chiral conformation of the ring incorporating achiral chromophore or from bonds closely connected to the chromophore, the chirality of the second or third sphere, respectively, should govern the CD of these compounds, as proposed by Snatzke [122]. According to this view, chirality or sector rules can correlate the stereochemistry around the chromophore with the sign of respective CD bands.

67

68

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ACKNOWLEDGMENTS The preparation of this chapter and the work of our research group (JF, AB, MW) described herein has been supported by the Ministry of Science and Higher Education, grants N N204 123507 and N N204 092935.

REFERENCES 1. Reviews and book chapters: (a) J. Gawronski, Houben-Weyl Methods of Organic Chemistry, Vol. E21a, 4th ed., G. Thieme, Stuttgart, 1995, pp. 499–533. (b) P. Salvadori, C. Rosini, L. Di Bari, in The Chemistry of Dienes and Polyenes, Vol. 1, Z. Rapoport, ed., John Wiley & Sons, New York, 1997, pp. 111–147. (c) J. Gawronski, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 305–335. (d) J. K. Gawronski, Tetrahedron 1982, 38 , 3–26. (e) N. Kirk, Tetrahedron 1986, 42 , 777–818. (f) J. Gawronski, Conformations, Chiroptical and Related Spectral Properties of Enones, in The Chemistry of Enones, S. Patai and Z. Rappoport, eds., John Wiley & Sons, New York, 1989, Chapter 3. (g) D. A. Lightner, J. E. Gurst, Organic Conformational Analysis and Stereochemistry from Circular Dichroism Spectroscopy, WileyVCH, New York, 2000, pp. 337–409. (h) G. Snatzke, Angew. Chem. Int. Ed . 1979, 18 , 363–377. (i) G. Snatzke, in Fundamental Aspects and Recent Developments in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli and P. Salvadori, eds., Heyden, London, 1973, pp. 109–124. 2. H. Suzuki, Electronic Absorption Spectra and Geometry of Organic Molecules, Methuen, London, 1963. 3. A. Rauk, H. A. Peoples, J. Comput. Chem. 1980, 1 , 240–256. 4. N. J. Bunce, J. P. Bergsma, M. D. Bergsma, W. De Graaf, Y. Kumar, L. Ravanal, J. Org. Chem. 1980, 45 , 3708–3713. 5. L. H. Hall, Group Theory and Symmetry in Chemistry McGraw-Hill, New York, 1969. 6. G. Herzberg, Molecular Spectra and Molecular Structure Van Nostrand, Princeton, NJ, 1966. 7. E. S. Stern, C. J. Timmons, Electronic Absorption Spectra in Organic Chemistry, St. Martin’s Press, New York, 1971. 8. T. Liljefors, N. L. Allinger, J. Am. Chem. Soc. 1976, 98 , 2745–2749. 9. H. Ziffer, C. H. Robinson, Tetrahedron 1968, 24 , 5803–5816. 10. J. K. Gawronski, T. Liljefors, B. Nord´en, J. Am. Chem. Soc. 1979, 101 , 5515–5522. 11. (a) A. Moscowitz, E. Charney, U. Weiss, H. Ziffer, J. Am. Chem. Soc. 1961, 83 , 4661–4663. (b) G. Wagniere, W. Hug, Tetrahedron Lett. 1970, 4765–4768 12. A. W. Burgstahler, R. C. Barkhurst, J. Am. Chem. Soc. 1970, 92 , 7601–7603. 13. A. W. Burgstahler, L. O. Weigel, J. K. Gawronski, J. Am. Chem. Soc. 1976, 98 , 3015–3016. 14. R. M. Moriarty, H. E. Paaren, U. Weiss, W. B. Whalley, J. Am. Chem. Soc. 1979, 101 , 6804–6810. 15. O. E. Weigang, Jr., J. Am. Chem. Soc. 1979, 101 , 1965–1975. 16. M. Duraisamy, H. M. Walborsky, J. Am. Chem. Soc. 1983, 105 , 3264–3269. 17. C. Djerassi, R. Records, E. Bunnenberg, K. Mislow, A. Moscowitz, J. Am. Chem. Soc. 1962, 84 , 870–872. 18. W. B. Whalley, Chem. Ind . 1962, 1024–1027. 19. (a) G. Snatzke, Tetrahedron 1965, 21 , 413–419. (b) G. Snatzke, Tetrahedron 1965, 21 , 421–438. (c) G. Snatzke, Tetrahedron 1965, 21, 439–448. (d) G. Snatzke, in Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry, G. Snatzke, ed. Sadtler Research Labs, Philadelphia, 1967, pp. 208–223.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

20. A. W. Burgstahler, D. L. Boger, N. C. Naik, Tetrahedron 1976, 32 , 309–315. 21. G. Giacomelli, L. Lardicci, C. Bertucci, A. M. Caporusso, Tetrahedron, 1978, 34 , 2015–2020. 22. H. J. C. Jacobs, E. Havinga, Rec. Trav. Chim. 1965, 84 , 932–941. 23. K. K. Cheong, A. Oshita, D. J. Caldwell, H. Eyring, Proc. Natl. Acad. Sci. USA 1970, 67 , 1727–1733 24. M. C. A. Donkersloot, H. M. Buck, J. Mol. Struct. 1988, 180 , 389–390. 25. H. Weiss, H. Ziffer, E. Charney, Tetrahedron 1965, 21 , 3105–3120. 26. E. Charney, H. Ziffer, U. Weiss, Tetrahedron 1965, 21 , 3121–3126. 27. W. Hug, G. Wagniere, Helv. Chim. Acta 1971, 54 , 633–649. 28. J. S. Rosenfield, E. Charney, J. Am. Chem. Soc. 1977, 99 , 3209–3214. 29. E. Charney, C.-H. Lee, J. S. Rosenfield, J. Am. Chem. Soc. 1979, 101 , 6802–6804. 30. D. A. Lightner, T. D. Bouman, J. K. Gawronski, K. Gawronska, J. L. Chappuis, B. V. Crist, A. E. Hansen, J. Am. Chem. Soc. 1981, 103 , 5314–5327. 31. A. Rauk, J. O. Jarvie, H. Ichimura, J. M. Barriel, J. Am. Chem. Soc. 1975, 97 , 5656–5664. 32. J. Lessard, L. Ruest, Ch. R. Engel, Can. J. Chem. 1972, 50 , 1433–1437. 33. A. E. Hansen, K. L. Bak, J. Phys. Chem. A 2000, 104 , 11362–11370. 34. R. Bucourt, Topics in Stereochemistry N. L. Allinger and E. L. Eliel, eds., 1974, 8 , 159. 35. P. J. Roberts, R. C. Patterson, G. M. Sheldrick, N. W. Isaacs, O. Kennard, J. Chem. Soc. Perkin II , 1973, 1978–1984. 36. B. Busetta, C. Courseille, F. Leroy, M. Hospital, Acta Crystallogr. Sec. B 1972, 28 , 3293–3299. 37. A. Cooper, G. Kartha, E. M. Gopalakrishna, D. A. Norton, Acta. Crystallogr. Sec. B 1969, 25 , 2409–2411. 38. A. Cooper, E. M. Gopalakrishna, D. A. Norton, Acta. Crystallogr. Sec. B 1968, 24 , 935–941. 39. H.-G. Kuball, S. Neubrech, A. Sch¨onhofer, Chem. Phys. 1992, 163 , 115–132. (b) H.-G. Kuball, B. Schultheis, M. Klasen, J. Frelek, A. Sch¨onhofer, Tetrahedron: Asymmetry, 1993, 4 , 517–528. (c) J. Frelek, W. J. Szczepek, H. P. Weiss, G. J. Reiss, W. Frank, J. Brechtel, B. Schultheis, H.-G. Kuball, J. Am. Chem. Soc. 1998, 120 , 7010–7019. (d) J. Frelek, W. J. Szczepek, S. Neubrech, B. Schultheis, J. Brechtel, H.-G. Kuball, Chem. Eur. J . 2002, 8 , 1899–1907. 40. L. Velluz, M. Legrand, R. Viennet, C. R. Acad. Sci. Paris. 1965, 261 , 1687–1691. 41. M. Kwit, J. Gawronski, D. R. Boyd, N. D. Sharma, M. Kaik, Org. Biomol. Chem. 2010, 8 , 5635–5645. 42. (a) F. Weinhold, in, Encyclopedia of Computational Chemistry, Vol. 3, P. v. R. Schleyer, N. L. Allinger, T. Clark, J. Gasteiger, P. A. Kollman, H. F. Schaefer III, and P. R. Schreiner, eds., John Wiley & Sons, Chichester, UK, 1998, pp. 1792–1811. (b) A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev . 1988, 88 , 899–926. (c) F. Weinhold, J. E. Carpenter, in, The Structure of Small Molecules and Ions, R. Naaman and Z. Vager, eds., Plenum, New York, 1988, pp. 227–236. (d) F. Weinhold, Computational Methods in Organic Photochemistry: Molecular and Supramolecular Photochemistry, A. G. Kutateladze, ed., Taylor & Francis/CRC Press, Boca Raton FL, 2005, pp. 393–476. 43. C. Diedrich, S. Grimme, J. Phys. Chem. A 2003, 107 , 2524–2539. 44. For the other definition of homo- and heterohelicity see, for example: M. Kwit, U. Rychlewska, J. Gawronski, New J. Chem. 2002, 26 , 1714–1717. 45. J. Gawronski, K. Gawronska, G. Buczak, A. Katrusiak, P. Skowronek, H. Suemune, Tetrahedron: Asymmetry 1996, 7 , 301–306. 46. (a) J. Gawronski, M. Kwit, D. R. Boyd, N. D. Sharma, J. F. Malone, A. F. Drake, J. Am. Chem. Soc. 2005, 127 , 4308–4319. (b) M. Kwit, J. Gawronski, Chirality 2011, 23 , 744–751.

69

70

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

47. M. Kwit, N. D. Sharma, D. R. Boyd, J. Gawronski, Chem. Eur. J . 2007, 13 , 5812–5821. 48. M. Kwit, N. D. Sharma, D. R. Boyd, J. Gawronski, Chirality 2008, 20 , 609–620. 49. (a) D. R. Boyd, G. N. Sheldrake, Nat. Prod. Rep. 1998, 15 , 309–324. (b) T. Hudlicky, D. Gonzalez, D. T. Gibson, Aldrichimica Acta 1999, 32 , 35–62. (c) R. A. Johnson, Org. React. 2004, 63 , 117–264. (d) D. R. Boyd, T. D. H. Bugg, Org. Biomol. Chem. 2006, 4 , 181–192. 50. E. Zysman-Colman, D. N. Harpp, J. Sulfur. Chem. 2004, 25 , 291–316. 51. P. Groner, C. W. Gillies, J. Z. Gillies, Y. Zhang, E. Block, J. Mol. Spectrosc. 2004, 226 , 169–181. 52. W. H. Green, A. B. Harvey, J. Chem. Phys. 1968, 49 , 3586–3595. 53. P. D’Antonio, C. George, A. H. Lowrey, J. Karle, J. Chem. Phys. 1971, 55 , 1071–1075. 54. O. Mundt, G. Becker, J. Baumgarten, H. Riffel, A. Simon, Z. Anorg. Allg. Chem. 2006, 632 , 1687–1709. 55. H. Lumbroso, J. M. Catel, G. Le Coustumer, C. G. Andrieu, J. Mol. Struct. 1999, 513 , 201–218. 56. R. Steudel, Y. Drozdova, K. Miaskiewicz, R. H. Hertwig, W. Koch, J. Am. Chem. Soc. 1997, 119 , 1990–1996. 57. W. N. Hubbard, D. R. Douslin, J. P. McCullough, D. W. Scott, S. S. Todd, J. F. Messerly, I. A. Hossenlopp, A. George, G. Waddington, J. Am. Chem. Soc. 1958, 80 , 3547–3554. 58. D. W. Scott, H. L. Finke, M. E. Gross, G. B. Guthrie, H. M. Huffman, J. Am. Chem. Soc. 1950, 72 , 2424–2430. 59. H. A. Khwaja, S. Walker, Can. J. Chem. 1983, 61 , 2107–2111. 60. R. L. Redington, J. Mol. Spectrosc. 1962, 9 , 469–476. 61. J. Hahn, P. Schmidt, K. Reinartz, J. Behrend, G. Winnewisser, K. M. T. Yamada, Z. Naturforsch. 1991, 46 , 1338–1342. 62. V. Aquilanti, M. Ragni, A. C. P. Bitencourt, G. S. Maciel, F. V. Prudente, J. Phys. Chem. A, 2009, 113 , 3804–3813. 63. D. Nori-Shargh, H. Yahyaei, J. E. Boggs, J. Mol. Graphics Modell . 2010, 28 , 807–813. 64. T. Shimizu, H. Isono, M. Yasui, F. Iwasaki, N. Kamigata, Org. Lett. 2001, 3 , 3639–3641. 65. G. Bergson, Ark. Kemi 1958, 12 , 233–237. 66. G. Bergson, Ark. Kemi 1962, 18 , 409–434. 67. J. Linderberg, J. Michl, J. Am. Chem. Soc. 1970, 92 , 2619–2625. 68. R. W. Woody, Tetrahedron 1973, 29 , 1273–1283. 69. M. Pericou-Cayere, M. Rerat, A. Dargelos, Chem. Phys. 1998, 226 , 297–306. 70. J. Seino, Y. Honda, M. Hada, H. Nakatsuji, J. Phys. Chem. A 2006, 110 , 10053–10062. ´ Ken´ez, S. Antus, L. Szil´agyi, T. Z. Illy´es, I. Szab´o, 71. T. Kurt´an, G. Pescitelli, P. Salvadori, A. Chirality 2008, 20 , 379–385. ´ 72. P. Skowronek, J. Scianowski, J. Gawronski, Tetrahedron: Asymmetry, 2006, 17 , 2408–2412. ´ 73. P. Skowronek, J. Gawronski, J. Scianowski, 12th International Conference on Circular Dichroism, Brescia 2009. 74. C. Djerassi, A. Fredga, B. Sj¨oberg, Acta Chem. Scand . 1961, 15 , 417–426. 75. B. Ringdahl, J. C. Craig, G. Zdansky, A. Fredga, Acta Chem. Scand . 1980, 34 , 735–738. 76. P. M. Scopes, R. N. Thomas, M. B. Rahman, J. Chem. Soc. C 1971, 1671–1676. 77. B. Ringdahl, J. C. Craig, A. Fredga, W. A. Bonner, Acta Chem. Scand . 1981, 35 , 467–471. 78. P. H. A. Laur, Proceedings of Third International Symposium on Organic Selenium and Tellurium Compounds, 1981, 217–229. 79. M. Michalska, G. Snatzke, Liebigs Ann. Chem. 1987, 179–181. 80. Y. Honda, A. Kurihara, Y. Kenmochi, M. Hada, Molecules 2010, 15 , 2357–2373.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

81. J. Mann, M. J. C. Crabbe, Bacteria and Antibacterial Agents, Oxford University Press, New York, 1996. 82. R. D. G. Cooper, Topics in Antibiotics Chemistry, Ellis Horwood, Chichester, 1980. 83. G. I. Georg, Bioorg. Med. Chem. Lett. (Symposia-in-print No 8) 1993, 3 , 2159. 84. A. G. Brown, D. Butterworth, M. Cole, G. Hanscomb, J. D. Hood, C. Reading, G. N. Rolinson, J. Antibiot. 1976, 29 , 668–669. 85. R. M. Sweet, L. I. Dahl, J. Am. Chem. Soc. 1970, 92 , 5489–5507. 86. D. B. Boyd, J. Am. Chem. Soc. 1972, 94 , 6513–6519. 87. D. B. Boyd, J. Med. Chem. 1973, 16 , 1195–1199. 88. R. Busson, H. Vanderhaeghe, J. Org. Chem. 1978, 43 , 4438–4441. 89. R. Busson, E. Roets, H. Vanderhaeghe, J. Org. Chem. 1978, 43 , 4434–4437. 90. H. Rehling, H. Jensen, Tetrahedron Lett. 1972, 27 , 2793–2796. 91. H. Ogura, H. Takayanagi, K. Kubo, K. Furuhata, J. Am. Chem. Soc. 1973, 95 , 8056–8059. 92. H. Ogura, K. Takeda, H. Takahashi, Chem. Pharm. Bull . 1975, 23 , 2469–2473. 93. V. Toome, B. Wegrzynski, G. Reymond, Biochem. Biophys. Res. Commun. 1976, 69 , 206–211. 94. J. McCann, A. Rauk, G. V. Shustov, H. Wieser, D. Yang, Appl. Spectros. 1996, 50 , 630–641. 95. D. Galle, M. Tolksdorf, M. Braun, Tetrahedron Lett. 1995, 36 , 4217–4720. 96. H. Wolf, Tetrahedron Lett. 1966, 5151–5156. 97. G. Barbaro, A. Battaglia, A. Guerrini, C. Bertucci, S. Geremia, Tetrahedron: Asymmetry 1998, 9 , 3401–3409. 98. Y. Chen, A. G. Petrovic, M. Roje, G. Pescitelli, M. M. Kayser, Y. Yang, N. Berova, G. Proni, Chirality 2010, 22 , 140–152. 99. T. Polonski, M. J. Milewska, Croat. Chim. Acta 1989, 62 , 129–134. 100. E. Ong, L. Cusachs, O. Weigang, J. Chem. Phys. 1977, 67 , 3289–3297. 101. A. Moscowitz, Advan. Chem. Phys. 1962, 4 , 67–112. 102. A. Moscowitz, Dansk Kemi 1962, 43 , 113–117. 103. A. Moscowitz, K. Mislow, M. A. W. Glass, C. Djerassi, J.Am.Chem.Soc. 1962, 84 , 1945–1955. 104. D. B. Boyd, J. P. Rhiel, F. S. Richardson, Tetrahedron 1979, 35 , 1499–1508. 105. I. Fric, P. Malon, M. Tichy, K. Blaha, Collect. Czechoslov. Chem. Comm. 1977, 42 , 678–686. 106. K. Blaha, M. Budesinsky, I. Fric, Z. Koblicova, P. Malon, M. Tichy, Tatrahedron Lett. 1978, 41 , 3449–3452. 107. G. Shustov, A. Kachanov, I. Chervin, R. Kostyanovsky, A. Rauk, Can. J. Chem. 1994, 72 , 279–286. 108. G. V. Shustov, G. K. Kadorkina, S. V. Varlamov, A. V. Kachanov, R. G. Kostyanovsky, A. Rauk, J. Am. Chem. Soc. 1992, 114 , 1616–1623. 109. M. Braun, D. Galle, Synthesis 1996, 819–820. 110. D. B. Boyd, C.-Y. Yeh, F. S. Richardson, J. Am. Chem. Soc. 1976, 98 , 6100–6106. 111. D. B. Boyd, J. P. Riehl, F. S. Richardson, Tetrahedron 1979, 35 , 1499–1508. 112. J.-C. Mueller, V. Toome, D. L. Pruess, J. Antibiotics 1983, XXXVI , 217–225. 113. R. Łysek, K. Borsuk, M. Chmielewski, Z. Kału˙za, Z. Urba´nczyk-Lipkowska, A. Klimek, J. Frelek, J. Org. Chem. 2002, 67 , 1472–1479. 114. J. Frelek, R. Łysek, K. Borsuk, J. Jagodzi´nski, B. Furman, A. Klimek, M. Chmielewski, Enantiomer 2002, 7 , 107–114. 115. A. Kozioł, J. Frelek, M. Wo´znica, B. Furman, M. Chmielewski, Eur. J. Org. Chem. 2009, 338–341.

71

72

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

116. T. T. Danh, W. Bocian, L. Kozerski, P. Szczukiewicz, J. Frelek, M. Chmielewski, Eur. J. Org. Chem. 2005, 429–440. 117. A. Kozioł, B. Furman, J. Frelek, M. Wo´znica, E. Altieri, M. Chmielewski, J. Org. Chem. 2009, 74 , 5687–5690. 118. J. Frelek, P. Kowalska, M. Masnyk, A. Kazimierski, A. Korda, M. Wo´znica, M. Chmielewski, F. Furche, Chemistry—Eur. J . 2007, 13 , 6732–6744. 119. M. Wo´znica, P. Kowalska, R. Rysek, M. Masnyk, M. G´orecki, M. Kwit, F. Furche, J. Frelek, Curr. Org. Chem. 2010, 14 , 1022–1036. 120. M. Chmielewski, M. Cierpucha, P. Kowalska, M. Kwit, J. Frelek, Chirality 2008, 20 , 621–627. 121. M. Wo´znica, M. Masnyk, S. Stecko, A. Mames, B. Furman, M. Chmielewski, J. Frelek, J. Org. Chem. 2010, 75 , 7219–7226. 122. G. Snatzke, Angew. Chem., Int. Ed. Engl . 1979, 18 , 363–377.

3 ELECTRONIC CD OF BENZENE AND OTHER AROMATIC CHROMOPHORES FOR DETERMINATION OF ABSOLUTE CONFIGURATION ´ Sandor ´ Tibor Kurtan, Antus, and Gennaro Pescitelli

3.1. BENZENE DERIVATIVES WITH CONTIGUOUS CHIRALITY CENTER; SECTOR AND CHIRALITY RULES The benzene chromophore is a common structural feature in numerous optically active synthetic and natural products, and hence Cotton effects (CEs) of the characteristic π –π ∗ electric transitions are regularly utilized for the determination of their absolute configurations. Above 175 nm, benzene shows three π –π ∗ electronic absorption bands, centered at 184, 204, and 254 nm and designated as 1 Ba,b (E1u ), 1 La (B1u ), and 1 Lb (B2u ), respectively [1–3]. Depending on the substitution pattern of the aromatic ring, the position and intensities of these absorption bands can be somewhat altered, but the spectrum is essentially unchanged. It is the longest wavelength CE belonging to the 1 Lb (B2u ) band, which is most often used to determine the absolute configuration of benzene derivatives. The 1 Lb transition is both electronically and magnetically forbidden in benzene, and its electronic absorption intensity derives from vibronic borrowing from the allowed 1 Ba,b transition. The 1 Lb band shows well-defined vibrational fine structure, and its CEs are associated with allowed transitions from the lowest-energy vibrational mode in the ground state to totally symmetric vibrational modes in the lowest-energy electronically excited state, and the lowest-energy CE corresponds to the 0–0 vibrational transition 1 Lb CE. Sometimes two distinct vibrational progressions appear with similar spacing and a small separation. The two progressions may have CEs with opposite signs, and in this situation the CD in the 1 Lb region appears as a succession of a sequence of minima and maxima or of maxima of alternating sign [4–6]. Upon substitution of the benzene ring, an additional intensity is attributed to the induced electric transition moment due to the substituent that destroys the symmetry of benzene [7]. In optically active benzene Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

73

74

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.1. 1 Lb Band CEs of (R)-Phenylmethylcarbinols in Methanola H

R C

OH CH3

Compound (R)-1 (R)-2 (R)-3 (R)-4 (R)-5 (R)-6 (R)-7 a Spectroscopic

CD [λ, nm (ε × 102 )

R H p-Cl p-CF3 m-Cl m-CF3 o-Cl o-CF3

268 276 268 274 271 273 270

Reference

(−17) (+2.5) (−12) (−28) (−15) (+6.7) (−13)

9 9 9 10 10 10 11

moments [8]: qCl = +6, q(CF3 ) = −9 [(cm mol)/L]−1/2 .

derivatives, the rotational strength of the 1 Lb CEs is influenced by the 1 Ba,b transitions through vibronic coupling as well as by the induced transition moments of the aromatic substituents. In previous papers [10, 12–16] and a review [17] on the ECD study of phenyl- and benzylcarbinols, phenyl- and benzylcarbinamines, and 1-substituted indans and tetralins, Smith applied an empirical sector rule to describe the vibronic contribution to the 1 Lb CE in monosubstituted benzene derivatives with a contiguous chirality center. This sector rule divides the space around the benzene chromophore into 12 sectors, but it is simplified to a quadrant rule for monosubstituted benzenes, since only sectors surrounding the chirality centers have to be considered (Figure 3.1). In the quadrant rule, the plane of the benzene ring defines a nodal plane, while the other one, perpendicular to the former, is allocated by the attachment bond of the benzylic carbon as shown in Figure 3.1. In substituted benzene derivatives with hydrogen atom at the contiguous chirality center, the benzylic hydrogen eclipses or nearly eclipses the plane of the benzene ring as supported by X-ray [18, 19], 1 H NMR [20, 21], and molecular modeling [22]. Since the observed chiroptical properties are dependent on both the conformation and absolute configuration, the knowledge of the proper conformation regarding the rotation about the benzylic

(a)

(b)

(c)

H C* R1

H

H C*

R2

R1

* R2

R1

R2

Figure 3.1. (a) Sector rule for third sphere contribution to the 1 Lb CE in monosubstituted benzene derivatives with contiguous chirality center. The plane of the benzene ring is also a nodal plane; signs are for the upper sectors. (b) Quadrant sector rule of monosubstituted benzene derivatives with benzylic chirality center. The plane of the benzene ring is also a nodal plane; signs are for the upper sectors. (c) Quadrant sector rule with signs of all the four sectors viewed from the direction of the benzylic carbon. Thick line represents the benzene ring, defining a nodal plane. Benzylic hydrogen eclipces or nearly eclipses the plane of the benzene ring.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

attachment bond is crucial for the correct application of the sector rule. If the benzylic hydrogen is not totally eclipsed with the benzene ring as observed by Pescitelli et al. [23] (vide infra) in the ECD analysis of PhCH(Me)t-Bu supported by TDDFT-ECD calculation, and by Butz and co-workers for α-phenylethyl alcohol (1) from gas-phase data [24], the application of the benzene sector rule becomes ambiguous (Table 3.1). If the benzylic hydrogen lies in the plane of the benzene ring (nodal plane), it does not have significant contribution, while the R1 and R2 groups are in front and rear sectors with negative and positive contributions, respectively. Based on experimental ECD data, a sequence of magnitudes of the contribution to the 1 Lb CE of various groups (related to group polarizabilities) has been determined [13]. For example, in (R)-α-phenylethyl alcohol [(R)-1, R1 = Me, R2 = OH, Figure 3.1], the methyl group, located in the negative lower front sector, has larger rotatory contribution than the hydroxyl group, which implies a negative 1 Lb band CE as found experimentally for (R)-1. Similarly, the sector rule can predict the sign of 1 Lb band CE and thus determine the absolute configuration from the experimental 1 Lb band CE for monosubstituted benzene derivatives with benzylic chirality centers if the sequence of rotatory contributions are known for the benzylic substituents R1 and R2 (Figure 3.1). It must, however, be stressed that the relative order of magnitude, for example, the methyl vs. the hydroxyl group contribution in 1 has been disputed [24]. While it is only the vibronic contribution (orientation and sequence of R1 and R2 ) that determines the 1 Lb band CE of monosubstituted benzene derivatives, with additional achiral ring substituents there is an additional induced rotatory contribution to the 1 Lb CE, which may have the same or opposite sign as that of the vibronic contribution. The magnitude of the induced electronic transition moment is related to the spectroscopic moment of the ring substituent, introduced by Platt [25] and Petruska [8]. Depending on the spectroscopic moment and ring position(s) of the substituent(s), the induced rotatory contribution can reinforce, decrease, or even override the vibronic contribution, whose effects are summarized by a chirality rule [17]. When (R)-α-phenylethyl alcohol [(R)1)] is substituted with a chlorine atom, having a negative spectroscopic moment, the induced rotatory contribution overrides the negative vibronic contribution and results in positive 1 Lb CEs for o- and p-substitution [(R)-6 and (R)-2, respectively], while for msubstitution the negative 1 Lb CE is preserved [(R)-4]. In contrast, when the trifluoromethyl group, having a positive spectroscopic moment, is the aromatic substituent, o-, m- and p-substituted derivatives [(R)-3, (R)-5, (R)-7] give equally negative 1 Lb CEs; i .e. the induced rotatory contribution of the trifluoromethyl substituent does not overshadow the negative vibronic contribution of the chirality center. It seems that phenylalkylcarbinols and phenylalkylcarbinamines ortho- or para-substituted by an atom or group with positive spectroscopic moment (Cl, Br, CH3 , OH, OMe) show 1 Lb CEs of opposite sign to that of the unsubstituted parent compound. In contrast, derivatives having a substituent with negative spectroscopic moment (CF3 , CN) in the meta position have the same sign of 1 Lb CEs as that of the unsubstituted parent. As a summary, the unambiguous configurational assignment of benzene derivatives with a benzylic chirality center by semiempirical rules has to meet the following conditions: 1. Benzylic hydrogen is eclipsed with the plane of the benzene ring or else the exact conformation must be known. 2. Reliable priority order for the rotatory contributions of the benzylic R1 and R2 substituents (vibronic contribution to the 1 Lb CE).

75

76

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3. Clear-cut allocation of the R1 and R2 substituents to sectors. 4. When vibronic and induced contributions have opposite signs, one has to know which is the dominant contribution. Even in the absence of auxochromic substituents on the phenyl ring, the consistency of the sector rule for benzene 1 Lb transition has been recently called into question by different reports having in common the use of high-level quantum-mechanics calculations for predicting ECD spectra. It must be stressed that ECD calculations offer a valuable chance for validating sector rules, the benefit of which the original authors of the rules didn’t have. In fact, provided that a reliable calculation method is chosen, the calculations allow one to observe the chiroptical response of a molecule in a certain specific geometry (that used as an input). Moreover, such a geometry may be manipulated accessing hypothetical conformations which cannot be physically observed. Pescitelli et al. have reported the ECD spectra of a homologous series of simple chiral aliphatic compounds (R)-PhCH(Me)R [(R)-8-11] with R = Et, nPr, i Pr and tBu, respectively [23]. Compound (R)-11 showed a negative 1 Lb CE in agreement with Smith’s rule; however, the lowest-energy conformation for this derivative had the C–H bond not coplanar with the ring as prescribed by the rule, thus the agreement was fortuitous. The lowest-energy conformations of the remaining compounds do have the prescribed conformation. Low-temperature ECD spectra (at 183 K), which should be dominated in all cases by the respective lowest energy conformer, consist of a series of more intense maxima with positive sign alternated by a series of weaker maxima with negative sign (Figure 3.2). The first series is allied with the allowed vibrational progression with spacing 920–1000 cm−1 , and its sign is at odds with that predicted by the 1 Lb CE sector rule. Time-dependent density functional theory (TDDFT) calculations reproduced instead the dominant sign of 1 Lb CE’s, although a more correct treatment would necessarily include vibronic effects [26, 27]. A further exception to the 1 Lb CE sector rule has been reported by Butz et al. [24] concerning (R)-α-phenylethyl alcohol [(R)-1]. According to these authors, the apparent consistency between the negative 1 Lb CE observed for (R)-1 and the sector rule is due to an incorrectly assumed conformation. The correct lowest-energy structure, found by geometry optimizations and gas-phase experiments, has the methyl bond roughly perpendicular to the plane of the phenyl, and the C–H bond is well oustide from the plane. To reconcile such a structure with the observed ECD spectrum, the sector signs for the rule must be reversed. Later, the authors substantiated their conclusions by TDDFT calculations also using a solvent model [28]. In this study, the relevant dihedral angles for 1 (Ph-Cα and Cα-O) were varied systematically and their impact on the sign of calculated 1 Lb CE was ascertained. Moreover, a rigorous theoretical approach was followed by Autschbach and co-workers in their critical evaluation of the same sector rule. The authors generated true nodal surfaces delimiting the sectors by placing a negative charge on a grid set on the top of a benzene ring at a fixed distance and varying its position systematically [29]. Since the benzene ring becomes dissymetrically perturbed by the charge, its transitions acquire non-negligible rotational strengths. In this case too, the signs obtained for the 1 Lb sector rule were opposite to the original ones. The above findings demonstrate that at least for simple benzenes with chiral substituents (i.e., endowed with third-sphere chirality in Snatzke’s terminology) [30, 31], the sector rule for 1 Lb CE is far from being generally valid and should be used with caution. Since in previous papers [13–15, 17], Smith reviewed the application of sector and helicity rules to benzene derivatives with contiguous chirality center in detail, the present chapter focuses mainly on benzene natural products with condensed carbocyclic and especially heterocyclic rings such as in tetralin, dihydrobenzo[b]furan, chroman, isochroman

77

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

1

1 (a)

×20

×20

(b) 0

0 ×2

293 k 183 k

–1

–1

×2

293 k 183 k

Δε

Δε

–2

–2

nPr

Et

–3

–3 (R)-(1-methylpropyl)-benzene [(R)-8]

(R)-(1-methylbutyl)-benzene [(R)-9]

–4

–4 190 200 210 220 250 260 270 280 λ (nm)

190 200 210 220 250 260 270 280 λ (nm)

1 (c)

0 (d)

×20

0

–1 Δε

–2

×2

293 k 183 k

×20 293 k 173 k

–4 Δε

–2 iPr

–6 tBu

–8

–3

–4

(R)-(1,2,-dimethylpropyl)-benzene [(R)-10] ×.5 190 200 210 220 250 260 270 280 λ (nm)

–10 ×.5 –12

(R)-(1,2,2-trimethylpropyl)benzene [(R)-11]

190 200 210 220 250 260 270 280 λ (nm)

Figure 3.2. Experimental ECD spectra of (R)-PhCH(Me)R derivatives (R)-8-11 at room temperature (solid lines) and at 173–183 K (dashed lines) in hexane or heptane. (Reprinted from reference 23, with permission from Elsevier).

and benzodioxane derivatives. In these derivatives the phenyl chromophore is embedded in a chiral ring, thus they exhibit second-sphere chirality [30, 31]. Some of Smith’s results have been outlined above to underscore the parallelism between the theories describing the two families of benzene derivatives. The major goal of the recent chapter is to provide guidelines for nonspecialists in the determination of absolute configuration of cyclic natural products with fused benzene ring by means of benzene semiempirical helicity rules.

3.2. TETRALIN AND TETRAHYDROISOQUINOLINE DERIVATIVES 3.2.1. Tetralins and Tetrahydroisoquinolines without Aromatic Ring Substituents; P/M Helicity → Positive/Negative 1 Lb CE In terms of chromophoric system, chiral tetralin and 1,2,3,4-tetrahydroisoquinoline derivatives belong to the benzene chromophores with chiral second sphere [30, 31]. In

78

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

6 X

P helicity

positive

5a

4 3

X

X2 8a 1 8 ωC5a,C4,C3,X < 0 X: CH2 tetralin X: NH tetrahydroisoquinoline negative M helicity sign of the 1Lb-band CE 7

ωC5a,C4,C3,X > 0

5

(b)

X

Figure 3.3. (a) Snatzke’s helicity rule or correlation between the sign of the second sphere contribution of tetralin (tetrahydroisoquinoline) and the 1 Lb band CE [30, 32]. The arrow indicates the direction of the overall spectroscopic moment. P- and M-helicity refer to the absolute conformation of the nonaromatic ring. (b) Sector rules for the third (fourth) sphere contributions to the 1 Lb band CE in tetraline (tetrahydroisquinoline). The plane of the benzene ring is a nodal plane, and signs refer to upper sectors.

these cases, the achiral benzene chromophore (first sphere) is chirally perturbed by the fused chiral ring (second sphere) and the substituents of the heterocyclic ring (third sphere), which gives rise to the observed Cotton effects. In tetralin derivatives having no substituents on the fused aromatic ring, the 1 Lb CE is determined mainly by the absolute conformation [30] of the fused nonaromatic chiral ring adopting usually a half-chair conformation–that is, the P - or M -helicity defined by the ωC5a,C4,C3,X torsional angle, which in turn is directed by the absolute configuration of the fused ring (Figure 3.3). Snatzke and Ho [32] developed a so-called helicity rule for the benzene chromophore of chiral tetralin and tetrahydroisoquinoline derivatives (Figure 3.3) according to which if the benzene ring is not further substituted, P -helicity of the nonaromatic ring leads to a positive CE within the 1 Lb band and, vice versa, M -helicity is manifested in a negative one. Moreover, a sector rule [33–36] was introduced to evaluate the contribution of the third sphere, namely, the presence of substituents attached to the nonaromatic ring. This sector rule is similar to the sector rule of benzene derivatives with benzylic chirality center (Figure 3.1a) except for that an additional plane is added, perpendicular to the benzene ring and coinciding the C2 axis, resulting in 16 sectors (Figure 3.3b). There is a simple relationship between the helicity rule and sector rule of tetralin; the two non-coplanar carbon atoms of tetralin lie in positive sectors with P -helicity affording the same prediction by the two rules. Since it is the chiral sphere nearest to the chromophores that generally determines the sign of the 1 Lb CE, the helicity of the nonaromatic ring has to be considered as the dominant contribution. If the relationship between the helicity of the nonaromatic ring and the sign of 1 Lb band CE is known, the chirality (absolute conformation) of the heterocyclic ring can be deduced from the measured ECD spectrum. Since the relative configuration of the substituents at the chiral centers as well as their equatorial or axial orientation can be obtained from NMR experiments (3 JH,H , 3 JC,H , NOE effects) or X-ray analysis, their absolute configurations can be also assigned. For instance, in the (2S , 3S )-12 tetraline derivative (Table 3.2), the trans-diequatorial arrangement of the C2 and C3 methyl groups and half-chair conformation of the carbocyclic ring can be determined from NMR experiments (e.g., large value of 3 J2H,3H ); and then according to the helicity rule, the measured positive 1 Lb CE suggests that the nonaromatic ring adopts P -helicity [36]. The combination of these two data allows the determination of the absolute configuration as (2S , 3S )-12. Similarly, the positive 1 Lb CEs of (2R)-13, (1R)-14, and (4aS , 9aS )15 derive from P -helicity of the dominant conformer with half-chair conformation. In

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

order to reduce van der Waals repulsion between the 1-Me and peri 8-H in (1R)-14, the 1-Me group preferably adopts a quasi-axial orientation with a distorted half-chair conformation [36]. The measured positive 1 Lb CE implies P -helicity, which is only feasible with (1R)-14 absolute configuration. With a half-chair conformation of the fused ring B, the benzylic 9-Me group of (4aS , 9S , 9aR)-16 would be in an equatorial orientation, however, C9 moves into the plane of the benzene ring to shift the 9-Me to a quasi-axial position and thus reduces the repulsion with peri 8-H and 1-H, resulting in a twist boat conformation. Since the nonaromatic ring adopts a conformation considerably different from the half-chair one, and other equilibrating conformers may also contribute, the helicity rule fails to determine the right absolute configuration in this case. In contrast, the epimeric (4aS , 9R, 9aR)17 corroborates well the helicity rule, since the fused cyclohexene ring has half-chair conformation with an axial 9-Me as the major conformer. The tetrahydroisoquinoline derivatives 18–21 behave similarly, and M -helicity results in a negative 1 Lb CE and vice versa.

3.2.2. Tetralins with Achiral Ring Substituents Snatzke et al. [31, 39] also showed that achiral substituents of the benzene ring with large spectroscopic moment {e.g., qOMe = +21 [(cm mol)/L]−1/2 } [8] in specific positions inverted the helicity rule. This inversion was attributed to the change of the direction of the sum spectroscopic moment [8, 25, 40, 41] vector which gives the electric transition moment vector (μ)–namely, the translation of the electron charge during the transition. This effect was called the induced rotatory contribution and was described by the chirality rule (vide supra) for benzene derivatives with contiguous chirality center. In Snatzke’s terminology, the achiral ring substituents can induce the inversion of the original helicity rule, which is the consequence of rotating the electric transition moment by approximately 30◦ . Figure 3.4a shows a polarization diagram of the tetralin chromophore, in which the addition of the spectroscopic moments oriented the electric transition moment along the direction of the C2 axis of the chromophore, which gives a positive 1 Lb -band CE for P -helicity of the nonaromatic ring (helicity rule of unsubstituted tetraline). The same helicity rule is valid for 6,7-dimethoxytetralins (Figure 3.4b), since in the presence of the two methoxy substituents at position 6 and 7, the direction of the electric transition moment does not change. In contrast, when tetralin has only one methoxy or hydroxy group at C6, the sum of the spectroscopic moments rotates the electric transition moment by approximately 30◦ , which leads to a sign inversion as shown in Figure 3.4c. Similarly, the inverse helicity rule holds for 5,7-dimethoxytetralins. A systematic study [31, 39] was carried out on substituted tetralin derivatives to reveal the effect of different substitution patterns on the sign of the 1 Lb band. This study clearly demonstrated that 5,8- and 6,7-disubstituted and 5,6,7-trisubstituted tetralins follow the same helicity rule as the unsubstituted tetralin, while 6-monosubstituted and 5,6- and 5,7-disubstituted tetralins obey the inverse one (Figure 3.5). The tetralin chromophore is found in pharmacologically active chiral synthetic derivatives such as melatoninergic ligands 22 [42] and 23 [43], the 5-HT1A receptor antagonist 24 [44], and 2-aminotetralin-2-carboxylic acids 25a,b [45, 46], the absolute configuration of which could have been determined by tetralin helicity rules (Chart 3.1). In 1-aryltetralin lignan natural products such as burseranin (26) [47], the 1-aryl group has exciton coupled interaction with the fused benzene ring [48], which determines the ECD spectrum (Chart 3.1). The absolute configurations of 1-aryltetralines are elucidated

79

80

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.2. Helicity and 1 Lb band CE of Tetralin and Tetrahydroisoquinoline Derivatives 8 7

R1 8a 1

6

5

5a

4

R1 3

4a

R2 2

6

1 10 R 4

5

7 8

3 R3

R2

R2

9a

9

R3

2

NH

1

H

R3

12 : R1 = H, R2, R3 = Me

15 : R1 = Me, R2 , R3 = H

18 : R1, R2 = H, R3 = Me

13 : R1, R3 = H, R2 = Me

16 : R1, R3 = H, R2 = Me

19 : R1, R3 = H, R2 = Me

1

2

3

1

14 : R = Me, R , R = H

2

3

17 : R , R = H, R = Me

20 : R1 = H, R2, R3 = Me 21 : R1, R3 = Me, R2 = H

Compound

Helicity of the Low-Energy Conformer H

(2S , 3S )-12

Me

1

Lb band CE [λ, nm (ε)a or [θ]b

Reference

272 (+0.191)a

36

264.5 (+0.248) Me H

(2R)-13

(1R)-14

H Me

H Me Me

(4aS , 9aS )-15 H2C

H

(4aS, 9S, 9aR)-16

H 2C

H

272 (+0.161)a 265 (+0.176)

36

272.5 (+0.226)a 264.5 (+0.233)

36

272.5 (+0.303)a 265 (+0.330)

36

272.5 (-0.058), 270 (+0.012)a 265 (−0.33), 257.5 (−0.027)

36

273 (+0.568)a 266 (+0.558)

36

266 (−290)b 273 (−800)b

37 38

270 (+292)b

37

CH2

CH2 H Me

H

(4aS, 9R, 9aR)-17

Me

H2C

H

H

H N

(1S )-18

H Me

H

(3S )-19 Me

N H

81

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

TAB L E 3.2. (Continued ) 1

Helicity of the Low-Energy Conformer

Compound

Lb band CE [λ, nm (ε)a or [θ]b

37

265 (+613)b

37

265

H

Reference

(+718)b

H

Me

(1R, 3S )-20 N H Me

H

H

(1R, 3R)-21

N H a CE b CE

Me

Me

reported as ε. reported as θ .

(a)

–

5

+

+

– –

(b)

4

6

3

7

2

+

+

+

MeO

8 1 P-helicity positive1LbCE

– –

+ P-helicity positive1LbCE

(d)

(c) MeO +

–

MeO

–

+

MeO +

MeO –

–

+

–

MeO –

– P-helicity negative1LbCE

+

+ OMe

OMe P-helicity negative1LbCE

Figure 3.4. Polarization diagram of the 1 Lb band, direction of the overall spectroscopic moment, and helicity rule of (a) tetralin (b) 6,7-dimethoxytetralin (c) 6-methoxytetralin, and (d) 5,7dimethoxytetralin.

from the sign of the relatively intense CE in the 270- to 290-nm region, governed by the exciton coupling of the two 1 Lb transitions, and thus the helicity rule cannot be applied for this type of tetralin derivatives [49, 50].

3.3. BENZENE CHROMOPHORES WITH FUSED HETEROCYCLIC RING In the following, the applicability of benzene helicity rules is to be discussed in O-heterocyclic natural products, in which the fused benzene ring is part of a 2,3dihydrobenzo[b]furan, isochroman, chroman, or 1,4-benzodioxane moiety (Chart 3.2). The correlation between the n –π ∗ CE and the absolute geometry will be also

82

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R1 R2

6

R3 7

R1 4

5

3

P-helicity positive 1Lb CE

2 8

tetralin: R1,R2,R3,R4= H 5,8-dimethoxytetralin: R1,R4= OMe, R2,R3= H 6,7-dimethoxytetralin: R1,R4= H, R2,R3= OMe 5,6,7-trimethoxytetralin:R1= H, R2,R3,R4= OMe

4 3

P-helicity negative1Lb CE

2

R3 7

1

R4

5

R2 6

8

1

6-methoxytetralin: R1,R3= H, R2= OMe 5,6-dimethoxytetralin: R1,R2= OMe, R3= H 5,7-dimethoxytetralin: R1,R3= OMe R2= H

Figure 3.5. Effect of achiral ring substituents of large spectroscopic moment (e.g., OMe) on the tetralin (tetrahydroisoquinoline) helicity rule.

O NHCOEt MeO

OH

NHC

n-Pr N

MeO

22

23

n-Pr

Me 24

O O O

NH2

O

COOH R

O O burseranin (26)

25a: R = H 25b: R = OH

Chart 3.1. Structures of tetralin derivatives 22–26.

R1

6

7 7a 1 O

5 4

4a 3

2 R2

2,3-dihydrobenzo[b]furan

R1

6

5 5a 4

7 8

8a 1 X

3 O2

R1

R2

X: H2 isochroman X: =O dihydroisocoumarin

R1

7 6

7 6

8 8a 1 O 2 R2 3 5 5a 4 X

X: H2 chroman X: =O chroman-4-one

8 8a 1 O 2 R2 3 5 5a O 4

1,4-benzodioxane

Chart 3.2. Chromophores with fused benzene ring.

addressed in some of the related carbonyl derivatives such as dihydroisocoumarins and chromanones. ECD spectroscopy has been often utilized for the elucidation of absolute configuration of these flavonoids by simply comparing the ECD spectra of similar derivatives without a deeper understanding of the factors that determine their chiroptical properties. We aimed to establish helicity rules for unsubstituted chromophores by the synthesis and

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

ECD study of derivatives with known absolute configurations to see whether the original or inverse tetralin helicity rule is valid for these chromophores. Since the benzene rings are substituted in most of the natural products containing these chromophores, the effect of aromatic ring substituents on the helicity rule has to be studied as well, which will help natural product chemists to choose appropriate ECD reference compound for the unambiguous determination of absolute configuration of novel natural products. The determination of absolute configuration of natural benzene-fused heterocycles has to follow the protocol outlined below: 1. Determination of the relative configuration and axial /equatorial orientation of the ring substituents by NMR methods (3 J H,H , 3 JC,H , NOE effect), X-ray single crystal diffraction, computational conformational analysis or their combinations. These studies also provide information on the conformation of the fused nonaromatic ring whether it has a half-chair, envelope, or boat conformation. 2. Determination of the absolute conformation (helicity) of the fused heterocyclic ring on the basis of benzene ECD helicity rules from the measured 1 Lb band CE. The helicity rule of the unsubstituted chromophore and the effect of ring substituents with large spectroscopic moment have to be known. The helicity of the heterocyclic ring is governed by the absolute configuration of the chirality centers and the preferred equatorial/axial orientation of the substituents. Large substituents prefer equatorial orientation due to 1,3-diaxial interaction, although benzylic substituents sometimes tend to favor axial position in order to reduce van der Waals repulsion with the peri aromatic hydrogen. 3. For a safe configurational assignment, a major conformer with known conformations and high population is required that dominates the ECD parameters. 4. By merging the information on the helicity, relative configuration, and axial/equatorial orientation of the substituents, the absolute configuration can be deduced.

3.3.1. Benzodioxane Chromophore; P/M-Helicity → Positive/Negative 1 Lb CE The benzodioxane chromophore occurs in chiral nonracemic natural flavanolignans [51–53] and neolignans [54–57] as well as in synthetic derivatives of pharmacological interest [58–62]. Antus et al. [51] prepared 1,4-benzodioxane steroid derivatives 27a–c of known absolute configuration and helicity, the ECD study of which showed that the same 1 Lb band helicity rule is valid for unsubstituted 1,4-benzodioxanes as for analogous tetralins; the P /M -helicity of the heteroring leads to a positive/negative 1 Lb band CE, respectively (Chart 3.3, Table 3.3). This result also afforded the configurational assignment of the natural flavanolignan (−)-silandrin and (−)-isosilandrin isolated from Silybum marianum [51, 53]. The unsubstituted 1,4-benzodioxane helicity rule was applied to deduce the absolute configurations of synthetic glycogen phosphorylase inhibitors 28a–b, having a 1,4-benzodioxane moiety connected to a N -(β-D-glucopyranosyl)amide unit (Chart 3.3) [60]. Their opposite 1 Lb band CE is governed by the helicity of the heteroring (Table 3.3), which in turn is dictated by the equatorial orientation of the C2 substituent and then the absolute configuration of the C2 chirality center. The chirality centers of the sugar unit are manifested only in the transitions of the amide chromophore, which are not in a mirror image fashion, but they do not interfere with the characteristic 1 Lb band CE. The ECD data of (2R, 3S )-29a,b, (S)-30a,b, and

83

84

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1 H O 2 3 O 4 H 2-H 27a β 27b α 27c β

OH

OH HO HO H 3-H β α α

R2 6

O

R1 7

O

1

OH

O

R2 7

O

R1

2 1

O 2-C 28a (R) 28b (S) OH

6

OH

R1 R2 (S)-30a CH2OAc H (S)-30b H CH2OAc

O

OH

5

O

NHCOMe

4

R2 H (2R,3S)-29a (CH2)2NHCOMe H (CH2)2NHCOMe (2R,3S)-29b

OMe

2 3

O

R1

O

1

O

1 2 3

OH R 7

2

H N

O

8

EtO

6

O 1

OH

OMe R (2S, 3S)-31a (CH2)3OH (2S, 3S)-31b CH=CHCH2OH

2

OH

O (S)-32

Chart 3.3. Structures of benzodioxane derivatives.

TAB L E 3.3. Helicity and ECD Data of Benzodioxane Derivatives cpd. 27ac 27bc 27cc (R)-28a (S )-28b (2R, 3S )-29a (2R, 3S )-29b (S )-30a (S )-30b (2S , 3S )-31a (2S , 3S )-31b (S )-32

Helicity

CE {λ, nm (ε)a or [θ]b}

Reference

M M P M P P P M M M M M

285 (−1.49)a 285 (−1.09)a 284 (+1.59)a 279 (−0.09)a,d 278 (0.04)a,d 280 (+389)b 280 (+400)b 284 (−0.18)a 285 (−0.21)a 299 (−1010)b 282 (+2549)b 315sh (−0.41), 289 (−0.58)a

51 51 51 60 60 56 56 61 61 54 54 e

reported as ε. reported as θ . c With cholestane skeleton. d Absolute configurations are erroneously shown in reference 60. e Unpublished ECD data of reference 61. a CE b CE

(2S , 3S )-31a, containing ring substituents with small spectroscopic moment in different positions, also corroborates the unsubstituted 1,4-benzodioxane helicity rule. One may expect that similarly to the isochroman chromophore (vide infra), ring substituents (devoid of any chirality center) do not invert the helicity rule, because the two benzodioxane oxygens fix the electric transition moment along the long axis of the chromophore. However, this expectation is contradicted by the example of the neolignan (2S , 3S )-31b chemically correlated to (2S , 3S )-31a, which has a C7 3-hydroxy1-propen-1-yl substituent and for which the positive 1 Lb CE derives from M helicity of its heteroring. Although (S )-32, bearing a conjugated C6 α,β-unsaturated ester moiety,

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

OMe OMe 1

O

7

1

8

O

O

OMe

2 3

2 3

Me

4

(2S,3S)-33 M-helicity negative 1Lb CE

6

5

OMe

O

Me

4

(2R,3R)-34a trans-eusiderin P-helicity negative 1Lb CE

Chart 3.4. Structures of synthetic (2S, 3S)-33 and natural trans-eusiderin [(2R, 3R)-34]. a Absolute configuration is shown as proposed in reference 63.

apparently follows the unsubstituted helicity rule, the application of this rule is prone to error in the presence of such conjugated chromophore. Compound (2S,3S )-33, a synthetic benzodioxane of known absolute configuration, was prepared from optically active 1-phenyl-1,2-epoxypropanes and showed negative 1 Lb CE pointing at M -helicity of its heteroring in agreement with the helicity rule (Chart 3.4) [63]. Its ECD data in the 220 to 250-nm 1 La region were opposite to that of the neolignan trans-eusiderin [(2R,3R)-34], on the basis of which the (2R,3R) absolute configuration was assigned to trans-eusiderin [63]. However, both (2S,3S )-33 and trans-eusiderin showed negative 1 Lb CEs, which suggests that they are homochiral if the ring substituents do not interfere. Since the C5 methoxy and C7 allyl substituents are not expected to invert the helicity rule, the absolute configuration of trans-eusiderin most likely has to be revised to (2S,3S ). The 1 La region is more sensitive to substituent effects and overlapping from other chromophores, which may explain the opposite CEs of (2S,3S )-33 and trans-eusiderin (34) in this region. These examples confirm that the benzodioxane helicity rule can be applied safely for unsubstituted derivatives or compounds in which the fused aromatic ring has alkyl or other substituents of low spectroscopic moments. However, ECD calculations are required whenever conjugated groups such as alkenyl or formyl are attached to the fused aromatic ring.

3.3.2. Isochroman Chromophore; P/M-Helicity → Positive/Negative 1 Lb CE Although the isochroman skeleton is far less common in natural products than the 2,3dihydrobenzo[b]furan one, there are several natural 3-alkylisochromans of remarkable biological activities whose absolute configurations have not been determined yet. For instance, absolute configurations of tricyclic derivatives 35a–c [64] and the anticoccidial optically active 3-methylisochroman derivative 36 [65], isolated from Penicillium sp., were not reported, as well as those of the topoisomerase II inhibitor CJ-12,373 (cis-37) [66] and the isochroman toxin trans-38 [67], natural 1,3-disubstituted isochromans with a benzylic hydroxy group (Chart 3.5). Moreover, there are several synthetic optically active isochroman derivatives reported with remarkable pharmacological activities such as selective 5-HT1D agonist [68–70], D1 agonist [71, 72], and D4 antagonist [73], which are promising for the treatment of migraine headache, Parkinson’s disease, and schizophrenia, respectively. It was shown that absolute configurations of these compounds play a decisive role in their pharmacological activities; the (S ) enantiomers of the 5-HT1D agonist 39a (PNU-109291) [68]

85

86

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H

R1

O

R2 3

R 35 a b c

R1 OH OH OH

CH3 R2 R3 H H OH H H OH

4

5

HO 6 7

3 Me 2

O 8

n-C7H15

HO O

HOOC

1

5

4

8

O

Me

OH OH trans-38 OH

R2

3

2 1 O

7

Me

OH OH cis-37

OMe 36 R1 6

MeO

HO N

O

N

NH2 HCl

39 R2 R1 a CONHMe OMe b H SO2NH2

(1R,3S)-40 (A68930)

Chart 3.5. Structures of natural (35a-c, 36-38) and synthetic (39a,b and 40) isochromans.

and the selective D4 antagonist sonepiprazole 39b (U-101387) [73] possessed a superior affinity for the binding site compared to the (R) enantiomers, while the (1R, 3S ) enantiomer of 40 (A68930) is almost exclusively responsible for the observed selective D1 agonist activity of the racemate [71]. In spite of the apparent importance of chirality in these derivatives, apart from the calculation of ECD parameters, there is no direct and general method for the configurational assignment of the isochroman skeleton available which can be used on μg quantity of a noncrystalline derivative. Thus, so far X-ray diffraction [72, 73] and correlations [68, 71] were applied to determine the absolute configurations of optically active isochromans. In order to establish a relationship between the helicity of the isochroman heteroring and the sign of the 1 Lb band CE and study the effect of ring substituents, rigid (41–43) and flexible (45a–g) isochroman derivatives with known absolute configuration and different ring substitution pattern were prepared and their ECD spectra were recorded (Figure 3.6, Scheme 3.1) [74]. Based on the synthetic steroid derivatives 41 and 42 (Figure 3.6a), the same helicity rule holding for tetralins and tetrahydroisoquinolines could be proposed for the isochroman chromophore having no aromatic substituents: P helicity of the heteroring leads to a positive Cotton effect (CE) in the 1 Lb band, and M -helicity is manifested in a negative

(b)

(a) 12 H H 11 O 12a H H 3 10a 6a 4 10 H H

H 12a

H 1 2

6a H

H 11 O

6a

O

11

12a H

P-helicity ωC–10a,C–11,O–12,C–12a>0

M-helicity ωC–10a,C–11,O–12,C–12a>0

positive1Lb CE

negative1LbCE

λmax [nm] Δε 6a–H 12a–H helicity α β M 274 (–0.05), 272 (+0.02), 270 (–0.08) α α P 272 (+0.33), 269 (+0.24) β β P 273 (–0.16), 268 (–0.08), 266 (–0.11) 1L CEs b

41 42 43

Figure 3.6. (a) Structures of steroid-fused isochroman derivatives 41–43 with the helicity of their heterorings and measured 1 Lb band CEs. (b) Helicity rule for the isochroman chromophore with no substituent on the aromatic ring represented on the example of 41 and 42.

87

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

one (Figure 3.6b) [74]. However, a negative 1 Lb band CE was measured for the 6aβ-H, 12aβ-H derivative 43 whose heteroring should adopt P -helicity, provided that ring A of the cholestane skeleton has chair conformation. This discrepancy was attributed to the contribution of a conformer with a boat conformation in ring A, but it made the former helicity rule ambiguous for the configurational assignment of natural isochromans. Thus 3-methylisochroman derivatives (+)-S -45a–g with different substitution pattern on the aromatic ring have been synthesized by ring closure of the (+)-(S )-1-arylpropane-2-ols 44a–f (Scheme 3.1), in turn prepared by kinetic resolution (44a–c) with Pseudomonas cepacia or chiral bioreduction (44d,e) [75]. Because the configuration of the chirality center is retained during the oxaPictet–Spengler ring closure of optically active alcohols (+)-(S )-44a–f, the absolute configurations of isochromans 45a–f can be determined readily, provided that the absolute configurations of the arylpropanols 44a–f are known. For this purpose, we envisaged the employment of the zinc porphyrin tweezer exciton chirality CD [76] and the Mosher’s NMR method [77], both developed for the configurational assignment of secondary alcohols [75]. The helicity and ECD data of the isochromans (S )-45a–g and (4aR,10bS )-46 are tabulated in Table 3.4. (S )-45a and (4aR,10bS )-46, bearing no substituents on their aromatic rings, have heterorings of P - and M -helicity (see Scheme 3.1 for definition and representation), respectively, wherein the C3 methyl group of (S )-45 is oriented equatorially (J3H,4H = 10.9 and 3.1 Hz) while the 4a–H and 10b–H of the trans-annulated (4aR,10bS )-46 have a trans-diaxial configuration [75]. Because the 1 Lb bands of (S )-45 of P -helicity and (4aR,10bS )-46 of M -helicity (Scheme 3.1, Table 3.4) show positive and negative CEs, respectively, it follows that the unsubstituted isochroman chromophore obeys indeed the helicity rule established for unsubstituted chiral tetralins and tetrahydroisoquinolines [32]: P /M -helicity of the heteroring results in positive/negative 1 Lb band CE , respectively. This corroborates well the similar spectroscopic moments of the hydroxymethyl (qCH2 OH = −5), aminomethyl (qCH2 NH2 = −5), and ethyl (qEt = +4.5) groups [8]. Since natural or synthetic isochroman derivatives of pharmacological interest often contain substituents with large spectroscopic moment (O-alkyl, hydroxy) on the fused

R4 R3

H

R4 Me

MeOCH2Cl

5

ZnCl2/Et2O 0°C

OR

R2

2

R2

R1 (+)-(S)-44a-f R2 R3 R1 H H a H OMe b OMe H H c OMe H H OMe OMe d -OCH2 OH e H OMe H f OMe H 45g H

44,45

H

4

R3

8

8a

H Me

Me

R5 H H H H H H H

O

1

2

ωC-8a,C-1,O-2,C-3 >0 P-helicity

4 H 5 3

1

4

O

R1 (+)-(S)-45a-g R4 H H H H H Br H

3

3

O 4a

6

10a 6a

2 10b 1 H 10

46

5

H 10b 7 8 9

O 6 4a

H

ωC-6a,C-6,O-5,C-4a 0 P-helicity

λmax [nm] Δε

266 (+0.03), 260 (+0.02), 255sh (+0.01) 287 (+0.59), 282 (+0.69), 275sh (+0.063)

Scheme 3.2. Preparation, preferred helicity, and ECD data of (−)-(1R,3S)-47a,b 1,3-disubstituted isochromans.

the heteroring has half-chair conformation [78]. The heteroring of (1R, 3S )-47a,b has P -helicity, and their positive 1 Lb CEs are practically the same as that of (S )-45a,d, which proved that the introduction of an axial benzylic alkoxy group does not change the isochroman helicity rule as long as the conformation or helicity of the heteroring remains the same. This result allows the configurational assignment of 1-alkoxy- or 1-hydroxyisochromans such as cis-37 and trans-38 from their ECD spectra. On the basis of the isochroman helicity rule, the absolute configuration of pseudoanguillosporin A (48a) and B (49), isolated from the endophytic fungus Pseudoanguillospora sp, could be deduced from their ECD spectra (Chart 3.6 and Figure 3.7) [79]. Since pseudoanguillosporin A (48a) shows a negative 1 Lb band CE [284 nm (ε = −0.4 nm)], its heteroring adopts M helicity (Figure 3.7, right), which implies a (3R) absolute configuration with equatorial C3 substituent. In contrast, the synthetic compound (3S )-45b [75] had a positive 1 Lb band CE and P -helicity. Both (3R)-48a and (3S )-45b showed a positive CE around 240 nm in the 1 La region, which is known to be more sensitive to the effect of achiral substituents of the aromatic ring. The absolute configuration of pseudoanguillosporin B (49) at C3 was deduced as (R), since its CD spectrum was nearly identical to that of 48a. Moreover, the absolute configuration of the C6 chirality center on the side chain of pseudoanguillosporin B, distant from the chromophore, was determined by the Mosher’s NMR method [79]. The interpretation of the ECD spectrum of compounds 48a and 49 (vide infra) is also supported by the calculation of the ECD spectrum of the model compound (R)-48b [79], which reproduced well the pattern of bands discussed above for (R)-48a (Figure 3.8). R4 3

R1

R O

HO

3

O OR R1 48a 48b

n-heptyl CH3

2O

6′

OH

1

2

R2 H H

1′

R3 R4 C-3 H CH3 R H CH3 R

OH (3R,6′R)-49

Chart 3.6. Structures of pseudoanguillosporins A (48a) and B (49), and the ECD model compound (R)-48b.

90

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O2

0,8

4

(3S)-45b

Δε [cm2 mmol–1]

1 R 3 H ωC-8a,C-1,O,C-2 > 0 P-helicity

0,4 0,0 (3R,6'R)-49

H 3 R

–0,4 1

(3R)-48a

4

–0,8 –1,2

O 2 220

240

260

280

300

ωC-8a,C-1,O,C-2 < 0 M-helicity

λ [nm]

Figure 3.7. Left: Measured ECD spectra of (3R)-48a, (3R, 6 R)-49 and (3S)-45b in acetonitrile. Right: P and M-helicity of the isochromane ring; pseudoanguillosporin A (3R)-48a and B (3R, 6 R)-49 have M-helicity.

2

Δε [cm2 mmol–1]

1 0 Absolute minimum –1 Experimental for (3R)-48a in ACN Calculated with BP86/TZVP on (R)-48b

–2

(Boltzmann-weighted over 4 structures at 300 K)

–3 –4

220

240

260 λ [nm]

280

300

320

Second minimum (+0.10 kcal mol–1)

Figure 3.8. Experimental CD spectrum of pseudoanguillosporin A (48a) compared with the ECD calculated on model compound (R)-48b with TDBP86/TZVP as Boltzmann average over four DFT-optimized structures (B3LYP/6-31G(d)); the two most stable ones are shown on the right.

3.3.3. 2,3-Dihydrobenzo[b]furan Derivatives Unsubstituted 2,3-Dihydrobenzo[b]furan Derivatives; P/M-Helicity → Negative/Positive 1 Lb CE. While a tetralin derivative may be described as a benzene derivative with two alkyl substituents having the same magnitude of spectroscopic moments, this condition is not given for the 2,3-dihydrobenzo[b]furan chromophore, since the spectroscopic moment of the alkoxy moiety is larger than that of the alkyl part based on the spectroscopic moment of methoxy (qOMe = +21) and ethyl group (qEt = +4.5) as determined by Petruska [8]. Therefore, it is expected that the electric transition moment vector (μ) does not lie in

91

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

TAB L E 3.5. Helicity of the Heteroring and 1 Lb Band CEs for Rigid and Flexible 2,3-Dihydrobenzo[b]furan Derivatives Compound

Standard Projection

(−)-50

O C1

1L

b

CE {λ, nm (ε),a [θ]b }

Reference

M

297 (+0.89)a ,c

80, 81

P

289 (−3.30)a ,c

80, 81

P

280 (−2.51)a ,d

81, 82

M

292 (+2.95)a ,e

81–83

M M

288 (+0.72)a ,e 292 (−0.05)a ,c

81, 83 81, 83

P

281 (−0.37)a ,e

84

H

P P

a ,e

285 (−1.61) 300 (+0.46)a ,e

84 84

CH3

M

282 (+2658)b ,e

85, 86

C4 H H

H

(+)-51

Helicity

C1 H

O

C4

(+)-(2S , 3S )-52

H Ph O

CH3 H

(−)-(2R, 3S )-53 (−)-(2R, 3S )-54 (+)-(2R, 3S )-55 (+)-(2S , 3S )-56f f

(+)-(2S , 3S )-57 (+)-(2S , 3S )-58f

H O Ar

CH2

H

OH

H Ph O

CH3

(−)-(2R, 3S )-59f O Ph

H H

CE reported as ε. CE reported as θ . Solvent of ECD measurement is c n-hexane, d ethanol, e methanol, f revised absolute configuration. Note: The standard projections show the definition of the helicity for the heteroring in the major conformer; M-helicity corresponds to negative ωC7a,O,C 2,C 3 torsional angle.

a b

the direction of the pseudo C2 axis of this chromophore, which can lead to an inversion of the original tetralin helicity rule. In order to obtain a correlation between the stereochemistry of chromophores (helicity of the heteroring) and the sign of the 1 Lb band CE, a stereocontrolled synthesis, conformational and ECD study of rigid [(−)-50, (+)-51] [80] and flexible [52–55 (Chart 3.7)] [82, 83] 2,3-dihydrobenzo[b]furan derivatives with known absolute configuration was performed. The 1 Lb band CD data and the preferred helicity of the five-membered O-heterocyclic ring in (−)-50, (+)-51, 52 (−)-53, 54 and (+)-55 are tabulated in Table 3.5. A comprehensive NMR study on the conformation of the dihydrofuran ring and ring A of the cholestane skeleton revealed that the heterocyclic ring of the cholestane derivatives (−)-50 and (+)-51 adopt M -and P -helicity, respectively. The known helicity of the heterocyclic rings in (−)-50 and (+)-51 and their measured 1 Lb band CE allowed us to set a helicity rule for the unsubstituted 2,3-dihydro-benzo[b]furan chromophore [80]. P /M -helicity of the heterocyclic ring leads to a negative/positive CE within the 1 Lb band; that is, the inverse form of the tetralin helicity rule is applicable, which is attributed to the large spectroscopic moment of furan O-1.

92

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Substituted 2,3-Dihydro-benzo[b]furans; Effect of Substitution. On the basis of the above helicity rule, the originally reported (2R, 3R) absolute configuration of trans-norneolignans (+)-56 and (+)-57, isolated from Krameria cystisoides [84], had to be revised as (2S,3S ) (Chart 3.7, Table 3.5). Although both (+)-56 and (+)-57 have a 3hydroxyprop-1-yl aromatic substituent at C5, this was not expected to invert the helicity rule due to the small spectroscopic moment of alkyl groups. Since the ECD spectra of (+)-56 and (+)-57, published by Achenbach et al. [84], exhibit a negative CE at 281 and 285 nm, respectively, their heteroring should adopt P helicity. Taking into account that the C2 and C3 substituents are equatorial in both cases (3 J2H3H 8.9 and 9.3 Hz), their absolute configurations are (2S ,3S ). Because (+)-conocarpan [(+)-58] was chemically correlated with (+)-(2S , 3S )-56, its absolute configuration had to be revised to (+)-(2S , 3S ) as well [80, 81]. However, (+)-conocarpan [(+)-58] has a positive 1 Lb CE at 300 nm (Table 3.5), which suggested that a conjugated C5 1-propen-1-yl substituent inverts the helicity rule. With (+)-(2S , 3S ) configuration, the heteroring of (+)-conocarpan preferably adopts P -helicity to ensure the low-energy quasi -equatorial arrangement of the C2 and C3 substitutents; that is, the positive 1 Lb CE derives from P -helicity of the heterocyclic ring. This observation corroborates well the large spectroscopic moment of the 1-propen-1-yl group (q = +15). Recently, conocarpan (58) possessing a wide range of biological activites [87–89] have attracted the attention of synthetic chemists and enantioselective synthesis of (−)-(2R, 3R)-conocarpan [90, 91], and its enantiomer [(+)-(2S,3S )-58] [86] was reported, which unequivocally confirmed our configurational assignment by ECD. (−)-Epi -conocarpan [(−)-59], the C2 epimer of (+)-conocarpan [(+)-58] isolated from roots of Piper regnelli [85], was synthesized and converted to (+)-(2S, 3S )-conocarpan [(+)-58] [86], which confirmed its (−)-(2R,3S ) absolute configuration (Chart 3.7, Table 3.5). (−)-(2R,3S )-epi -conocarpan has a positive 1 Lb CE at 282 nm and a heteroring of P -helicity with an equatorial orientation of the C2 aryl group, which does not follow the expected helicity rule. The cis orientation of the C2 and C3 substituents forces the dihydrofuran ring into a nearly planar conformation and thus the third sphere contribution and especially that of the pseudo-axial C3 methyl group most likely overrides the second sphere contribution—that is, the helicity of the heteroring. Flexible unsubstituted 2,3-dihydrobenzo[b]furans (+)-52, (−)-53 and (−)-54 prepared in a stereocontrolled manner also confirmed the validity of the helicity rule

6 5

7 7′a

4

1

H 1 O

4′a

O

2A 34 5

H

H

6 5

H

H

(−)-50

H

(+)-51

O

(+)-(2S,3S)-52

7

A

R2

1

O2 3

B

R3

CH2OH (−)-(2R,3S)-53: R1,R2,R3= H (−)-(2R,3S)-54: R1 = H,R2 = OMe, R3 = OH (+)-(2R,3S)-55: R1 = OMe,R2 = OMe, R3 = OH

R1 OH

HO

R1

7 7a 1 O2 Ph 3 4a 4 CH3

O

CH3

CH3

(+)-(2S,3S)-56: R1 = H (+)-(2S,3S)-57: R1 = OMe

(+)-(2S,3S)-58 (+)-conocarpan

O OH

OH CH3 (–)-(2R,3S)-59 (–)-epi-conocarpan

Chart 3.7. Structures of 2,3-dihydrobenzo[b]furan derivatives for Table 3.5.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

Ph O

93

CH3

H CH3 H

P-helicity 2,3-trans relative configuration pseudoequatorial Ph and Me groups

O H

H Ph

M-helicity 2,3-trans relative configuration pseudoaxial Ph and Me groups

Figure 3.9. Equlibrating conformers of (+)-(2S, 3S)-52 with the P-helicity conformer as the dominant one.

established for the unsubstituted chromophore based on the steroid derivatives (−)-50 and (+)-51. The P - or M -helicity of the heterocyclic ring is controlled by the equatorial arrangement of the phenyl group at C2 (Figure 3.9), whose contribution to the ECD is less significant compared to that of the 2,3-dihydrobenzo[b]furan chromophore, possibly because of increased mobility and distance relatively to the chirality elements. Thus, the substitution pattern of the C2 phenyl (ring B) does not influence significantly the 1 Lb band CE either; (−)-53 and (−)-54 have consistently positive 1 Lb band CE. On the contrary, the substitution of the aromatic ring A at C7 position by a methoxy group (qOMe = +21) [8] changes the sign of the 1 Lb band CE, since the homochiral (−)-54 and (+)-55 show opposite signs for the M -helicity of the heterocyclic ring. The weak negative 1 Lb band CE of (+)-55 proved that a substituent at C7 such as a methoxy group possessing a large spectroscopic moment (q) also reversed the helicity rule; that is, P /M -helicity of the heterocyclic ring leads to positive/negative 1 Lb band CD, respectively. Although chiroptical methods (ECD, ORD, and optical rotations) are extensively used in the configurational assignment of natural 2,3-dihydrobenzo[b]furan neolignans, the number of publications in which the absolute configuration of a neolignan was determined independently, by X-ray or chemical correlation, and its ECD was also measured, are very limited. Rare examples were presented by Yuen et al. [92], who characterized synthetic neolignans 60, 61, 64, and 65 by ECD and also determined their absolute configurations unambiguously by X-ray analysis and chemical correlations (Chart 3.8). The negative 1 Lb band CE of (−)-(2R, 3S )-60 confirmed our results regarding the effect of a methoxy group at C7 (Table 3.6), since its heterocyclic ring adopts M -helicity due to the equatorially oriented bulky aryl group at C2. The 7-methoxy group is quite a common substituent in numerous dihydrobenzo[b]furan neolignans, and thus its effect on the sign of the 1 Lb band CE could lead to many erroneous configurational assignment if not properly taken into account. The ECD data of the homochiral analogue (−)-(2R, 3S )-61 revealed that the introduction of an additional conjugated 3-hydroxy-1-propen-1-yl group at C5 does not induce a further change in the sign of the 1 Lb band CE in the presence of a 7-methoxy group. The ECD data of (−)-(2R, 3S )-60 and (−)-(2R, 3S )-61 also proved that the published absolute configurations of (−)-62 [93] and (+)-63 [84] are incorrect. In fact, ring B substituents do not influence the sign of the 1 Lb CE; therefore measured positive 1 Lb band CE for (−)-62 and (+)-63 should stem from P -helicity, which implies (−)-(2S , 3R) and (+)-(2S , 3S ) absolute configuration, respectively [81]. The neolignan licarin B, differing from (+)-63 in the substitution of ring B (3,4-methylenedioxy instead of 4-hydroxy), was used by Achenbach et al. [84] as an ECD reference, but its absolute configuration published by Aiba et al. [97] also has to be revised. In (−)-(2R, 3S )-64 and (−)-(2R, 3S )-65, a C5 α,β-unsatured carbonyl moiety is conjugated with the benzene ring and the resultant chromophore cannot be considered identical with the previous ones. If the intense negative CEs around 330 nm are

94

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OMe

OMe

OMe

7

O OMe HO

HO (CH2)3

5

O

R1

3

OH

2

CH2OH

CH2OH R

(−)-(2R,3S)-60

1

CH3 2

(−)-(2R,3S)-61: R = OH, R = H (−)−(2S,3R)-62: R1,R2 = OMe (woorenogenin)

OMe

(+)-(2S,3S)-63

OMe

HO 6

O

O R

OMe

OMe

1

O2

CH2OH

OH MeOOC 5

OH HO CH3

OH

(CH2)3 5

(+)-(2S,3S)-66

(−)-(2R,3S)-64: R = CHO (−)-(2R,3S)-65: R = COOMe

O

(2S)-67 (S)-hexahydromarmesin R2

OMe R1O

O

O Ar

HO

OH (2S)-68 wutaiensol

MeO CH3 trans-(2R,3R)-69 R1 = H or allyl R2 = H or allyl

Chart 3.8. Structures of 2,3-dihydrobenzo[b]furan derivatives for Table 3.6. Ar in (2R, 3R)-69 is O-methyl-O, O-methylenepyrogallyl, piperonyl, or tri-O-methylpyrogallyl group.

TAB L E 3.6. Helicity of the Heteroring and 1 Lb Band CEs for 2,3-Dihydrobenzo[b]furan Derivatives Compound (−)-(2R, 3S )-60c (−)-(2R, 3S )-61c (−)-(2S , 3R)-62d (+)-(2S , 3S )-63d (−)-(2R, 3S )-64c (−)-(2R, 3S )-65c (+)-(2S , 3S )-66d (2S )-67c (2S )-68 (2R, 3R)-69d

Helicity M M P P M M P P P M

CE {λ, nm (ε)a or [θ]b} 296 (−1739)b 279 (−11372)b 265 (+1.0)a 295 (+1.36), 270 (+5.30)a 335 (−9862), 279 (+2240)b 376 (+806), 327 (−9884)b 263sh (+2600)b 292 (−2435)b 272 (+3223)b negative CE from ORD

Reference 92 92 93 84 92 92 85 94 95 96

reported as ε. reported as θ . c absolute configuration was determined independently from ECD data. d Revised absolute configuration. a CE b CE

taken into account, these compounds apparently follow the inverse form of the helicity rule but accompanying weaker opposite CEs indicate that the n –π ∗ transitions can make the assignment ambiguous. The neolignan (+)-(2S , 3S )-66 (Chart 3.8, Table 3.6) possesses a C5 methoxycarbonyl group, which has a large negative spectroscopic moment

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

(qCOOMe = −17) [8]. It was shown earlier that a substituent with large spectroscopic moment at C5 such as a 1-propen-1-yl group (q = +15) inverts the helicity rule of the unsubstituted chromophore. Thus a similar behavior may be expected from a C5 methoxycarbonyl substitutent; that is, P-helicity results in a positive 1 Lb CE as shown in Table 3.6. This assumption is supported by the fact that (+)-66 was isolated together with (+)-(2S , 3S )-conocarpan (58) from Piper regnellii [85], which suggests that they are most likely homochiral derivatives. The absolute configuration of (S )-hexahydromarmesin [(2S )-67] was deduced unequivocally by a chemical correlation to (S )-marmesin [94], and the negative 1 L band CE corresponds to P -helicity of its heteroring with pseudoequatorial C2 b substituent. This implies that a C6 hydroxy substituent (qOH = +20) does not invert the helicity rule of the unsubstituted chromophore. The related (2S )-68 neolignan, wutaiensol, has a C5 3 -hydroxy-1-propen-1-yl and a C5 OMe substituent, the presence of which inverts the unsubstituted helicity rule as shown earlier. Thus the published (S ) absolute configuration is in accordance with our findings, since it shows positive 1 Lb CE and its heteroring has P -helicity [95]. Gottlieb et al. [96] classified benzofuranoid neolignans into structurally homogeneous groups by constitution and ORD curves and proposed their configurational assignment. The structure (2R, 3R)-69 represents a group of neolignans with 5,6-dioxygenated benzo[b]furan chromophore. Because we have already shown that the presence of a C5 substituent with large spectroscopic moment inverts the unsubstituted helicity rule, the same is expected for the 5,6-dioxygenated chromophore. Accordingly, negative 1 Lb CEs of 69 derivatives originates from M -helicity and (2R, 3R)-69 absolute configuration. As a consequence, the absolute configurations assigned to these neolignans may have to be revised as denoted in Chart 3.5. There are numerous publications that used incorrect absolute geometries and ECD results of the presented structures for the configurational assignment of isolated dihydrobenzo[b]furan neolignans. Whenever the effect of an aromatic substitution pattern on the sign of the 1 Lb CE was not taken into account and an improper ECD reference compound was chosen for comparison, the determination of absolute configuration by ECD data led to incorrect absolute geometry [84, 85, 93, 96–104]. In contrast, there are recent publications that use the dihydrobenzo[b]furan helicity rule [80, 81] properly taking care of the substitution pattern [54, 105–107]. As an example, the C2 and C3 absolute configuration of difengpiol A (70), a neolignan with a nonprecedented C2 cyclohexenediol and a C7 methoxy substituent, was determined as (2S , 3R) on the basis of the positive 1 Lb CE and P -helicity of its heteroring (Chart 3.9). The 9,10-dihydrophenanthrofuran derivative (−)-pleionesin A (71) was isolated from the orchid Pleione yunnanensis together with three related derivatives [108] and the dihydrobenzo[b]furan helicity rule [80, 81] was applied to determine the absolute configuration. However, (−)-pleionesin A (71) contains an inherently chiral biphenyl chromophore with an axial chirality element along the biaryl axis instead of the benzene chromophore of dihydrobenzo[b]furans, and therefore the helicity rule cannot be utilized safely for the configurational assignment. Similarly, the helicity rule should not be applied whenever there is an additional benzylic chirality center in the dihydrobenzo[b]furan lignan [109, 110] as exemplified by (−)-radulignan (72) [110], in which the contribution of the benzylic chirality center may override that of the dihydrobenzo[b]furan moiety rendering the assignment ambiguous.

95

96

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Our observations in Tables 3.5 and 3.6 can be concluded in the following points and are summarized in Figure 3.10: 1. P /M -helicity of the heterocyclic ring results in negative/positive 1 Lb CE, respectively, if the aromatic ring has no other ring substituent (we refer to this statement as the unsubstituted helicity rule). 2. Substituents with low spectroscopic moment such as an alkyl or 3-hydroxypropan1-yl at C5 and possibly in other positions do not change the unsubstituted helicity rule. 3. A hydroxyl or alkoxy group at C6 does not invert the unsubstituted helicity rule even in the presence of a C4 alkyl group. 4. 2,3-dihydrobenzo[b]furans with C5 substituents having a large spectroscopic moment such as 1-propen-1-yl, CHO, COOMe, OH, OMe follows the inverse helicity rule. 5. A C7 methoxy group in the presence or absence of C5 1-alken-1-yl substituent also causes inversion. 6. Cis-dihydrobenzo[b]furans do not follow the expected helicity rule because the contribution of the pseudoaxial benzylic substituent, belonging to the third sphere, determines the sign of the 1 Lb CE.

3.3.4. Chroman Chromophore; P/M-Helicity → Negative/Positive 1 Lb CE In chroman derivatives, the magnitude of the spectroscopic moment belonging to the alkoxy part of the molecule (qOMe = +21) [8] is approximately five times larger than that of the alkyl moiety (qEt = 4.5) [8]. Therefore, similarly to the unsubstituted dihydrobenzo[b]furan chromophore, the sum vector is rotated by more than 30◦ and a sign inversion of the tetralin helicity rule is expected (Figure 3.11). On the basis of the spectroscopic moment belonging to the N -methyl group (qNHMe = 27) [8], one can predict that the same rule holds for tetrahydroquinoline chromophore as well. Structures, ECD data, helicity, and references of selected synthetic and natural chroman and tetrahydroquinoline derivatives are tabulated in Chart 3.10 and Table 3.7. The 1 Lb band CEs of rigid synthetic steroid model compounds (−)-73 and 74 and the flexible (−)-(S )-flavan [(S )-75, see conformation in Figure 3.12a], prepared with known absolute

OH OMe HO

7

6 5

1

O 2 1′ 3

4

4′

OH

3′ 2′ CH2OH OH

HO HO

O O O

(+)-(2S,3R,3′S,4′R)-70 difengpiol A 286 nm (+0.49) HO

OMe

OH

MeO

OMe

(−)-pleionesin A (71) 290 nm (+0.53)

Chart 3.9.

O O OH

OH

O

CH2OH OH (−)-radulignan (72) 290 nm (+5.0)

97

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

configuration, confirmed unambiguously the above assumptions. In fact, P/M-helicity of the chroman or tetrahydroquinoline chromophore having no aromatic ring substituents results in a negative/positive 1 Lb band CE, respectively (Chart 3.10, Table 3.7) [81, 111]. The helicity rule is also applicable to the cis-4-hydroxyflavan (−)-(2R,4R)-76, in which both the C2 phenyl and C4 hydroxy groups adopt equatorial orientation if the heteroring has half-chair conformation with M -helicity (Figure 12b) affording a positive 1 Lb band CE [81, 112]. Since the chroman chromophore frequently has aromatic ring substituents of large spectroscopic moment (such as hydroxy or methoxy groups) in natural products, the effect of ring substitution on the chroman helicity rule has to be addressed as well. Although the corresponding 5- and 7-hydroxy or -methoxy substituted rigid model compounds have not been synthesized, the effect of substitution on the chiroptical properties can be assessed from the published ECD data of natural chroman derivatives [113]. The comparison of the ECD data of (+)-catechin [(2R, 3S )-77], (−)-rubinetinidol [(2R, 3S )78], (+)-afzelechin [2R, 3S )-79], all of which have P -helicity and negative 1 Lb band CE (Figure 3.12c), with the ECD data of unsubstituted derivatives [(−)-73, (−)-75] of P -helicity proves that neither 5-hydroxy nor 5- and 7-dihydroxy substitutions of the ring A have an influence on the sign of the 1 Lb CE [81]. The same conclusion can be made

R

4

R O

R6 7 6

5

5

O1

2

3 4 R3

R1

R2

O

1

P-helicity positive 1Lb CE inverse helicity rule

P-helicity negative Lb CE unsubstituted helicity rule −R3,R4,R5,R6 = H

−R4 = 1-propen-1-yl, R3,R5,R6 = H

−R4=

−R6 = OMe, R3,R4,R5 = H or alkyl

3

alkyl or 3-hydroxypropan-1-yl 5

6

−R6 = OMe, R4 = 3-hydroxyprop-1-en-1-yl

R ,R ,R = H −R5

4

3

6

R3,R5 = H

= OH, R = alkyl, R ,R = H

−R4 = CHO or COOMe, R3, R5,R6 = H* −R4,R5 = OH or OMe, R3,R6 = H

Figure 3.10. Dependence of the 2,3-dihydrobenzo[b]furan helicity rule on the substitution pattern of the aromatic ring. *Benzaldehyde or benzoic acid methyl ester chromophore instead of the benzene.

X X chroman: X = O tetrahydroquinoline: X = NH

X μ = Σq

P-helicity negative 1Lb CE

Figure 3.11. Platt polarization diagram of the 1 Lb band for chroman (X = O) and tetrahydroquinoline (X = NH) chromophores having no substituents on the aromatic ring. P-helicity results in a negative 1 Lb band CE. Smaller arrows represent the spectroscopic moment vectors (q), while the longer one represents the electric transition moment vector (μ).

98

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H1

1 H

2 3

R1

O

4

4

R1

R4

O

OAc

H

H

R2

(+)-(2R,3S)-80: R1,R2,R4 = H, R3 = OMe (−)-(2R,3S)-81: R1,R3 = OMe, R2,R4 = H (+)-(2R,3S)-82: R1,R2,R3,R4 = OMe

OAc

5

R2

4 H

B

R5

OH

R1 (2R,3S)-77: R1,R2,R3,R4 = OH, R5 = H (+)-catechin R4 (2R,3S)-78: R1,R3,R4,R5 = OH, R2 = H (−)-rubinetinidol (2R,3S)-79: R1,R2,R4 = OH, R3,R5 = H (+)-afzelechin

HO 6

R1 R

6

A

1 H O C3 2

(+)-(2S,3S)-83: R1,R2,R4 = H, R3 = OMe (+)-(2S,3S)-84: R1,R3 = OMe, R2,R4 = H (+)-(2S,3S)-85: R1,R2,R3,R4 = OMe

O

HO 7

R1 7

3

OH (−)-(2R,4R)-76

(−)-(S)-flavan (S)-75 R3

H

H

R2

2 3

H4 H (−)−73: X = O 74: X = NH

R4

O

2

X

R2

1 H

O

O

2

(R)-86: R1,R2 = H (S)-87: R1= H,R2 = OMe (S)-88: R1 = Me,R2 = OMe R1

O C 3

4′

8′

O

8

CH3

(−)−(R)-89

A

2

OH H

R2

(2R,4′R,8′R)-90 δ-tocopherol H6

O 2 1 Me R3 (−)-(S)-91: R1,R2,R3 = H (+)-(2R,6S)-93 (+)-(S)-92: R1,R2,R3 = OMe H

B

Chart 3.10. Structures of chroman derivatives 73–93.

on the basis of ECD data of synthetic trans-3-acetoxyflavans (2R, 3S )-80–82. Cis-3acetoxyflavans (2S , 3S )-83–85 would apparently also follow the helicity rule if their heterorings had half-chair conformation with M -helicity and equatorial C2 aryl group. However, the small coupling constants of 2-H and 3-H [for (2S , 3S )-83, 3 J2,3 = 1.5 Hz, 3 J3,4 = 4.5 and 3.0 Hz) [114] suggest that the heteroring preferably adopts a boat or twist boat conformation with axial or pseudo-axial C2 aryl group. The comparison of the ECD data of (−)-(S )-75 with those published for (R)-7-hydroxy-flavan 86 [115], having M helicity and a positive 1 Lb band CE, revealed that 7-hydroxy substitution does not invert the original rule either [81]. The positive 1 Lb CE of (R)-86 served as a reference to determine the correct absolute configuration of flavans (S )-87 and (S )-88 (Figure 3.12a), although the possible effect of an additional C5 methoxy group was not considered [115]. The ECD data of (−)-(R)-89 [116] and δ-tocopherol [(2R, 4 R, 8 R)-90] [117] corroborate the chroman helicity rule and the case of δ-tocopherol also demonstrates that the presence of a C6 hydroxy substituent does not put a limit to the empirical rule (Figure3.12d,e). Reported ECD data of synthetic [118] and natural [120] isoflavans such as those of (−)-(S )-91 and (+)-(S )-92 would suggest that isoflavans surprisingly follow an inverse helicity rule with respect to flavans; P /M -helicity of the heteroring gives positive/negative 1 Lb CE [121]. For instance, the heteroring of synthetic derivative (+)(S )-92 has a half-chair conformation of M -helicity with an equatorial C3 aryl group (Figure 3.12f), and it shows an intense negative CE at 288 nm accompanied by weak

99

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

TAB L E 3.7. Helicity and ECD Data of Chroman Derivatives Compound (−)-73c 74c (−)-(S )-75 (−)-(2R, 4R)-76 (+)-(2R, 3S )-77 (−)-(2R, 3S )-78 (+)-(2R, 3S )-79 (+)-(2R, 3S )-80 (−)-(2R, 3S )-81 (+)-(2R, 3S )-82 (+)-(2S , 3S )-83 (+)-(2S , 3S )-84 (+)-(2S , 3S )-85 (R)-86 (S )-87 (S )-88 (−)-(R)-89 (2R, 4 R, 8 R)-90 (−)-(S )-91 (+)-(S )-92 (+)-(2R, 6S )-93

Helicity

CE {λ, nm (ε)a or [θ]b}

References

M M P M P P P P P P

277 (+1.83)a 308 (+3.14)a 276 (−1.01)a 276 (+1.28)a 282 (−0.36)a 284 (−1.50)a 270 (−0.44)a 272 (−6800)b 284 (−10000)b 279 (−5100)b 274 (+6800)b 273 (+2600)b 271 (+1100)b Positive 1 Lb CE Negative 1 Lb CE Negative 1 Lb CE 278 (+0.09), 270 (+0.06)a 298 (−0.33)a 282sh (−0.60), 275 (−0.63)a,f 300 (−78), 288 (−4780), 274 (+17)b 283sh (−0.57), 276 (−0.67)

81, 111 81, 111 81, 111 81, 112 113 113 113 114 114 114 114 114 114 115 115 115 116 117

e e e

M P P M P M M M

d

118 119

CE reported as ε. CE reported as θ . c with cholestane skeleton. d unpublished ECD data. e not determined, conformation of heteroring is different from half-chair. f1 La CE: 228 (−1.39). a b

transitions at higher and shorter wavelength (Table 3.7). This finding is contradictory to the expectations, since the preferred conformation of the heteroring does not deviate from the half-chair, there is no axial benzylic substituent, and the contribution of the third sphere is not considered significant. In order to explain this discrepancy, TDDFT calculations were carried out on the synthetic isoflavan (−)-(S )-91, and, for comparison, on flavan (−)-(S )-75. In both cases, DFT-optimized geometries (at B3LYP/6-31G(d) level) were employed as input structures, which showed the heteroring in the expected halfchair conformation with helicity depicted in Figure 3.12a,3.12f. Surprisingly, none of the low-energy transitions of isoflavan (S )-91 are of pure 1 Lb character as found for flavan (S )-75. On the contrary, three transitions predicted by B3LYP/TZVP calculations between 247 and 257 nm result from a combination of exciton-coupled excitations centered on the two aromatic rings, plus charge-transfer transitions. Since exciton-coupled interactions are determined by the relative orientation of the electric transition moments, which is in turn are influenced by the substitution pattern of ring A and B, the ECD of isoflavans with different substitutions cannot be compared easily and ECD calculations should be considered for a safe configurational assignment. On the basis of the above considerations, the determination of absolute configuration for some recently reported isoflavan derivatives 94–98 using the inverse chroman helicity rule of isoflavans [121] is possibly prone to error (Chart 3.11) [122–125]. This is especially true in the presence of a conjugated double bond ring substituent as in desmodin A [(−)-94] [122] and glabridin (96)

100

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H

(a) Ar O

H

OH

O Ph

(e)

OH

(g)

H

2

Ar

Me (2R,4′R,8′R)-90 P-helicity negative 1Lb CE

(R)-89 M-helicity positive 1Lb CE

H (2R, 3S)-3-hydroxyflavans 77-79 (2R, 3S)-3-acetoxyflavans 80-82 P-helicity negative 1Lb CE

(f) O R

O

Ar O OR

H H (2R,4R)-76 M-helicity positive 1Lb CE

(S)-flavans 75,87,88 P-helicity negative 1Lb CE (d)

H

(c)

(b)

6

O (S)-isoflavans 91,92 M-helicity negative 1Lb CE

Me O

(2R,6S)-93 M-helicity negative 1Lb CE

Figure 3.12a–g. Preferred low-energy conformations and helicity of chroman derivatives viewed from the direction of the fused aromatic ring.

O

O

O

O

Me

O

O

O

OMe

OH

Me OMe

OH

desmodin A [(−)-94]

OMe

OH

OH

desmodin B [(+)-95]

glabridin (96)

OMe

O

O

OH

HO OH isoflavan-4-ol (97)

(+)-98

Chart 3.11. Structures of isoflavans 94–97 and the homoisoflavan derivative (+)-98.

[123] or with an additional chiral dihydrobenzo[b]furan ring as in desmodin B [(+)-95] [122]. The bridged tricyclic tetrahydro-2,6-methano-2H -1-benzoxocine derivative (+)(2R, 6S )-93 also follows the inverse helicity rule; that is, its negative 1 Lb CE (Table 3.7) corresponds to a fixed half-chair conformation of M -helicity (Figure 3.12g) [119]. Due to the bridged structure, the C2–C3 and C6–C5 bonds are axially oriented and the contribution of the third sphere presumably overrides that of the second sphere (helicity of the heteroring), thus inverting the helicity rule. As a summary, the same helicity rule (P /M -helicity negative/positive 1 Lb CE) was found for the unsubstituted chroman chromophore as for the unsubstituted dihydrobenzo[b]furan, but, in contrast to the latter, methoxy or hydroxy ring substituents do not change the correlation. The chroman helicity rule can be safely utilized for the configurational assignment of flavans, cis-4-hydroxyflavans, trans-3-oxygenated flavans, and 2-alkylated chromans [81, 121], while the helicity of the heteroring is not decisive for the 1 Lb CE of isoflavan and bridged tetrahydro-2,6-methano-2H -1-benzoxocine derivatives.

101

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

3.4. TETRALONE, DIHYDROISOCOUMARIN, AND CHROMAN-4-ONE CHROMOPHORES In this section, the carbonyl derivatives of tetralin, isochroman, and chroman—that is, tetralones, dihydroisocoumarins, and chroman-4-ones—are discussed, in which besides the π –π * transition, the carbonyl n –π * transition is frequently used for determination of absolute configuration.

3.4.1. Tetralone Derivatives Tetralone (3,4-dihydronaphthalen-1(2H )-one) derivatives, containing both a tetraline and acetophenone chromophore, are widespread in nature [126–134], and their n –π * transition of the conjugating carbonyl group above 300 nm may be used for the determination of absolute configuration instead of the 1 Lb band, since its sign is expected to be independent from the substitution pattern of the aromatic ring [112]. The fused carbocyclic ring of tetralones most likely adopts an envelope conformation with C3 out of the plane of the benzene ring [127], the conformation (helicity) of which is expected to determine the sign of the n –π * CE as found also for natural flavanones and isoflavanones [121]. ECD data and preferred conformation of the carbocyclic ring of natural tetralones (Chart 3.12) are tabulated in Table 3.8. For tetralones 99–103 and 105, 106 having diverse substitution pattern, the M -helicity of the fused carbocyclic ring with envelope conformation results in positive n –π* CEs, although (2S , 4R)-100 had a very weak negative low-energy CE. In contrast, tetralones 104 and 107–110 show contradictory correlations; P -helicity of the heteroring affords positive n –π * CEs. Although absolute configurations were determined independently from ECD study by either Mosher’s NMR method or X-ray analysis for 99, 100, 103, 104, 108 and 109, some of these derivatives also show an inconsistent relationship between the helicity of the nonaromatic ring and the sign of the n –π * CE. This finding suggests that in contrast to the case of flavanones and isoflavanones [121], there is no straightforward general correlation between the sign of the n –π * CE and conformation of the nonaromatic ring. A possible explanation can be that conformers other than the represented envelope one as well as the position and nature of the substituents may play a non-negligible role, which renders the configurational assignment difficult if based only on the helicity of the nonaromatic ring or a simple comparison of the ECD

O

O

OH O OH

CH2 OH

H3C

O

CH3

OH

R1 R2 R1 R2 callianthone A [(2S,4R)-101]: R1= OH, R2= H (2S,4S)-99: R1= H, R2 = OH (2S,4R)-100: R1 = OH, R2 = H callianthone B [(2S,4S)-102]: R1= H, R2= OH O

O H3C

O

O

O OH (2R,4S)-104

hemiculone [(3S,4S,1′R)-103]

OH O

OH O Me

H3C

Me OH (3S,4S)-105

Me OH

OH

pyrolone A [(3S,4S)-106]

Me R1 R2 OH pyrolone B cis-isoshinanolone 1 2 [(2R,4S)-107] (3R,4R)-108: R = OH, R = H trans-isoshinanolone (3R,4S)-109: R1= H, R2= OH

O

Ar

OH

Chart 3.12. Structures of tetralone derivatives for Table 3.3.

(3R)-110 Ar: 3,4-dihydroxyphenyl

102

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.8. Conformation of the Carbocylic Ring with Orientation of the Substituents and Their n –π * CEs Low-Energy Conformera Helicityb CE {λ, nm (ε),c [θ],d θ e}

Compound (2S , 4S )-99

f

(2S , 4R)-100f

301 (+272)d

126

OH

M

341 (−285)d 304 (+520)

126

OH

M

324 (+5.29)e

127

H

M

345 (+1.68)e

127

M

301 (+1.5)c

128

P

321 (+4160)d

131

M

333 (+1200), 323 (+1070)d

131

M

338 (+2.4), 316 (+0.8)e

130

P

323 (+2.4)

130

P

331 nm positive CEi

129

P

positive CEi

129

M

320 (−2.2)a

133

H

O H

OH

HO O H

HO O H

CH3

(2S , 4S )-102

M

HO

H

(2S , 4R)-101

Reference

HO O

OH

CH3

(3S , 4S , 1 R)-103f

H

OH O

O

H

(2R, 4S )-104f

H

H

O

OH

Me

(3S , 4S )-105g

H

Me

O

H

OH

(3S , 4S )-106 H

(2R, 4S )-107

H

O

OH

Me

(3R, 4R)-108h

OH O OH

H

H

(3R, 4S )-109h

H O OH

OH

H H

(3R)-110

Ar O

a

Envelope conformer is in equilibrium with the distorted chair but their helicities are identical. Defined by the sign of the ωC5a,C4,C3,C2 torsional angle. c CE reported as ε. d CE reported as θ . e CE reported as θ ε. f Absolute configuration was determined by Mosher’s method. g The reported absolute configuration is probably wrong and should be reassigned as (3R, 4R)-105. h Determination of absolute configuration is based on X-ray diffraction analysis of a related derivative. i From HPLC-CD. b

103

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

spectra. In this case, the configurational assignment is especially prone to errors and a thorough conformational analysis and ECD calculation cannot be spared. As an example to demonstrate what was stated, we analyzed the conformation and ECD spectra of compounds 100, 104, 105, and 109 with molecular modeling (MMFF conformational search and DFT optimization at B3LYP/6-31G (d) level) and TDDFT ECD calculations. For tetralones 100, 104, and 109, the substituents of the carbocyclic ring adopt a preferred equatorial position in the lowest-energy conformer (populations above 85% at 300 K) which constrains the ring in a more or less fixed envelope conformation with well-defined helicity (Table 3.9). The result of ECD calculations run with TDDFT method, B3LYP/TZVP level, on the three compounds are also shown in Table 3.9. Apparently, ECD calculations reproduce experimental data for 100, 104, and 109, apart from a wavelength shift, and confirm the assigned absolute configuration. However, at least in the case of tetralone (2S , 4R)-100 the assignment based on the n –π *

TAB L E 3.9. Summary of Geometry Optimizationsa and CD Calculationsb for Selected Tetralones Compound (2S , 4R)-100

Lowest-Energy Conformerc (Population)c

Structure OH O

1L

Helicity

n –π * CE λ (sign)

b CE λ (sign)

M

285 (+)

300 (−)

P

317 (+)

261 (−)

M

331 (−)f

273 (−)f

P

298 (+)a

295 (−)a

OH

HO

OH

O H

(86%)d

OH

(2R, 4S )-104

H

O

H

H

O

OH

Me (89%)d

OH

(3S , 4S )-105c

O

H

Me

O

H

OH (85%)e

OH

(3R, 4S )-109

OH

O

H O OH OH a With

OH

H (87%)d

B3LYP/6-31G(d) on MMFF-calculated low-energy minima. B3LYP/TZVP. c Optimizations with B3LYP/6-311+G(d,p) in methanol (IEF-PCM); TDDFT calculations with B3LYP/TZVP in methanol. d At 300 K, using internal energies. e Using free energies; includes two 4-OH rotamers. f Boltzmann-weighted average for four low-energy structures. The result for the absolute lowest-energy minimum was similar. b With

104

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

helicity rule must be considered fortuitous. In fact, the relative position of aromatic 1 Lb and carbonyl n –π * bands varies for the three compounds due to to the presence of OH groups hydrogen-bonded to the carbonyl. On passing from 104 (unbound C = O) to 109 (one C = O· · ·HO bond) and to 100 (two C = O· · ·HO bonds), the n –π * transition experiences the expected hypsochromic effect moving from 317 nm (104) to 298 nm (109) and to 285 nm (100). In this latter case the first calculated transition (corresponding to the negative CE observed at 341 nm) turns out to be the 1 Lb and not the n –π * one, while for tetralone 109 the two transitions are almost degenerate (and heavily mixed). Moreover, the calculated rotational strength for the n –π * transition for (2S , 4R)-100 is positive, in contrast with the M -helicity of the carbocylic ring. The conformational situation for tetralone (3S , 4S )-105 is less clear-cut because the 3-methyl and 4-hydroxy substituents are not allowed to occupy simultaneously an equatorial position. On the basis of relative conformational energies [135], the methyl group is expected to dictate the conformation of the ring, showing a stronger preference for the equatorial position than the hydroxyl group. This is confirmed by NMR experiments [131] and by our calculations. In this case, we run B3LYP geometry optimizations using a larger basis set (6-311+G(d , p)) and including a solvent model (IEF-PCM) for methanol, followed by frequency calculations to estimate true free energies. The two lowest-energy conformers show an equatorial 3-methyl group (they differ for the rotation of 4-OH) and amount to an overall population of 85%. They are followed by two other minima (overall 15% population) with axial 3-methyl and equatorial 4-OH group. ECD calculations were run with B3LYP/TZVP including again IEF-PCM for methanol and considering all four minima. Very interestingly, the average TDDFT-calculated ECD spectrum for (3S , 4S )-105 (which is dominated by the lowest-energy structure at long wavelengths) shows negative CEs for both n –π * and 1 Lb , in contrast with the experimental data (compare Tables 3.8 and 3.9). According to this outcome, the reported absolute configuration for (3S , 4S )-105 is wrong and should be reassigned as [(+)-ECD(333)]-(3R, 4R)-105. The above calculation results demonstrate that the observed n –π * CEs for tetralones depend heavily on the substitution pattern of both the aromatic and carbocyclic rings; therefore we discourage the use of the relative helicity rule for configurational assignments.

3.4.2. Dihydroisocoumarin Chromophore Optically active synthetic isochromans (S )-45a,d,c (Scheme 3.1) were converted to the corresponding dihydroisocoumarins (S )-111a–c by oxidation with Jones reagent [136] or dimethyldioxirane (DMDO) [137] as shown in Scheme 3.3 [75]. The (S )-dihydroisocoumarins 111a–c have very similar ECD patterns; positive π –π * and n –π * transitions at 278–307 nm and 252–268 nm, respectively, followed by a negative and positive band in the high-energy region (Table 3.10). Their ECD TAB L E 3.10. ECD Data for Dihydroisocoumarins (S )-111a–c Compound (S )-111a (S )-111b (S )-111c

CD λmax [nm] (ε) π → π ∗ : 289sh (+1.08), 278sh (+2.00); n → π ∗ : 252 (+4.19); 230 (−4.73), 204 (+13.83). π → π ∗ : 307sh (+1.10), 300 (+1.28), 294sh (+1.20); n → π ∗ : 268 (+7.62); 244 (−4.64), 226 (+10.48), 204 (−6.22). π → π ∗ : 304sh (+2.44), 296 (+2.48); n → π ∗ : 258 (+4.17); 239 (−1.29), 229sh (+0.67), 206 (+11.21).

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

data confirmed that regardless of the substitution pattern of the aromatic ring, the positive n –π * transition of 3-alkyldihydroisocoumarins derives from P -helicity of the heteroring and thus (S ) absolute configuration, in accordance with previous ECD studies on synthetic steroidal dihydroisocoumarin derivatives of 41–43 [74] (Figure 3.6a). The heteroring of (S )-dihydroisocoumarins 111a–c adopts a half-chair or envelope conformation with P -helicity, as defined by the positive ωC-5a,C-4,C-3,O torsional angle, to ensure the favorable equatorial orientation of the C3 methyl group (Scheme 3.3). The dihydroisocoumarin n –π * transition was consistently applied to the configurational assignment of synthetic derivatives [138] and natural products [139–144] in agreement with the above helicity rule. In order to check the applicability of the dihydroisocoumarin helicity rule for 3,4disubstituted dihydroisocoumarins, the 3,4-cis-dimethyl-dihydroisocoumarin derivative 113 was prepared by catalytic hydrogenation of (S )-(+)-ascochin (112), a natural product isolated from the endophytic fungus Ascochyta sp. (Scheme 3.4), and its chiroptical data were measured and reproduced by calculation [145]. The addition occurred with cis diastereoselectivity due to the inherent (4S ) chirality center (the formyl group was also reduced during the hydrogenation), resulting in (3S , 4S ) absolute configuration for 113. The ECD spectrum of 113 shows a negative CE at 307 nm and a positive one at 267 nm (Figure 3.13). Thus, according to the literature [74, 146], the lactone n –π * CD band should be assigned to the latter band. Interestingly, the synthetic (3S )-3-methyldihydroisocoumarin derivative 111b (Scheme 3.3) has also a positive CE at 268 nm [75], and except for this transition, its ECD curve was almost the mirror image of that of 113 (Figure 3.13). In order to confirm the position of the lactone n –π * ECD transition and hence the semiempirical rule of dihydroisocoumarins, a TDDFT calculation was carried out

(S)-45a Jones reagent (S)-45c (S)-45d or DMDO/ dry acetone

H

H 5a 4 Me 2 3 1 O

R3 R2

R1

1

4

O

O

R2 H OMe H

3

Me

R3 H OMe H

O

2

positive n-π* CE in the 252 – 268 - nm range

H

(S)-111a-c R1 111a H 111b H 111c OMe

3

Me

1

4

O

ωC-5a,C-4,C-3,O >0 P-helicity

Scheme 3.3.

O

11 6 5

HO 7

HO 11 10

10 4

9 3

1 O

HO 6 H2-Pd/C 7

8

OH O (S)-(+)-ascochin (112)

5 4 8

H 9

3

1 O

OH O (3S,4S)-tetrahydroascochin (113)

Me H4 Me

3

1

O

O

2

P-helicity ωC-8a, C-1, O, C-3 > 0

Scheme 3.4. Conversion of (S)-(+)-ascochin (112) to 113. P-helicity of 113 and DFT-calculated most stable conformation.

105

106

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

12 9 n→ π*

Δε and R [10–40 cgs]

6 3

(S)-111b

0

×10 113

–3 –6 –9 –12

200

220

240

260

280

300

320

340

λ (nm)

Figure 3.13. Measured ECD spectra of 113 (solid line) and (S)-111b (dotted line), and TDDFTcomputed rotational strengths R (vertical bars) for the absolute minimum of (3S, 4S)-113 found by DFT (Scheme 3.4).

on the absolute minimum-energy structure computed for 113 [145]. The conformational analysis of 113 resulted in a hydrogen-bonded (O-11 → HO-7) conformer as the most stable structure (1.7 kcal/mol lower DFT-energy than the second minimum), as shown in Scheme 3.4. The TDDFT CD calculation on this conformer demonstrated that the lactone n –π * CD transition of 113 appears with positive CE as the third computed transition from the red at 246 nm, namely as part of the 250 to 270-nm ECD band, overlapped with aromatic π –π * transitions (Figure 3.13). In particular, the most redshifted transition computed at 288 nm is of the 1 Lb -type, and it is apparently responsible for the weak ECD signal above 280 nm. Since 113 has (3S , 4S ) absolute configuration and P -helicity of the heteroring (Scheme 3.4), its computed positive n –π * transition is in accordance with the semiempirical rule. (S )-111b has again P -helicity, also resulting in positive n –π * CE at 268, although all the other corresponding transitions have opposite signs to those of 113, due to the different substitution pattern of the aromatic ring. Since the lactone n –π * transition is only one of the contributors to the 267-nm ECD band among several π –π * transitions, its application for a safe configurational assignment is endangered in the current case by overlapping transitions. Another example for the combined use of the dihydroisocoumarine helicity rule and TDDFT calculations is offered by natural products phomolactone A (114) and B (115), isolated from Phomopsis sp. (Chart 3.13) [147]. The n –π * CE of 114 at 262 nm is

R1 R2 6

5

5a

4

R4 3

7 3

R

8

O2

8a

OH

1

O

R1 R2 R3 114 115 116 117

OH Cl OH H OH Cl OH H

R4

H n-Pr H n-Pr H Me H Me

n-π* CE 262 (−5.84) 258 (−2.08)

Chart 3.13.

107

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

8 6

Δε [cm2 mmol–1]

4 2 0 –2 Experimental CD of 113 Experimental CD of 114 Calculated CD of 116

–4 –6 200

220

240

260

280

300

320

340

360

λ (nm)

Figure 3.14. Measured (MeCN) ECD spectrum of 114 (solid line) compared with TDDFT-calculated ECD of (R)-116 (gray dotted line) and experimental ECD of 113 (dashed–dotted line).

negative and its high-energy ECD transitions are also opposite to those of (3S , 4S )tetrahydroascochin (113), which suggests that 114 has a heteroring with M helicity and hence (3R) absolute configuration (Figure 3.14). Dihydroisocoumarin 115 showed the same −/+/–/+ ECD pattern from the low-energy to the high-energy region as 114, which allowed its assignment as (3R) as well. TDDFT ECD calculations were employed to confirm the configurational assignment and support the ECD correlation discussed above. The calculations confirmed that neither the length of the alkyl chain nor its conformation affects the shape of ECD bands allied with the dihydroisocoumarin chromophore. Their sign is entirely determined by the ring A chirality, which is in turn dictated by the absolute configuration of C3. This finding corroborated the ECD correlation for n –π * CE of dihydroisocoumarin discussed above, and also simplified the treatment of compounds 114 and 115, instead of which we could consider the methyl analogues 116 and 117 (Chart 3.13). Figure 3.14 shows the TDDFT-calculated ECD spectrum (B3LYP/TZVP) as Boltzmann-weighted average over two DFT-optimized geometries for (R)-116, in a good agreement with the experimental spectrum of 114 below 300 nm, which confirms the absolute configuration established above as (R)-114.

3.4.3. Chroman-4-one Chromophore The chroman-4-one chromophore is found in natural flavanones, 3-hydroxyflavanones, 2-alkylchroman-4-ones, and isoflavanones exemplified by compounds 118–121 (Figure 3.15). Snatzke established a relationship between the chirality of cyclic aryl ketones and their high-wavelength n –π * CEs [148] which was extended to correlate the helicity of the heteroring and the sign of the n –π* CE in flavanones [149], 3-hydroxyflavanones [149], 2-alkylchromanones [150], and isoflavanones (Figure 3.15) [121, 151]. According to this rule, P -helicity of the heteroring adopting envelope conformation is manifested in a positive n –π * CE above 300 nm, such as in (S )-flavanone, (2R, 3R)-3-hydroxyflavanone, (R)-2-methylchroman-4-one and (R)-isoflavanone.

108

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1

O 4

H

H

1

O

2 3

4

R

O (S)-flavanone [(S)-118]: R = H (2R,3R)-3-hydroxyflavanone [(2R,3R)-119: R = OH

1

Me

O

2 3

4

2 3

O (R)-isoflavanone [(R)-121]

O (R)-2-methylchroman-4-one [(R)-120] H

H R

Ph O

Ph

Me O

O

O

O

O

H

H envelope conformation P-helicity positive n-π* CE

envelope conformation P-helicity positive n-π* CE

envelope conformation P-helicity positive n-π* CE

Figure 3.15. Correlation between the helicity of the heteroring in (S)-flavanone, (2R, 3R)-3hydroxyflavanone, (R)-2-methylchroman-4-one, and (R)-isoflavanone and the sign of the n–π * CE.

1

O

1

O

O

8a

2 4 5

OR

Me O

(R)-122: R = H (R)-123: R = Me

Br

C

3

3

4 2

O

H ωC8a,O1,C2,C3 > 0 P-helicity positive n→π* negative 1Lb (π→π*)

O

O

O 124

Chart 3.14. Structures of chromanones 122–124 and preferred conformation of the heteroring for (R)-122 and (R)-123.

A recent application of the chromanone n –π * helicity rule is demonstrated by the 2-methyl-chroman-4-one derivatives 122 and 123 isolated from the endophytic fungus Nodulisporium sp. with 6% enantiomeric excess (Chart 3.14) [152]. The separations of their enantiomers were carried out with HPLC using a chiral stationary phase; and then their LC/CD spectra were recorded on-line, which allowed their configurational assignment on the basis of their long-wavelength n –π * CEs (Figure 3.16). Since the first-eluted enantiomers of both 122 and 123 have positive n –π * and negative π –π * CEs around 340 and 310 nm, respectively, their heterorings adopt P -helicity, which implies (R) absolute configuration, provided that the methyl group is equatorial ly oriented [152]. In fact, the X-ray data of the p-bromobenzoate of 122 (compound 124) showed that the chromanone heteroring has an envelope conformation with torsion angles ωC8a,O1,C2,C3 44.9(7)◦ and ωC5,C4a,C4,O − 1.3(7)◦ . The equatorial position of the methyl group is further in agreement with the large coupling constant of J = 12.8 Hz observed for the transdiaxial protons 2-Hax and 3-Hax . It must be stressed that similarly to tetralone derivatives, intramolecular hydrogen bonding of the carbonyl group may cause considerable blue shift of the characteristic n –π * transition, and thus the unambiguous assignment of the n –π *

109

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

9 6

(S)-122

n π* (R)-122

π π* 3 0 Δε

–3 (S)-123

–6 –9 –12

240

270

300

330

360

390

Wavelength (nm)

Figure 3.16. LC/CD spectra of (S)-122 (solid line), (R)-122 (dotted line), and (S)-123 (dashed line) in hexane/isopropanol 9:1.

and π –π * transitions may require ECD measurements in solvents of different polarity or excited-state calculations.

3.5. CONCLUSION The large amount of ECD data compiled on natural and synthetic derivatives containing a fused benzene chromophore makes it often possible to compare the measured ECD data of a new compound with those of analogues with known absolute configuration, allowing a fast determination of absolute configuration. The present chapter aimed to give guidelines for the scope and limitation of these correlations by discussing the safe applications and pitfalls of semiempirical helicity rules on benzene derivatives. As demonstrated by some examples, the application of high-level quantum-mechanics calculations may be extremely useful, especially in ambiguous cases, to explain the background and limitation of these semiempirical rules.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

L. Doub, J. M. Vandenbelt, J. Am. Chem. Soc. 1947, 69 , 2714–2723. H. B. Klevens, J. R. Platt, J. Chem. Phys. 1949, 17 , 470–484. J. R. Platt, H. B. Klevens, Chem. Rev . 1947, 41 , 301–310. J. Horwitz, E. H. Strickland, C. Billups, J. Am. Chem. Soc. 1969, 91 , 184–190. H. E. Smith, T. C. Willis, J. Am. Chem. Soc. 1971, 93 , 2282–2290. N. Berova, B. Kurtev, G. Snatzke, Tetrahedron 1983, 39 , 1371–1378. N. Berova, B. Kurtev, G. Snatzke, Croat. Chem. Acta 1987, 60 , 243–262. J. Petruska, J. Chem. Phys. 1961, 34 , 1120–1136.

110

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

44. 45.

L. P. Fontana, H. E. Smith, J. Org. Chem. 1987, 52 , 3386–3389. S. T. Pickard, H. E. Smith, J. Am. Chem. Soc. 1990, 112 , 5741–5747. L. A. Mitscher, F. Kautz, J. LaPidus, Can. J. Chem. 1969, 47 , 1957–1963. R. J. Lorentzen, J. H. Brewster, H. E. Smith, J. Am. Chem. Soc. 1992, 114 , 2181–2187. H. E. Smith, L. P. Fontana, J. Org. Chem. 1991, 56 , 432–435. H. E. Smith, J. R. Neergaard, J. Am. Chem. Soc. 1996, 118 , 7694–7701. H. E. Smith, J. R. Neergaard, J. Am. Chem. Soc. 1997, 119 , 116–124. D. R. Michals, H. E. Smith, Chirality 1993, 5 , 20–23. H. E. Smith, Chem. Rev . 1998, 98 , 1709–1740. O. Korver, S. Dejong, T. C. Vansoest, Tetrahedron 1976, 32 , 1225–1229. A. O. Patil, W. T. Pennington, I. C. Paul, D. Y. Curtin, C. E. Dykstra, J. Am. Chem. Soc. 1987, 109 , 1529–1535. T. Schaefer, G. H. Penner, R. Sebastian, Can. J. Chem. 1987, 65 , 873–877. U. Berg, T. Liljefors, C. Roussel, J. Sandstrom, Acc. Chem. Res. 1985, 18 , 80–86. J. L. Broeker, R. W. Hoffmann, K. N. Houk, J. Am. Chem. Soc. 1991, 113 , 5006–5017. G. Pescitelli, L. Di Bari, A. M. Caporusso, P. Salvadori, Chirality 2008, 20 , 393–399. P. Butz, G. E. Tranter, J. P. Simons, Phys. Chem. Comm. 2002, 5 , 91–93. J. R. Platt, J. Chem. Phys. 1951, 19 , 263–271. M. Dierksen, S. Grimme, J. Chem. Phys. 2006, 124 . N. Lin, F. Santoro, X. Zhao, A. Rizzo, V. Barone, J. Phys. Chem. A 2008, 112 , 12401–12411. N. A. Macleod, P. Butz, J. P. Simons, G. H. Grant, C. M. Baker, G. E. Tranter, Isr. J. Chem. 2004, 44 , 27–36. M. D. Kundrat, J. Autschbach, J. Chem. Theory Comput. 2009, 5 , 1051–1060. G. Snatzke, Angew. Chem. Int. Ed . 1979, 18 , 363–377. G. Snatzke, M. Kajt´ar, F. Snatzke. Relationship of chiroptical properties with structure. I. Organic compounds. Aromatic chromophores, in Fundamental Aspects and Recent Development in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli and P. Salvadori, eds. Hyden & Sons, London, 1973, pp. 148–172. G. Snatzke, P. C. Ho, Tetrahedron 1971, 27 , 3645–3653. N. J. Luche, A. Marquet, G. Snatzke, Tetrahedron 1972, 28 , 1677–1682. E. Dorhege, G. Snatzke, Tetrahedron 1970, 26 , 3059–3067. J. Schulte, G. Snatzke, Chem. Ber. 1989, 122 , 1373–1374. S. Hagishita, K. Kuriyama, Bull. Chem. Soc. Jpn. 1982, 55 , 3216–3224. J. Toda, S. Matsumoto, T. Saitoh, T. Sano, Chem. Pharm. Bull . 2000, 48 , 91–98. J. C. Craig, S. Y. C. Lee, R. P. K. Chan, I. Y. F. Wang, J. Am. Chem. Soc. 1977, 99 , 7996–8002. G. Snatzke, M. Kajt´ar, F. Werner-Zamojska, Tetrahedron 1972, 28 , 281–288. J. Sagiv, Tetrahedron 1977, 33 , 2303–2313. J. Sagiv, Tetrahedron 1977, 33 , 2315–2320. S. Kloubert, M. Mathe-Allainmat, J. Andrieux, S. Sicsic, M. Langlois, Bioorg. Med. Chem. Lett. 1998, 8 , 3325–3330. S. Yous, S. Durieux-Poissonnier, E. Lipka-Belloli, H. Guelzim, C. Bochu, V. Audinot, J. A. Boutin, P. Delagrange, C. Bennejean, P. Renard, D. Lesieur, Bioorg. Med. Chem. 2003, 11 , 753–759. A. C. Trillat, M. Mathe-Allainmat, B. Bremont, I. Malagie, C. Jacquot, A. M. Gardier, M. Langlois, Eur. J. Med. Chem. 1998, 33 , 437–444. A. P´eter, G. T´oth, E. Olajos, F. F¨ul¨op, D. Tourwe, J. Chromatogr. A 1995, 705 , 257–265.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

46. A. P´eter, G. T¨or¨ok, D. W. Armstrong, G. T´oth, D. Tourwe, J. Chromatogr. A 2000, 904 , 1–15. 47. T. Nakanishi, Y. Inatomi, H. Murata, K. Shigeta, N. Iida, A. Inada, J. Murata, M. A. P. Farrera, M. Iinuma, T. Tanaka, S. Tajima, N. Oku, Chem. Pharm. Bull . 2005, 53 , 229–231. 48. P. B. Hulbert, W. Klyne, P. M. Scopes, J. Chem. Res., Synop. 1981, 27–27. 49. M. Klaes, T. Ellendorff, T. J. Schmidt, Planta Med . 2010, 76 , 719–U724. 50. C. Zhao, A. Nagatsu, K. Hatano, N. Shirai, S. Kato, Y. Ogihara, Chem. Pharm. Bull . 2003, 51 , 255–261. 51. S. Antus, E. Baitz-G´acs, G. Snatzke, T. S. T´oth, Liebigs Ann. Chem. 1991, 633–641. 52. Z. Samu, S. Nyiredy, E. Baitz-G´acs, Z. Varga, T. Kurt´an, Z. Dinya, S. Antus, Chem. Biodivers. 2004, 1 , 1668–1677. 53. I. Szilagi, P. Tetenyi, S. Antus, O. Seligmann, V. M. Chari, M. Seitz, H. Wagner, Planta Med . 1981, 43 , 121–127. 54. T. H. Kim, H. Ito, K. Hayashi, T. Hasegawa, T. Machiguchi, T. Yoshida, Chem. Pharm. Bull . 2005, 53 , 641–644. 55. Y. W. Chin, J. W. Kim, Tetrahedron Letters 2004, 45 , 339–341. 56. N. Noda, S. Kubota, Y. Miyata, K. Miyahara, Chem. Pharm. Bull . 2000, 48 , 1749–1752. 57. M. Z. Xu, W. S. Lee, J. M. Han, H. W. Oh, D. S. Park, G. R. Tian, T. S. Jeong, H. Y. Park, Bioorg. Med. Chem. 2006, 14 , 7826–7834. 58. Z. Varga, I. Seres, E. Nagy, L. Ujhelyi, G. Balla, J. Balla, S. Antus, Phytomedicine 2006, 13 , 85–93. 59. P. M´atyus, K. Horv´ath, Med. Res. Rev . 1997, 17 , 523–535. 60. Z. Czak´o, L. Juh´asz, A. Ken´ez, K. Czifr´ak, L. Soms´ak, T. Docsa, P. Gergely, S. Antus, Bioorg. Med. Chem. 2009, 17 , 6738–6741. 61. K. K´onya, R. Ferenczi, A. Czompa, A. Kiss-Szikszai, T. Kurt´an, S. Antus, Arkivoc 2008, 200–210. 62. A. Czompa, Z. Dinya, S. Antus, Z. Varga, Arch. Pharm. 2000, 333 , 175–180. 63. A. Arnoldi, L. Merlini, J. Chem. Soc. Perkin Trans. 1 1985, 2555–2557. 64. L. Han, X. S. Huang, I. Sattler, U. Moellmann, H. Z. Fu, W. H. Lin, S. Grabley, Planta Med . 2005, 71 , 160–164. 65. N. Tabata, H. Tomoda, Y. Iwai, S. Omura, J. Antibiot. 1995, 48 , 83–84. 66. T. Inagaki, K. Kaneda, Y. Suzuki, H. Hirai, E. Nomura, T. Sakakibara, Y. Yamauchi, L. H. Huang, M. Norcia, L. M. Wondrack, J. A. Sutcliffe, N. Kojima, J. Antibiot. 1998, 51 , 112–116. 67. J. Malmstrom, C. Christophersen, J. C. Frisvad, Phytochemistry 2000, 54 , 301–309. 68. M. D. Ennis, N. B. Ghazal, R. L. Hoffman, M. W. Smith, S. K. Schlachter, C. F. Lawson, W. Bin Im, J. F. Pregenzer, K. A. Svensson, R. A. Lewis, E. D. Hall, D. M. Sutter, L. T. Harris, R. B. McCall, J. Med. Chem. 1998, 41 , 2180–2183. 69. I. Bouchelet, B. Case, A. Olivier, E. Hamel, Br. J. Pharmacol . 2000, 129 , 501–508. 70. A. Bhattacharya, K. W. Schenck, Y. C. Xu, L. Nisenbaum, E. Galbreath, M. L. Cohen, J. Pharmacol. Exp. Ther. 2004, 309 , 825–832. 71. M. P. Deninno, R. Schoenleber, K. E. Asin, R. Mackenzie, J. W. Kebabian, J. Med. Chem. 1990, 33 , 2948–2950. 72. M. P. Deninno, R. J. Perner, H. E. Morton, S. Didomenico, J. Org. Chem. 1992, 57 , 7115–7118. 73. R. E. TenBrink, C. L. Bergh, J. N. Duncan, D. W. Harris, R. M. Huff, R. A. Lahti, C. F. Lawson, B. S. Lutzke, I. J. Martin, S. A. Rees, S. K. Schlachter, J. C. Sih, M. W. Smith, J. Med. Chem. 1996, 39 , 2435–2437. 74. S. Antus, G. Snatzke, I. Steinke, Liebigs Ann. Chem. 1983, 2247–2261.

111

112

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

75. G. Kerti, T. Kurt´an, T. Z. Illy´es, K. E. K¨ov´er, S. S´olyom, G. Pescitelli, N. Fujioka, N. Berova, S. Antus, Eur. J. Org. Chem. 2007, 296–305. 76. T. Kurt´an, N. Nesnas, Y. Q. Li, X. F. Huang, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2001, 123 , 5962–5973. 77. J. M. Seco, E. Quinoa, R. Riguera, Tetrahedron-Asymmetry 2001, 12 , 2915–2925. 78. K. Inoue, Y. Makino, N. Itoh, Tetrahedron-Asymmetry 2005, 16 , 2539–2549. 79. I. Kock, S. Draeger, B. Schulz, B. Elsasser, T. Kurt´an, A. Ken´ez, S. Antus, G. Pescitelli, P. Salvadori, J. B. Speakman, J. Rheinheimer, K. Krohn, Eur. J. Org. Chem. 2009, 1427–1434. 80. T. Kurt´an, E. Baitz-G´acs, Z. M´ajer, S. Antus, J. Chem. Soc., Perkin Trans. 1 2000, 453–461. 81. S. Antus, T. Kurt´an, L. Juh´asz, L. Kiss, M. Holl´osi, Z. M´ajer, Chirality 2001, 13 , 493–506. 82. L. Juh´asz, L. Szil´agyi, S. Antus, J. Visy, F. Zsila, M. Simonyi, Tetrahedron 2002, 58 , 4261–4265. 83. L. Juh´asz, J. Visy, M. Simonyi, K. Krohn, S. Antus, Tetrahedron: Asymmetry 2002, 13 , 1219–1222. 84. H. Achenbach, J. Gross, X. A. Dominguez, G. Cano, J. V. Star, L. D. Brussolo, G. Munoz, F. Salgado, L. Lopez, Phytochemistry 1987, 26 , 1159–1166. 85. P. J. C. Benevides, P. Sartorelli, M. J. Kato, Phytochemistry 1999, 52 , 339–343. 86. Y. Natori, H. Tsutsui, N. Sato, S. Nakamura, H. Nambu, M. Shiro, S. Hashimoto, J. Org. Chem. 2009, 74 , 4418–4421. 87. D. C. Chauret, C. B. Bernard, J. T. Arnason, T. Durst, H. G. Krishnamurty, P. SanchezVindas, N. Moreno, L. SanRoman, L. Poveda, J. Nat. Prod . 1996, 59 , 152–155. 88. P. S. Luize, T. Ueda-Nakamura, B. P. Dias, D. A. G. Cortez, C. V. Nakamura, Biol. Pharm. Bull . 2006, 29 , 2126–2130. 89. M. P. De Campos, V. Cechinel, R. Z. Da Silva, R. A. Yunes, S. Zacchino, S. Juarez, R. C. Cruz, A. B. Cruz, Biol. Pharm. Bull . 2005, 28 , 1527–1530. 90. D. L. J. Clive, E. J. L. Stoffman, Chem. Commun. 2007, 2151–2153. 91. D. L. J. Clive, E. J. L. Stoffman, Org. Biomol. Chem. 2008, 6 , 1831–1842. 92. M. S. M. Yuen, F. Xue, T. C. W. Mak, H. N. C. Wong, Tetrahedron 1998, 54 , 12429–12444. 93. K. Yoshikawa, H. Kinoshita, Y. Kan, S. Arihara, Chem. Pharm. Bull . 1995, 43 , 578–581. 94. H. Ishii, F. Sekiguchi, T. Ishikawa, Tetrahedron 1981, 37 , 285–290. 95. H. Ishii, T. Ishikawa, I. S. Chen, S. T. Lu, Tetrahedron Lett. 1982, 23 , 4345–4348. 96. O. R. Gottlieb, J. C. Mourao, M. Yoshida, Y. P. Mascarenhas, M. Rodrigues, R. D. Rosenstein, K. Tomita, Phytochemistry 1977, 16 , 1003–1006. 97. C. J. Aiba, R. G. C. Correa, O. R. Gottlieb, Phytochemistry 1973, 12 , 1163–1164. 98. I. R. Nascimento, L. M. X. Lopes, L. B. Davin, N. G. Lewis, Tetrahedron 2000, 56 , 9181–9193. 99. C. J. Aiba, J. B. Fernandes, O. R. Gottlieb, J. G. Soaresmaia, Phytochemistry 1975, 14 , 1597–1604. 100. H. Achenbach, W. Utz, A. Usubillaga, H. A. Rodriguez, Phytochemistry 1991, 30 , 3753–3757. 101. F. Li, X. W. Yang, Helv. Chim. Acta 2007, 90 , 1491–1496. 102. I. R. Nascimento, L. M. X. Lopes, Phytochemistry 1999, 52 , 345–350. 103. H. Achenbach, W. Utz, B. Lozano, E. M. G. Touche, S. Moreno, Phytochemistry 1996, 43 , 1093–1095. 104. L. Duan, H. W. Tao, X. J. Hao, Q. Q. Gu, W. M. Zhu, Planta Med . 2009, 75 , 1241–1245. 105. L. Fang, D. Du, G. Z. Ding, Y. K. Si, S. S. Yu, Y. Liu, W. J. Wang, S. G. Ma, S. Xu, J. Qu, J. M. Wang, Y. X. Liu, J. Nat. Prod . 73 , 818–824.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

106. S. Garcia-Munoz, M. Alvarez-Corral, L. Jimenez-Gonzalez, C. Lopez-Sanchez, A. Rosales, M. Munoz-Dorado, I. Rodriguez-Garcia, Tetrahedron 2006, 62 , 12182–12190. 107. S. M. O. Van Dyck, G. L. F. Lemiere, T. H. M. Jonckers, R. Dommisse, L. Pieters, V. Buss, Tetrahedron-Asymmetry 2001, 12 , 785–789. 108. H. Dong, C. Wang, Y. Li, S. Guo, J. Yang, Magn. Reson. Chem. 2010, 48 , 256–260. 109. F. Martin, A. E. Hay, V. R. Q. Condoretty, D. Cressend, M. Reist, M. P. Gupta, P. A. Carrupt, K. Hostettmann, J. Nat. Prod . 2009, 72 , 852–856. 110. J. G. Sena, S. L. Nimmo, H. S. Xavier, J. M. Barbosa, R. H. Cichewicz, J. Nat. Prod . 2009, 72 , 1344–1347. 111. S. Antus, E. Baitz-G´acs, J. Kajt´ar, G. Snatzke, A. L. Tok´es, Liebigs Ann. Chem. 1994, 497–502. 112. G. Snatzke, F. Znatzke, A. L. Tok´es, M. R´akosi, R. Bogn´ar, Tetrahedron 1973, 29 , 909–912. 113. O. Korver, C. K. Wilkins, Tetrahedron 1971, 27 , 5459–5465. 114. H. vanRensburg, P. S. vanHeerden, D. Ferreira, J. Chem. Soc. Perkin Trans. 1 1997, 3415–3421. 115. G. Cardillo, L. Merlini, G. Nasini, P. Salvadori, J. Chem. Soc. (C) 1971, 3967–3970. 116. A. Czompa, T. Kov´acs, S. Antus, J. Het. Chem. 2000, 37 , 991–995. 117. F. Mazzini, G. Pescitelli, L. Di Bari, T. Netscher, P. Salvadori, Chirality 2009, 21 , 35–43. 118. M. Versteeg, B. C. B. Bezuidenhoudt, D. Ferreira, Tetrahedron 1999, 55 , 3365–3376. 119. T. Kurt´an, E. Baitz-G´acs, Z. M´ajer, A. B´enyei, S. Antus, J. Chem. Soc., Perkin Trans. 1 2002, 888–894. 120. L. Verbit, J. W. Clark-Lewis, Tetrahedron 1968, 24 , 5519–5527. 121. D. Slade, D. Ferreira, J. P. J. Marais, Phytochemistry 2005, 66 , 2177–2215. 122. C. Lima, G. Zappia, G. Delle Monache, B. Botta, Eur. J. Org. Chem. 2006, 5445–5448. 123. M. Kim, S. U. Kim, Y. U. Kim, J. Han, Bull. Korean Chem. Soc. 2009, 30 , 415–418. 124. D. Won, B. K. Shin, S. Kang, H. G. Hur, M. Kim, J. Han, Bioorg. Med. Chem. Lett. 2008, 18 , 1952–1957. 125. S. Awale, T. Miyamoto, T. Z. Linn, F. Li, N. N. Win, Y. Tezuka, H. Esumi, S. Kadota, J. Nat. Prod . 2009, 72 , 1631–1636. 126. E. Couche, A. Fkyerat, R. Tabacchi, Helv. Chim. Acta 2003, 86 , 210–221. 127. Y. L. Chen, C. H. Tan, J. J. Tan, S. J. Qu, S. H. Jiang, D. Y. Zhu, Planta Med . 2008, 74 , 1826–1828. 128. D. Laurent, G. Guella, I. Mancini, M. F. Roquebert, F. Farinole, F. Pietra, Tetrahedron 2002, 58 , 9163–9167. 129. G. Bringmann, K. Messer, W. Saeb, E. M. Petes, K. Peters, Phytochemistry 2001, 56 , 387–391. 130. L. Liu, A. L. Li, M. B. Zhao, P. F. Tu, Chem. Biodivers. 2007, 4 , 2932–2937. 131. K. Kagawa, K. Tokura, K. Uchida, H. Kakushi, T. Shike, H. Nakai, Chem. Pharm. Bull . 1992, 40 , 2083–2087. 132. K. Machida, E. Matsuoka, T. Kasahara, M. Kikuchi, Chem. Pharm. Bull . 2005, 53 , 934–937. 133. N. Allouche, B. Morleo, O. Thoison, V. Dumontet, O. Nosjean, F. Gueritte, T. Sevenet, M. Litaudon, Phytochemistry 2008, 69 , 1750–1755. 134. J. Thiem, M. Gerken, G. Snatzke, Liebigs Ann. Chem. 1983, 448–461. 135. S. E. Boiadjiev, D. A. Lightner, J. Am. Chem. Soc. 2000, 122 , 11328–11339. 136. P. S. Steyn, C. W. Holzapfel, Tetrahedron 1967, 23 , 4449–4461. 137. P. Bovicelli, A. Sanetti, R. Bernini, P. Lupattelli, Tetrahedron 1997, 53 , 9755–9760. 138. P. Salvadori, S. Superchi, F. Minutolo, J. Org. Chem. 1996, 61 , 4190–4191.

113

114

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

139. K. Krohn, R. Bahramsari, U. Florke, K. Ludewig, C. KlicheSpory, A. Michel, H. J. Aust, S. Draeger, B. Schulz, S. Antus, Phytochemistry 1997, 45 , 313–320. 140. K. G. Guimaraes, J. Filho, T. R. Dos Mares-Guia, F. C. Braga, Phytochemistry 2008, 69 , 439–444. 141. Y. Asakawa, Y. Tada, T. Hashimoto, Phytochemistry 1994, 37 , 233–235. 142. K. Fujimoto, T. Fujimaki, E. Okuyama, M. Yamazaki, Chem. Pharm. Bull . 1999, 47 , 1426–1432. 143. K. Baba, Y. Q. Xiao, M. Taniguchi, H. Ohishi, M. Kozawa, Phytochemistry 1991, 30 , 4143–4146. 144. M. Taniguchi, M. Yanai, Y. Q. Xiao, T. Kido, K. Baba, Phytochemistry 1996, 42 , 843–846. 145. K. Krohn, I. Kock, B. Els¨asser, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2007, 1123–1129. 146. K. Krohn, R. Bahramsari, U. Fl¨orke, K. Ludewig, C. KlicheSpory, A. Michel, H. J. Aust, S. Draeger, B. Schulz, S. Antus, Phytochemistry 1997, 45 , 313–320. 147. H. Hussain, N. Akhtar, S. Draeger, B. Schulz, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, K. Krohn, Eur. J. Org. Chem. 2009, 749–756. 148. G. Snatzke, Tetrahedron 1965, 21 , 439–448. 149. W. Gaffield, Tetrahedron 1970, 26 , 4093–4108. 150. W. J. McGahren, G. A. Ellestad, G. O. Morton, M. P. Kunstman, J. Org. Chem. 1972, 37 , 1636–1639. 151. C. Galeffi, P. Rasoanaivo, E. Federici, G. Palazzino, M. Nicoletti, B. Rasolondratovo, Phytochemistry 1997, 45 , 189–192. 152. J. Dai, K. Krohn, U. Fl¨orke, S. Draeger, B. Schulz, A. Kiss-Szikszai, S. Antus, T. Kurt´an, T. van Ree, Eur. J. Org. Chem. 2006, 3498–3506.

4 ELECTRONIC CD EXCITON CHIRALITY METHOD: PRINCIPLES AND APPLICATIONS Nobuyuki Harada, Koji Nakanishi, and Nina Berova

4.1. INTRODUCTION AND HISTORICAL OVERVIEW The electronic CD exciton chirality method enables one to determine the absolute configuration (AC) of various chiral compounds in a nonempirical manner without reference to compounds with known AC [1–7]. Namely, if a compound contains two identical chromophores in a chiral position, where each chromophore undergoes an intense π –π ∗ transition, these two electrically allowed π –π ∗ transitions interact with each other to generate the so-called bisignate intense CD Cotton effects (CEs). The sign of the bisignate CEs reflects the absolute arrangements of the two chromophores in the molecule. By observing the exciton coupled CD, one can determine the AC of chiral compounds. The exciton CD is very intense and its generation mechanism is simple. Since the exciton chirality rule can be proven by derivation of quantum mechanical equations without numerical calculation as shown below, the CD exciton chirality method (ECM) is classified as a nonempirical rule. The nature and mechanism of the CD exciton coupling are explained in this chapter. The science of stereochemistry started when Louis Pasteur first succeeded in the so-called “optical resolution” of racemic tartaric acid in 1848 [8, 9], and the theory of “tetrahedral carbon atom” was then proposed independently by J. H. van’t Hoff and J. A. Le Bel in 1874 to explain the enantiomeric structures of optically active compounds (Figure 4.1) [9–12]. In 1895, A. Cotton discovered an anomalous dispersion effect in the optical rotation phenomena, which became known as the Cotton effect (CE) [13]. However, because it was not possible to determine the ACs of chiral compounds at that time, E. Fischer proposed a convention where the (D)-AC was arbitrarily assigned to (+)-glyceraldehyde as the standard of chiral organic compounds [14]. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

115

116

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 4.1. Historical overview of the determination of AC.

Quantum mechanics appeared upon entering the twentieth century; and in 1928, L. Rosenfeld reported the equation expressing the rotational strength R, a parameter governing the sign and magnitude of optical rotation, which is equal to the imaginary part of the scalar product between electric transition moment and magnetic transition moment [15]. In the 1930s, physicists and physical chemists challenged to determine the AC of chiral compounds by the calculation of optical rotation, developing their own theories. For example, W. Kuhn proposed the coupled-oscillator concept based on the classical theory [16]. E. U. Condon, W. J. Kauzmann, and H. Eyring applied the quantum mechanical one-electron theory [17], while J. G. Kirkwood developed the polarizability theory [18]. However, none were sufficiently convincing to claim that ACs of chiral compounds could be determined. For example, the coupled-oscillator theory, later extended to the exciton CD theory, had been applied to compounds with regular substituents, but not to compounds with two or more chromophores. The optical rotations of target compounds were small and not suited for determining ACs. The concept of “exciton coupling” was later developed by A. S. Davydov in 1948 for studying the UV spectra of molecular crystals [19]. The history of AC determination of chiral compounds could have been quite different, if physicists and physical chemists had realized in the 1930s that the coupled-oscillator theory would be more suited for the compounds with two identical chromophores and collaborated with organic chemists to synthesize such compounds. That is, it would have been possible to determine the AC of chiral organic compounds prior to Bijvoet’s discovery in X-ray crystallography. Unfortunately, at that time, the concept of the chiral interaction between two identical chromophores had not yet been adopted. It was later that the UV exciton theory was developed by A. S. Davydov. In 1951, the AC of a chiral compound was first determined by a totally different method. Namely, J. M. Bijvoet succeeded to determine the L-(2R,3R) AC of (+)-tartaric acid by using the anomalous scattering effect of heavy atoms in X-ray crystallographic diffraction experiments [20]. It was fortunate that the AC arbitrarily selected in Fischer’s convention agreed with the results of the X-ray Bijvoet method. Later in the field of chiroptical spectroscopy, the relation between helical structures of biopolymers and chiroptical spectra was rationalized by using the so-called coupledoscillator mechanism. In 1956, W. Moffitt extended the theory to study the OR and CD

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

of proteins [21]. In 1962, I. Tinoco and co-workers applied the theory to the CD of DNAs and oligonucleotides [22]; the same year, J. A. Schellman applied the theory to helical polypeptides [23]. In these studies, the ACs of biopolymers or oligomers were already known through the AC of monomeric units determined by X-ray crystallography. In the field of organic stereochemistry, the coupled-oscillator theory was first applied in 1962 by S. F. Mason to calycanthine, a natural dimeric alkaloid with C2 symmetry [24]. The CD spectrum of calycanthine shows a positive CE at 259 nm and negative CE at 240 nm in the 1 La transition (252 nm) of the aniline chromophores, from which the AC was determined; the AC was later confirmed by X-ray crystallography [25]. It should be noted that this is the first application of the coupled-oscillator mechanism to a chiral organic compound. S. F. Mason and co-workers applied the same method to other chiral compounds, but later it became clear that some compounds were not suited for the coupled oscillator mechanism. For example, the AC of Troger’s base assigned by the coupled-oscillator method [26] was later revised by X-ray crystallography [27]. In 1969, N. Harada and K. Nakanishi reported the dibenzoate chirality rule for determining the AC of chiral glycols [28]. Here the absolute helicity between two benzoate chromophores (i.e., AC of the original glycols) could be unambiguously determined from the bisignate CEs of dibenzoates. This dibenzoate chirality rule was based on the coupled-oscillator mechanism, and it opened a general protocol for determining ACs as the CD exciton chirality method [1, 2]. In the history of the absolute configurational assignment, there were many controversies, and among them the biggest one was raised in 1972–1973 [29]. Namely, it was claimed that the ACs determined by the X-ray Bijvoet method disagreed with those assigned by the exciton coupling mechanism and that the ACs determined by the Bijvoet method had to be reversed because of an error in the Bijvoet theory. If this was true, all organic chemistry textbooks would have to be revised. In 1973, Y. Saito immediately pointed out that there is no error in the theory of the X-ray Bijvoet method [30], and H. H. Brongersma and P. M. Mul experimentally confirmed the Bijvoet method [31]. In the same year, S. F. Mason reported that the dipole velocity treatment of CD led to the correct AC [32]. In 1976, N. Harada reported the synthesis and CD of an ideal compound connecting the X-ray Bijvoet and CD exciton chirality methods (Section 4.4) [33]. It is now established that both methods lead to the same and correct AC.

4.2. OUTLINE AND PRINCIPLE OF CD EXCITON CHIRALITY METHOD The CD exciton chirality method (ECM) has been successfully applied to a variety of natural products and synthetic chiral compounds to determine their ACs. This method enables one to determine the AC of a chiral compound without any reference, that is, it is a nonempirical method [1–7]. For example, the CD and UV spectra of cholest-5ene-3β,4β-diol bis(p-bromobenzoate) 1 are illustrated in Figure 4.2, where UV spectrum shows an intense band of the allowed π –π ∗ transition at 244.0 nm, which is polarized along the long axis of the p-bromobenzoate chromophore [5]. In this region, the CD shows intense negative first and positive second CEs (λext 243.6 nm, ε −30.4: λext 236.2 nm, ε +21.2; A = −51.6) [5]; the CE at longer wavelength is called the first CE, while the shorter wavelength extremum the second CE. The exciton CD shows two CEs of similar intensity but opposite signs (Figure 4.2), which are called “bisignate” CEs. The exciton bisignate CD reflects the “exciton chirality”—that is, helical sense between two electric transition dipole moments (ETDMs) involved in the excitation:

117

118

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+40

+20

Br

group i 3β-equatorial

CD

Δε

O

3 O

0

–20

–40 UV

CD 243.6 (–30.4) 236.2 (+21.2) A = –51.6

ε × 10–4

Br

O

H

4 O O

O O

1

O group j 4β-axial

Br

4β-axial group j in rear Br

6 4

UV 244.0 (41,800)

200

300 λ (nm)

250

2

X-ray crystallographic stereoview

3β-equatorial group i in front

0

(a)

(b)

Figure 4.2. (a) CD and UV spectra of cholest-5-ene-3β,4β-diol bis(p-bromobenzoate) 1: UV in 0.3% 1,4-dioxane/EtOH; CD in 10% 1,4-dioxane/EtOH. (b) Negative exciton chirality between long axes of two p-bromobenzoate chromophores: Newman projection and X-ray crystallographic stereoview. (Redrawn from reference 5, with permission.)

(i) If the exciton CD shows negative first and positive second CEs, the two ETDMs constitute a counterclockwise screw sense as in the case of bis(p-bromobenzoate) 1 (Figure 4.2). (ii) If the exciton CD shows positive first and negative second CEs, the two ETDMs constitute a clockwise screw sense. From this relation, the AC of the target compound, namely, cholest-5-ene-3β,4β-diol, can be determined. This is the electronic CD exciton chirality method [1–7]. In the case of bis(p-bromobenzoate) 1, the π –π ∗ transition at 244 nm is polarized along the long axis of the p-bromobenzoate chromophore, and the exciton chirality corresponds to the helicity between the long axes of two chromophores. As illustrated in the Newman projection, the two long axes constitute anticlockwise screw sense generating negative first and positive second CEs. The counterclockwise screw sense between two p-bromobenzoate groups is directly observed in the X-ray crystallographic stereoview in Figure 4.2 [5].

4.2.1. Principles and Nonempirical Nature of Exciton Chirality Method The second example used for explaining the principle of the ECM is 5α-cholestane-2β,3βdiol bis(p-dimethylaminobenzoate) 2 (Figure 4.3). When two identical chromophores i and j , with intense UV π –π ∗ transition (ground state 0 → excited state a), exist in a molecule, two chromophores interact with each other to split the excited state into two energy levels (α and β states), while the ground state (0) remains unsplit [1]. This phenomenon, the exciton coupling or exciton interaction, generates two electronic transitions, from ground state 0 to excited states α and β—that is, transitions 0 → α and 0 → β. The wavefunction, energy, dipole strength, and rotational strength for the α-state

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

α−state:

N

wave function, energy, dipole strength, rotational strength,

O group j O H O

H

N

β−state:

2

O group i

5α-Cholestane-2β,3β-diol bis(p-dimethylaminobenzoate) a

0 group i

β α

a

0 0 total system group j

wave function, energy, dipole strength, rotational strength,

⎯ 2) (fiafj0 – fi0fja) yaa = (1/√ E a = Ea – Vij D a = (1/2)(li0a – lj 0a)2 R a = +(1/2)ps0 Rij • ( li 0a × lj0a) yab = (1/√2) (fiafj0 + fi0fja) Eb = Ea + Vij D b = (1/2)(li 0a + lj 0a)2 R b = –(1/2)ps0 Rij • ( li 0a × lj 0a)

interaction energy, Vij = Rij –3{li0a lj0a – 3Rij–2( li 0aRij) (lj0aRij)} If Vij > 0, the a-state is lower in energy than the b-state. α-state, longer wavelength side => 1st Cotton effect. β-state, shorter wavelength side => 2nd Cotton effect.

Figure 4.3. Theoretical summary of the CD ECM.

and β-state are summarized in Figure 4.3, where Vij is defined as the interaction energy between two electric transition moments μi 0a and μj 0a . If Vij is positive, the α-state corresponds to the transition at longer wavelength, while the β-state corresponds to the transition at shorter wavelength. As shown in Figure 4.3, the rotational strength R α of the α-state is opposite in sign to that of the β-state, R β , but their absolute values are equal. Note that the sign and magnitude of R α and R β are governed by the triple product R ij • (μj 0a × μj 0a ) [1]. These equations were next applied to bis(p-dimethylaminobenzoate) 2 in Figure 4.4, where electric transition moments μi 0a and μj 0a of the benzoate chromophores were assigned as shown. Since vectors μi 0a and μj 0a are set in-phase, the interaction energy Vij becomes positive, and hence the α-state is lower in energy than the β-state. Two vectors μi 0a and μj 0a constitute a clockwise screw, and so the resultant vector (μj 0a × μj 0a ) becomes parallel to the distance vector R ij . Therefore the triple product R ij • (μj 0a × μj 0a ) becomes positive, and R α is positive, while R β is negative. This leads to the CD in Figure 4.4b, where the CE at longer wavelength (1st CE) is positive and that at shorter wavelength (2nd CE) is negative [1]. Figure 4.4c shows the UV and CD spectra of bis(p-dimethylaminobenzoate) 2, which has an intense π –π ∗ transition (λmax 307 nm, ε 54,300) polarized along the long axis of the chromophore. The CD spectrum shows positive 1st and negative 2nd CEs in agreement with the theoretical conclusion: 1st CE, λext 320 nm, ε +61.7 and 2nd one, λext 295 nm, ε −33.2. The amplitude of the exciton CD is defined as A = ε1 − ε2 , where ε1 and ε2 are ε values of 1st and 2nd CEs, respectively. In the case of dibenzoate 2, A = +94.9. From these results, the AC of the original glycol is readily determined. In Figure 4.4a, the in-phase combination of vectors μi 0a and μj 0a was considered. But we can theoretically choose another case of in-phase relation and two cases of outof-phase combination. What will happen in these cases? Due to the self-consistency of the exciton CD theory, the theoretical results agree with that in Figure 4.4b; that is, the exciton CD depends only on the mutual absolute arrangements of two long

119

120

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N

μi0a

O O H 2

N O

Vector product μi0a × μj0a

O

+40

O H

Parallel

Rij

group i (1) Interaction energy Vij > 0 (2) Parallel, so R α > 0,and R β < 0

H

O H O

O 3

N +20

2

Δε

(a)

OBz-dma OBz-dma

ε × 10–4

μj0a

group j

N

320 (+61.7)

+60

CD

0 307 (54,300)

in EtOH

Positive 1st Cotton –20

295 (–33.2)

λ –40 Negative 2nd Cotton

6 4 2

UV

0 200

250

300

350

λ (nm)

(b)

(c)

Figure 4.4.

Application of the CD ECM to 5α-cholestane-2β,3β-diol bis(p-dimethyl-

aminobenzoate) 2: CD and UV spectra in EtOH. (Redrawn from reference 28g, with permission.)

axes of benzoate chromophores. In application of the CD ECM for AC determination, it is unnecessary to consider the in-phase or out-of-phase combination of ETDMs. The exciton chirality governing the sign and intensity of CEs is defined as shown in Table 4.1 [1]. As discussed, the nonempirical nature of the CD ECM is easily proved, indicating the simplicity of exciton CD mechanism. Further details of the quantum mechanical molecular exciton CD theory are described in the Section 4.3.

TAB L E 4.1. Definition of Exciton Chirality Qualitative Definition

Quantitative Definition

CEs

Positive exciton chirality

R ij • (μioa × μjoa )Vij > 0

Negative exciton chirality

R ij • (μioa × μjoa )Vij < 0

Positive first (at longer wavelength) and negative second (at shorter wavelength) Cotton effects Negative first (at longer wavelength) and positive second (at shorter wavelength) Cotton effects

Source: Redrawn from reference 28g, with permission.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

121

The qualitative definition of exciton chirality is simple: If two ETDMs constitute a clockwise screw sense, CD shows positive first and negative second CEs, and vice versa. In general, intense exciton CD CEs are observed at the long axis-polarized transition, and the essential points of the ECM are summarized as follows [1]. 1. If the long axes of two interacting chromophores constitute a clockwise screw sense, the CD shows a positive first CE at longer wavelength and a negative second CE at shorter wavelength (Table 4.1 and Figure 4.5). 2. If they constitute a counterclockwise screw sense, a negative first CE at longer wavelength and a positive second CE at shorter wavelength result (Table 4.1 and Figure 4.5). Pertinent features of the exciton CD follow. 1. The intensity of the exciton CD (A value) is inversely proportional to the square of the interchromophoric distance Rij , provided that the remaining angular part is the same [1]. A(= ε1 − ε2 ) ∝ Rij −2 2. The A value of exciton split CD is a function of the dihedral angle between two transition moments. In vicinal glycol dibenzoates, the sign of the exciton split CEs remains unchanged from 0◦ to 180◦ . The qualitative definition shown in Table 4.1 is applicable to a dibenzoate with the dihedral angle of more than 90◦ . The maximum A value is around 70◦ [1]. 3. The A value is proportional to the square of absorption coefficient ε of the chromophore. Therefore, it is advisable to use chromophores undergoing intense π –π ∗ transition. In general, a weak transition along the short axis of chromophores is unsuitable.

Figure 4.5. Exciton coupled CD CEs and UV absorption band. In general, the CD zero-crossing point corresponds to λmax of UV band. (Redrawn from refernece 1, with permission.)

122

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4. The CEs of α and β states have identical rotational strength of opposite signs, and two CEs are conservative, satisfying the sum rule. Thus the integrated areas of positive and negative CEs are equal. 

Rk = 0

5. Rotational strength R should be origin-independent because it is a physically observable quantity. Equations in Figure 4.3 satisfy the origin-independence of rotational strength.

4.3. THEORY OF EXCITON CD SPECTROSCOPY 4.3.1. CD Spectra and Rotational Strength of CE The rotational strength R, a parameter representing the sign and intensity of a CE, is experimentally obtained from the observed CD spectra as shown in Eq. (4.1) [34]. R = 2.296 × 10−39

 ε(σ )/σ dσ

(cgs unit)

(4.1)

where σ is wavenumber. The rotational strength R is theoretically formulated by Eq. (4.2) as proposed by Rosenfeld [15]. R = Im{< 0|μ|a > • < a|M |0 >}

(4.2)

where Im denotes the imaginary part of the terms in brackets, < > denotes the integration over configuration space, μ and M are operators of electric and magnetic moment vectors, respectively. The dot stands for scalar product of two vectors, 0 and a are wavefunctions of ground and excited states, respectively. Rotational strength R is thus equal to the imaginary part of the scalar product of electric and magnetic transition moments. A Gaussian distribution approximation of a CD Cotton effect curve leads to Eq. (4.3) [1, 34]. ε(σ ) = εmax exp{−((σ − σo )/σ )2 }

(4.3)

where εmax is the maximum intensity of the CE, σo is the central wavenumber of the CE, and σ is half the band width at 1/e peak height of the Gaussian curve. From eq. (4.1) and (4.3), we obtain √ R = 2.296 × 10−39 π εmax σ/σo

(4.4)

From Eqs. (4.3) and (4.4), the CD curve is formulated as √ ε(σ ) = (σo /(2.296 × 10−39 π σ ))R exp{−((σ − σ0 )/σ )2 }

(4.5)

where σ can be evaluated from observed UV–Vis spectra. Provided that rotational strength R is calculated by Eq. (4.2), a CD spectrum can be reproduced by theoretical calculation [1, 34].

123

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

4.3.2. Molecular Exciton Theory of a Binary System with Two Chromophores According to the quantum mechanical exciton theory, in the exciton coupling system composed of two identical chromophores i and j , exciton wavefunctions are expressed by (4.6) and (4.7), where each chromophore undergoes excitation 0 → a [1]: Ground state :

φi 0 ,

φj 0

(4.6)

Excited state :

φia ,

φja

(4.7)

The Hamiltonian operator of the coupling system is formulated as H = Hi + Hj + Hij

(4.8)

where Hi and Hj are the Hamiltonian of groups i and j , respectively, and Hij is the interaction energy term between two groups i and j . The ground-state wavefunction and energy of a binary system are expressed as ψ0 = φi 0 φj 0

(4.9)

E0 = 0

(4.10)

The singly excited state of the binary system splits into two energy levels, α and β states. For the α state, wave function : Energy :

√ ψaα = (1/ 2){φia φj 0 − φi 0 φja } α

E = Ea − Vij

(4.11) (4.12)

where Vij is the interaction energy between two groups i and j and is approximated by the point dipole approximation: Vij = μi 0a μj 0a Rij−3 {ei • ej − 3(ei • eij )(ej • eij )}

(4.13)

where μioa , μjoa , and Rij are absolute values of vectors μioa , μjoa , and R ij , respectively; ei , ej , and eij are unit vectors of μioa , μjoa , and R ij , respectively. For the terms μioa , μjoa , and R ij , see Eqs. (4.18), (4.19), and (4.31), respectively. For the β state, wave function: Energy:

√ ψaβ = (1/ 2){φia φj 0 + φi 0 φja }

(4.14)

E β = Ea + Vij

(4.15)

These equations indicate that the binary system has two electronic transitions, 0 → α and 0 → β, in UV–Vis spectrum. If Vij > 0, the α state is lower in energy than the β state, and therefore the transition 0 → α locates at longer wavelengths, while the transition 0 → β is at shorter wavelengths. The electric dipole moment operator μ of a whole system is defined as μ=



μi =



er is

(4.16)

124

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

where μi is the electric dipole moment operator of group i , e is the elementary charge, and r is is the distance vector of electron s in group i from the origin. The electric transition moment α of the transition 0 → α is formulated as  √ (4.17) < 0|μ|a>α = μo aα = ψo μψ aα d τ = (1/ 2)(μioa − μjoa ) where  μioa =

φio μi φia d τi

(4.18)

φjo μj φja d τj

(4.19)

 μjoa =

Those are electric transition moments of transition 0 → a in groups i and j , respectively. The electric transition moment of the transition 0 → β is similarly expressed as < 0|μ|a>β = μoa β =



√ ψo μψ a β d τ = (1/ 2)(μioa + μjoa )

(4.20)

The magnetic moment operator M of a whole system is formulated as M =



M i = (e/2mc)



r is × p is

(4.21)

where m is the mass of electron, c is the velocity of light, p is is the linear momentum of electron s in group i , and × stands for vector product of two vectors. The magnetic moment operator is further changed as M = (e/2mc)



Ri × pi +



mi

(4.22)

where R i is distance vector of group i from the origin, p i and m i are linear momentum and internal magnetic moment operators of group i , respectively. The magnetic transition moment < a|M |0>α of the excitation 0 → α is calculated as  < a|M |0>α = ψaα M ψo d τ √ = (1/ 2){(e/2mc)R i × p iao + m iao − (e/2mc)R j × p jao − m jao } (4.23) where p iao and m iao are linear momentum and internal magnetic moment of group i , respectively. For the linear momentum of a group, the next equation is useful [35]: p oa = −(2π imc/e)σo μoa

(4.24)

where i is the symbol of imaginary, σo is excitation energy expressed in wavenumber units, and μoa is electric transition moment of transition 0 → a. For group i , p iao = (2π imc/e)σo μioa

(4.25)

125

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

Accordingly, √ α = (1/ 2){i π σo R i × μioa − i π σo R j × μjoa + m iao − m jao }

(4.26)

In a similar manner, the magnetic transition moment of excitation 0 → β is calculated as √ β = (1/ 2){i π σo R i × μioa + i π σo R j × μjoa + m iao + m jao } (4.27) The dipole strength D α of a binary system for the excitation 0 → α is expressed as D α = (1/2)(μioa − μjoa )2

(4.28)

D β = (1/2)(μioa + μjoa )2

(4.29)

Similarly,

Rotational strength R α of the α state is derived from Eqs. (4.2), (4.17), and (4.26): R α = Im{< 0|μ|a>α • α } = (1/2)Im{(μioa − μjoa ) • (m iao − m jao )} + (1/2)π σo R ij • (μioa × μjoa )

(4.30)

where R ij is the interchromophoric distance vector from group i to group j and is defined as (4.31) R ij = R j − R i Similarly, β

R β = Im{< 0|μ|a>β • < a|M |0 >} = (1/2)Im{(μioa + μjoa ) • (m iao + m jao )} − (1/2)π σo R ij • (μioa × μjoa )

(4.32)

In the case of π → π ∗ transition of common molecules, internal magnetic transition moments m iao and m jao are negligible. Therefore, rotational strengths are approximated as: (4.33) R α = +(1/2)π σo R ij • (μioa × μjoa ) R β = −(1/2)π σo R ij • (μioa × μjoa )

(4.34)

These equations indicate that the CEs of α and β states have equal intensity but of opposite signs. Thus exciton CD satisfies the sum rule.  (4.35) RA = R α + R β = 0 The rotational strength is proportional to the triple product of interchromophoric distance and electric transition moments of groups i and j . Therefore, provided that chromophores exhibiting intense π –π ∗ transitions are used, intense exciton CEs are observable. The rotational strengths of exciton CD satisfy the origin-independence as shown in Eqs. (4.31), (4.33), and (4.34).

126

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4.3.3. Exciton CD of N-mer and Dimer: Quantitative Definition of Exciton Chirality The exciton theory is applicable to UV–Vis and CD spectra of N -mer having N identical chromophores [1]. When N chromophores undergoing intense π –π ∗ transition (0 → a) interact with one another, the excited state splits into N energy levels. The wavenumber σk of k th excitation is formulated as  Cik Cjk Vij (4.36) σk − σo = 2 where coefficients Cik and Cjk are obtained by solving the N th-order secular equation. The rotational strength R k is expressed as R k = −π σo



Cik Cjk R ij • (μioa × μjoa )

(4.37)

For the N -mer, CD curve is formulated as  √ ε(σ ) = {σo /(2.296 × 10−39 πσ )} R k exp{−((σ − σk )/σ )2 }

(4.38)

The Taylor expansion of Eq. (4.38) against σk /σ around σo /σ gives the second term of expansion as √ ε(σ ) = {2σo /(2.296 × 10−39 π σ )} exp{−((σ − σo )/σ )2 }{(σ − σo )/σ }  × R k {(σk − σo )/σ } (4.39) From Eqs. (4.36), (4.37), and (4.39), the CD equation of N -mer is obtained: √ ε(σ ) = {4 π σo2 /(2.296 × 10−39 σ 2 )}{(σo − σ )/σ } exp{−((σo − σ )/σ )2 }        (4.40) × Cik Cjk R ij • (μioa × μjoa ) Cik Cjk Vij It should √ be noted that for α state √ in the case of a binary system, √ the coefficients √ are always 1/ 2 and −1/ 2, and for β state they are 1/ 2 and 1/ 2, for any mutual configuration of two identical chromophores. Therefore, Eq. (4.40) is simplified as √ ε(σ ) = {2 π σo2 /(2.296 × 10−39 σ 2 )}{(σo − σ )/σ } exp{−((σo − σ )/σ )2 } × R ij • (μioa × μjoa )Vij

(4.41)

This is the exciton CD equation of a binary system. The next term of Eq. (4.41), √ {2 π σo2 /(2.296 × 10−39 σ 2 )}{(σo − σ )/σ } exp{−((σo − σ )/σ )2 } represents an anomalous dispersion curve with positive and negative extrema. The sign and intensity of exciton CD depend on the quadruple term R ij • (μioa × μjoa )Vij . Therefore, the term R ij • (μioa × μjoa )Vij is adopted as the quantitative definition of exciton chirality. This term is also expressed as R ij • (μioa × μjoa )Vij = Dioa Djoa Rij−2 eij • (ei × ej ){ei • ej − 3(ei • eij )(ej • eij )} (4.42)

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

where Dioa and Djoa are transition dipole strengths of groups i and j , respectively. From this equation, it is suggested to use chromophores undergoing intense π –π ∗ transition in the UV–Vis spectrum for obtaining intense exciton CD. This equation also indicates that the exciton CD amplitude is inversely proportional to the square of the interchromophoric distance Rij .

4.3.4. Theoretical Simulation of Exciton CD As discussed above, the ECM is simple in mechanism, and exciton CD spectra were simulated by various theoretical methods. There were reported examples of simulation by the DeVoe coupled oscillator method [1, 6, 36], the π -electron SCF-CI-DV MO method [1; See Chapter 5, this volume], and more recent ab initio and related MO methods—for example, TDDFT B3LYP/6-32G(d) [37]. For complex CD spectra and molecules, the TD-HF/6-31G(d), TDDFT B3LYP/6-32G(d), and CAM-B3LYP/6-32G(d) methods would be very useful (see theoretical chapters). The simulations are important for confirming the ACs determined by the ECM, and these theoretical methods provide better understanding of the CD generation mechanism.

4.4. THE CONSISTENCY BETWEEN X-RAY BIJVOET AND CD EXCITON CHIRALITY METHODS These methods are based on totally different physical phenomena, but they should give the same AC for a specific compound. However, it was claimed in 1972 that the ACs determined by X-ray and CD exciton methods disagreed and that the ACs determined by the X-ray Bijvoet method should be revised [29]. This claim was based on the X-ray and CD analyses of compounds (–)-5 and (+)-6 in Figure 4.6, where the CD of the weak 1 Lb transition (∼290 nm) of aniline chromophore polarized along the short axis was analyzed as an exciton couplet. However, this claim was subsequently retracted. Note that the ECM should be applied to an intense UV transition (Section 4.5), but not to a weak UV transition. In 1976, the synthesis and CD spectrum of a chiral cage compound (+)-3 with two anthracene chromophores as an ideal model for ECM were reported [33]; the results unambiguously proved the consistency between X-ray Bijvoet and CD exciton methods (Figure 4.6). Compound (+)-3 was synthesized from diester (+)-4, which was chemically correlated with compounds (−)-5 and (+)-6. The ACs of (−)-5 and (+)-6 had been determined by the Bijvoet method [38]. As expected, compound (+)-3 showed strong exciton-coupled CD CEs at the strong 1 Bb transition of anthracene polarized along the long axis: λext 268.0 nm (ε +931.3), 249.7 (−720.8), A = +1652.1 (Figure 4.6), showing that the strong UV transition gives rise to intense the exciton coupled CD. Since the UV π –π ∗ transition at 267.2 nm is polarized along the long axis of the anthracene chromophore, the exciton split CEs at 268.0 nm and 249.7 nm are generated by the exciton coupling between two ETDMs of anthracene groups. The long axes of the anthracene moieties constitute a clockwise screw sense leading to positive first and negative second CEs, and therefore the AC of compound (+)-3 was determined as shown. This result agrees with that determined by the X-ray Bijvoet method (Figure 4.6) [33]. Another controversy regarding the ACs of clerodin (7) and related diterpenes should be noted. In 1962, the AC of clerodin 7 [39] was determined by the X-ray Bijvoet method, which was opposite to the AC shown in Figure 4.7 [40]. Since this AC was believed to

127

128

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 4.6.

CD and UV spectra of (6R,15R)-(+)-6,15-dihydro-6,15-ethanonaphtho-[2,3c]pentaphene 3 in dioxane/EtOH, and chemical correlation between compound 3 and related compounds. (Redrawn from reference 33a, with permission.)

(−)-clerodin 7

3-epicaryoptin derivative bis(p-Cl-benzoate) 9

steroidal model 11

Figure 4.7. Correct ACs of (–)-clerodin 7 and related compound 9, and comparison of exciton CD of compounds 9 and 11.

be correct, clerodin 7 had been used as a reference of AC for newly isolated compounds of the clerodane family for many years. For example, in 1974 the AC of 3-epicaryoptin 8 was determined by comparison of CD and chemical correlation with 7 [41]. At the same time, an unexpected result was reported; the CD ECM was applied to 3-epicaryoptin derivative 3,6-bis(p-Cl-benzoate) 9, but the observed positive couplet was opposite to the negative couplet expected from the AC of 9, which was based on the X ray of 7. To explain the discrepancy, it was postulated that one of the benzoate groups is twisted by an intramolecular H bond, generating a positive exciton chirality, and the result was reported as an exception of the CD ECM [41]. To solve the problem of this unexpected discrepancy, the steroidal model compound 11 was synthesized in 1978 [42], because compounds 9 and 11 have the same relative

129

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

configurations at key positions (Figure 4.7). The CD spectrum of 11 showed a positive couplet, concluding that the conformation of the benzoate group was not twisted by an H bond. Since 3,6-bis(p-Cl-benzoate) 9 and 11 showed CD couplets of the same sign, 9 should have the same AC as 11. From the exciton CD and molar rotation data, the ACs of clerodane diterpenes 7 and 8 were reversed [42]. The revised ACs were confirmed by X-ray analysis later [43]. Unfortunately, the incorrect AC of clerodin 7 had been used as a reference of the clerodane diterpene family for about 16 years.

4.5. SUITABLE CHROMOPHORES FOR THE CD ECM AND EXAMPLES It is essential to select suitable chromophores for the application of the CD ECM. The following issues should be considered: (i) intense π –π ∗ transition, (ii) unambiguous direction of ETDM, and (iii) symmetrical structure. Figure 4.8 shows typical chromophores useful for the CD ECM, in which arrows show the ETDM direction leading to exciton coupled CD. In general, the long-axis polarized transitions are suitable for exciton CD, because of the larger UV intensity. (a) Para-Substituted Benzoate Chromophores for Glycols The intramolecular CT or 1 La transition (230–310 nm) of para-substituted benzoate chromophores has been used for determining the AC of many glycols [1, 3]. The intramolecular CT transition is polarized along the long axis of the benzoate chromophore, which is almost parallel to the alcoholic C–O bond. Therefore, the AC of

Chromophores for -OH groups: O X O X = H, Br, OMe, NMe2 230–310 nm O

Ph O

X

N

X = OMe, NMe2 O 300–360 nm

Ph N

270 nm fluorescent

M = 2H, Zn, Mg

For -C=C- groups:

O

N

O

O 260 nm fluorescent

280 nm fluorescent

230 nm fluorescent

Ph

O O

Me2N

MeO

N 305 nm

O 260 nm

O 350 nm

O

420 nm fluorescent

For -COOH groups:

O

Me2N

N Ph

O

For -NH2 groups:

O

O

O 235 nm fluorescent

N M

HN

N

N M

Ph N

N

M = 2H, Zn, Mg Ph 420 nm fluorescent

Figure 4.8. Chromophores useful for the CD ECM, where arrows show the direction of ETDM. (Redrawn from reference 5, with permission.)

130

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the glycol part can be determined from the exciton CD data. In contrast, ortho- and meta-substituted benzoate chromophores are not suited, because their ETDMs are not parallel to the alcoholic C–O bond. (b) Cinnamate, β-Naphthoate, and Other Chromophores for Glycols These chromophores are characterized by their strong absorption at longer wavelengths [44]. (c) Tetraphenyl-Porphyrin-Carboxylic Acid (TPP-COOH) Tetraarylporphyrins and their zinc and copper analogues deserve special attention as chromophores. They possess a very intense, sharp, and narrow Soret band, which is red-shifted to ∼420 nm and has ε ∼450,000–550,000. These porphyrins and metalloporphyrins belong to the most powerful and versatile CD chromophores [45–47]. One of the first examples for application of porphyrin in structural studies includes the red-tide toxin brevetoxin, where the CD exciton coupling was observed over a long distance of ˚ [45, 47, 48]. A detailed discussion on the application of porphyrins as CD ∼40–50 A reporter groups as well as an account of the theoretical analysis of porphyrin-porphyrin exciton interactions can be found in references 45–47. The Soret band originates from the two degenerate transitions Bx and By (Figure 4.9), which are perpendicular to each other; theoretically the porphyrin Soret band should be considered as a circular oscillator [46, 47]. However, due to the rotational flexibility around the meso porphyrin 5-C-phenyl junction (librational averaging), the transitions Bx and By can be represented by one effective transition moment along the 5–15 axis (Figure 4.9), and the exciton CD reflects the chirality between two effective transition moments. The large ε value and red shift of Soret band above 400 nm where most other chromophores do not absorb make the TPP-COOH extremely useful and versatile chromophore for exciton CD analysis. The typically very intense couplet can be measured with a high S /N ratio at very low concentration; thus it allows reliable studies on a microscale and, as mentioned above, also in cases where a very long-range coupling is involved [45, 48, 49]. Bx

15 N N 5 Ph M N N Effective transition moment

Ph

By X

3 O O

3α,17β

13, M: Zn2+ . UV-Vis: 419 nm ε 550,000 (CH2Cl2) fluorescence: λem 646, Φf 0.10

OR

TPP-COOH

O 17 O

12, M: H, H. UV-Vis: 418 nm ε 440,000 (CH2Cl2) fluorescence: λem 650, Φf 0.12

O

Δε +200 CD

+100

A = +270

0 –100

416 nm (–117) 419 nm (914,000)

–200

419 nm (550,000) monomer

350

1

bis (Zn-TPP) derivation CH2Cl2

UV-Vis

X Rij

14, X = H e = 15,000 Rij = 13.6 Å no coupling

424 nm (+153)

e × 10–6

Ph

400

0 450 λ (nm) 500

15, X = NMe2 e = 28,000 Rij = 13.6 Å A = +21 16, X = TPP e = 440,000 Rij = 24.4 Å A = +193 17, X = Zn-TPP e = 550,000 Rij = 24.4 Å A = +270

Figure 4.9. (Top) UV–Vis, and fluorescence data for TPP-COOH and its derivatives. (Bottom) CD and UV–Vis data of steroidal bis(tetraarylporphyrin) derivatives 16 and 17 together with corresponding benzoate derivatives 14 and 15. (Redrawn from reference 5, with permission.)

131

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

The large steric size of porphyrin chromophores, however, calls for attention when they are introduced at vicinal positions, such as in diols, polyols, diamines, and amino alcohols. The steric repulsion between bulky porphyrins may cause conformational changes upon porphyrin derivatization; therefore additional conformational studies by NMR or molecular modeling are recommended [47]. Figure 4.9 illustrates the striking increase in the A value seen in tetraarylporphyrin (TPP) and its Zn derivative (Zn-TPP) compared to p-dimethylaminobenzoate at the Rij ˚ (Figure 4.9). Other examples for efficient porphyrin–porphyrin CD distance of 24.0 A ˚ can be found [48]. coupling over 40–50 A (d) Benzamide, Phthalimide, and 2,3-Naphthalenedicarboximide Chromophores for Diamines and Amino Alcohols The CD ECM is also applicable to the intramolecular CT band of benzamide groups [1]. The transition is polarized along the long axis of the chromophore. However, in some cases such as N -methyl benzamides, the benzamide moiety exists as a mixture of (E ) and (Z ) isomers, and therefore, the mutual orientation of the ETDMs is uncertain. Thus in these cases, one should be cautious in assigning AC by exciton CD [50]; see Section 4.10 (1). The C2v -symmetrical phthalimide and 2,3-naphthalenedicarboximide chromophores are ideally suited for the ECM application to diamines and amino alcohols, because their long axis-polarized ETDMs are exactly parallel to the amine C–N bond [51, 52]. For example, the 2,3-naphthalenedicarboximide chromophore exhibits an intense 1 Bb transition around 260 nm, which is polarized along the long axis of the chromophore. Figure 4.10 shows the CD and UV spectra of trans-1,2-cyclohexanediyl bis(2,3-naphthalenedicarboximide) (1R,2R)-(−)-18, where the UV 1 Bb transition shows +100 in 10% 1,4-dioxane-EtOH

+50

O –50 Δε

H N O O

CD

–100

H N O (1R,2R)-(–) CD 264.8 (–154.6) 257.5 ( 0.0) 251.9 ( +38.9) UV A = –193.5 UV 259.0 (94,600) 252.2 (97,400)

–150

–200

200

250

300

350

ε x10–4

0

20

O

H

N H N

OO

H 18

(A) NMR: J1-H,2-H = 11.9 Hz

O

O

N

O

15 10

O

N

H O

(B)

5 0

400

λ (nm)

Figure 4.10.

CD and UV spectra of (1R,2R)-(−)-trans-1,2-cyclohexanediyl bis(2,3-naph-

thalenedicarboximide) 18 in 10% 1,4-dioxane/EtOH. (Redrawn from reference 52.)

132

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

two split bands. This split represents the exciton coupling, because the UV spectrum of mono(2,3-naphthalenedicarboximide) shows a single band in this region. The CD showed negative first and positive second CEs (λext = 264.8 nm, ε −154.6; λext = 251.9 nm, ε +38.9; A = −193.5). From the negative A value, a counterclockwise helicity between two long axes of 2,3-naphthalenedicarboximide chromophores was assigned to compound (−)-18, leading to the (1R,2R)-AC [52]. Note that compound (−)-18 can take diequatorial conformer (A) and diaxial conformer (B) as shown in Figure 4.10. In general, conformer (A) is considered to be more stable than conformer (B). However, if the electric repulsion between two polar groups is strong, conformer (B) may become more stable. To confirm the conformational preference of (A), the measurement of 1 H NMR coupling constant J1,2 is crucial. However, since these two protons are equivalent, it is not possible to determine J1,2 by regular 1 H NMR spectroscopy. Hence, the 13 C satellite signal method was used to give J1,2 = 11.9 Hz, indicating that conformer (A) was more stable than (B) [53]. Figure 4.11 illustrates a submicroscale chemical protocol developed for the analysis of sphingosines and dihydrosphingosines isolated from new cell lines. First the NH2 group of d-erythro-sphingosine 19 was blocked as a naphthalimide group yielding a derivative 20. Then the OH groups were converted to 2-naphthoate groups yielding derivative 21 that could be sensitively detected by HPLC, mass spec, CD and fluorescence analyses. The relative and ACs were assigned by comparing the observed CD with the standard CD curves of erythro- and threo-sphingosines/dihydro-sphingosines [54]. (e) Chromophores for Carboxylic Acids and Olefin Compounds The chromophores suitable for chiral carboxylic acids are listed in Figure 4.8. The application of the ECM to olefin compounds is unique. The isolated olefin group shows a π –π ∗ transition below 200 nm, and therefore the exciton method is not applicable in a straightforward manner. However, the double bond can be transformed via olefin metathesis into chromophores (see Figure 4.8) that absorb at longer wavelength, so that the entire exciton couplet can be conveniently measured [55]. (f) Compounds with Preexisting Chromophore that Interfere with Exciton CD: Use of Red-Shifted Chromophores When a natural product has a preexisting chromophore that could interfere with observation of exciton CD, chromophores with a longer wavelength UV λmax than the preexisting chromophore can be used to avoid overlap of CEs. Red-shifted chromophores for derivatization of hydroxyl groups are shown in Figure 4.12. For amino groups, the red-shifted

fluorescent fluorescent

NH2 HO

C13H27

OH D-erythro-sphingosine 19 50 μg

O N O

O N O

λmax 258 nm λem 370 nm

C13H27

HO OH

20

C13H27

O O λmax 234 nm λem 360 nm

O

O 21

Figure 4.11. By a selective two-step microscale chemical derivatization procedure, two different types of chromophores are introduced in D-erythro-sphingosine 19.

133

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

Δe 263 (+25) +20

N O O

chrom-1

O chrom-3

N O chrom-2

O

UV λmax 382 nm (ε 27,000)

+10

CD

O

UV λmax 382 nm (e 34,000) UV λmax 410 nm (e 37,000) RO OR

389 (+16) e × 10–4

Me2N

0 –10 –20

22a 412 274 (58,000) (63,000)

22b 8 455 (–25)

O O-Cin H OAc : 22a, R = H : 22b, R = chrom-3 ester

6 4

H

UV 2 200

300

400

500 λ (nm)

0

Figure 4.12. Red-shifted chromophores and application to taxinine. (Redrawn from reference 57, with permission.)

chromophores, such as Schiff bases and cyanine dyes, are also useful for exciton CD analysis, in particular, when it is desirable to avoid possible interference with other electronic transitions present in the substrate [56]. As shown in Figure 4.12, taxinine derivative α-glycol 22a exhibits an intense CE around 263 nm due to the π –π ∗ transition of the highly strained enone group. In the previous application of the ECM, unsubstituted benzoate chromophores were used where a negative exciton couplet was clearly observed despite the overlap with the enone CE [28a]. However later, to avoid the overlap, a red-shifted chromophore (chrom-3) was used for derivatization to yield ester 22b. As expected, the CD of 22b showed a clearer negative exciton couplet indicating a counterclockwise screw sense between the two hydroxyl groups in full agreement with the previous report of the AC [57].

4.6. THE USE OF PREEXISTING CHROMOPHORES IN NATURAL PRODUCTS FOR EXCITON COUPLING Some natural products already have one or two chromophores such as those shown in Figure 4.13, which can be used in exciton CD to determine their ACs.

Figure 4.13. Exciton CD chromophores found in natural products.

134

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a) Substituted Benzene and Polyacene Chromophores for the CD ECM As demonstrated in Section 4.4, the long-axis polarized 1 Bb transition of polyacene chromophores is ideally suited for observing exciton coupled CD, because of large ETDM. The UV data of polyacenes with D2h symmetry are shown in Figure 4.13. In the polyacene systems, there is no ambiguity for determining the long and short axes, and therefore the CD ECM leads to clear-cut conclusions. (b) Conjugated Dienes, Enones, Ene-Esters, Ene-Lactones, and DieneEsters as Exciton CD Chromophores The moieties in Figure 4.13 are useful chromophores for the CD ECM. The transition moment of their π –π ∗ band is almost parallel to the chromophoric long axis. (c) Natural Products with Two Chromophores Showing Exciton CDs The exciton coupling CD mechanism is applicable also to compounds already having two different chromophores, which exhibit long-axis polarized π –π ∗ transitions at different wavelengths—that is, i.e., nondegenerate system. The ACs of some natural products, such as those in Figure 4.14, were established by direct analysis of their CD spectra. In such cases the interaction of at least two preexisting chromophores leads to exciton split CDs. For abscisic acid (23), the opposite AC was once assigned, but it was later revised as shown by several studies. One was the application of exciton CD to the interaction between the enone and diene-carboxylic acid chromophores showing a positive couplet [58]. The case of quassin (24) is unique because of the exciton coupling between two identical preexisting α-methoxy-enone groups [59]. The AC of dendryphiellin F (25) was determined by exciton CDs from the interaction between diene and diene-carboxylate chromophores [60], while that of arnottin II (26) was determined from the dehydrotetralone and phthalide chromophoric interaction [61]. The AC of a derivative of vinblastine 27, a dimeric alkaloid, was originally determined by X-ray crystallography (Figure 4.15) [62]. Experimental and theoretical CD studies were carried out to clarify the CD mechanism of vinblastine and related compounds [63, 64]. The molecule 27 can be considered as containing two moieties, that of cleavamine 27a and that of vindoline 27b, both of which exhibit the intrinsic CD CEs due to their own local chirality [64]. In addition, an exciton-coupled through-space interaction between them takes place (Figure 4.15). To obtain the net exciton CD, the intrinsic CD bands were subtracted from the CD of 27. The resulting “difference CD” appeared as an intense positive couplet around 220 nm (Figure 4.15d), generated by the positive exciton chirality between 1 Bb transition moments of the indole and indoline chromophores at 225 nm and 215 nm, respectively, and by that confirming the (S ) AC at 16’-C.

Figure 4.14. Natural products with two preexisting chromophores showing exciton CD.

135

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(a)

(c)

27 λ (nm)

(b)

(d)

λ (nm)

Figure 4.15. (a) Vinblastine 27 consists of cleavamine 27a and vindoline 27b. (b) ETDMs of indole and indoline chromophores. (c) The sum CD spectrum (dotted line) = CD(27a) + CD(27b), and CD spectrum of 27. (d) Difference CD = CD(27) − {CD(27a) + CD(27b)}. (Redrawn from reference 64, with permission.)

4.7. SUPRAMOLECULAR APPROACH IN THE ECM: APPLICATION OF PORPHYRIN TWEEZERS The use of tetraarylporphyrins and their metal derivatives as CD chromophores has initiated a new supramolecular approach for the determination of the AC of chiral compounds containing a single stereogenic center and one site for chromophoric derivatization. This group includes various natural products carrying only a single functionality, such as secondary hydroxyl, primary or secondary amino, and carboxyl groups. These compounds are unsuitable for application of the conventional ECM where at least two intramolecularly interacting chromophores are required. The supramolecular approach employs a dimeric zinc porphyrin reagent 31, now commercially available as “Zn-tweezers,” which is capable of forming 1:1 host–guest complexes upon adding a solution of N,N -bidentate conjugate 30, prepared by reacting the chiral substrate 28 with an achiral carrier 29 as shown in Figure 4.16 [65]. The application of the porphyrin tweezers method to (S )-α-(2-naphthyl)ethanol is shown in Figure 4.16b. The formation of 1:1 sandwiched chiral host–guest complex 34

136

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

M N N Zn N N

H chiral XH substrate 28

M L X = O, NH

HO

M L

X

H N

N N Zn N N

NH2

Zn

L H O

:N H

P-1 HH N:

Zn

conjugate 30 (guest)

+ O

O

H

P-2 O

H N carrier 29

O O O O 1:1 host-guest complex 32

Zn porphyrin host 31 “tweezer”

NH2 (a)

+100 M H3C O

P-1

O :N H

P-2

P-1

P-2

H tweezer

M

L

M

L

HH N:

chiral conjugate 33 of (S)-abs. config

+50 Δε 0 –50

host-guest complex 34 –100

preferred conformation

433 nm (+91)

CD A = +170

e × 10–5

L

423 nm (–79)

15 10

422.2

416.6

UV/Vis 400

420

440 λ (nm)

5 0

(b)

Figure 4.16. Porphyrin tweezers method: (a) Preparation of bidentate conjugate 30 from chiral substrate 28 and carrier 29; Formation of a 1:1 host guest complex 32 with Zn-porphyrin tweezers 31. (b) Example of (S)-α-(2-naphthyl)ethanol conjugate 33: two conceivable conformations of complex 34; observed CD spectrum of 34 in methylcyclohexane. (Redrawn from reference 65b, with permission.)

from conjugate 33 proceeds under steric control and leads to positive first and negative second CEs in the Soret region. The origin of the intense exciton CEs is due to a preferred conformer with a clockwise interporphyrin twist, where the larger group L (2-naphthyl) protrudes from the binding pockets in order to avoid unfavourable steric interactions. The interporphyrin twist in the complex is thus dictated by the steric orientation of L (2-naphthyl) and M (methyl) at the stereogenic center of the substrate. When there is no ambiguity in the assignment of L and M groups, the sign of the couplet determines the AC at this center. Recently, a more reliable discrimination of preferred interporphyrin helicity of the host–guest complexes by theoretical analysis of both steric and electronic factors involved in stereocontrolled complexation process has become possible. This analysis relied on molecular mechanics calculations by Merck Molecular Force Field (MMFFs) or OPLS2005 approaches coupled to Monte Carlo-based conformational analysis and quantum mechanical treatment of free conjugates [66–68]. The porphyrin tweezers method is now well established and has allowed successful determinations of AC of some natural products, such as isotomenoic acid 35, an irregular diterpene [69], and bovidic acid 37, an 18-carbon hydroxyfuranoid acid [70] (Figure 4.17). More recently, other types of porphyrin-based tweezers have been developed. Structural changes in the tweezers, such as introduction of various substituents at the aryl groups and in the bridge between the two porphyrins, allow for tuning the complexation ability

137

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

O H OH O

s P-2

15

Zn Isotomenoic acid 35

H O N H

5

O O

MeO P-1 NH2 15′ HO Zn 5′

36

O O

O

H H O 6

OH

9

4

O

P-2 O 5

15

Zn Bovidic acid 37

6

H

N H

15′

N Zn H2

5

O O

P-1

O

5′

38

O O

Figure 4.17. Applications of the porphyrin tweezers method to natural products.

of the tweezers and extension of its application to other types of chiral substrates [67, 71, 72].

4.8. APPLICATION OF THE CD ECM: FUNDAMENTAL EXAMPLES As shown, there exist numerous applications of ECM. In the following, some selected examples of fundamental application are explained.

4.8.1. Exciton Coupling Between Polyacene and Related Chromophores Polyacenes (e.g., naphthalene and anthracene) are ideal chromophores for exciton coupling, because the long and short axes are clearly assigned, and the long-axis polarized transition has a large ETDM, yielding intense bisignate CEs. Chiral 1,1 -binaphthyls are typical atropisomers showing exciton CEs. For example, (S )-1,1 -binaphthyl-2,2 -dimethanol 39 shows intense positive first and negative second CEs (λext 231.3 nm, ε +342.4: λext 224.3 nm, ε −329.0; A = +671.4) in the longaxis 1 Bb transition (λmax 224.4 nm, ε 132,700) (Figure 4.18). The positive A value leads to an (S ) AC [1]. The exciton CD sign depends on the dihedral angle between two naphthalene planes. From the theoretical viewpoint, the exciton CD of 1,1 -binaphthalene compounds undergoes sign changes around 110–120◦ [73]. Therefore, in the application of the ECM, information of the dihedral angle is necessary. In most cases, the dihedral angle is distributed around 90◦ , which was supported by X-ray crystallography [74] and MO calculations. Chiral 1,1 :4 ,1 -ternaphthalene-2,2 -dimethanol (aS,aS )-(+)-40 with three naphthalene groups in positions of axial chirality is a unique atropisomer [75]. Enantiopure compound (+)-40 was prepared by the CSDP acid method. As shown in Figure 4.18, the CD spectrum of (aS,aS )-(+)-40 shows intense negative first and positive second CEs (λext 231.7 nm, ε −333.9: λext 223.4 nm, ε +225.4; A = −559.3) in the long-axis 1 B transition (λ b max 224.0 nm, ε 186,500). In Figure 4.18, chromophores 1 and 2 constitute a counterclockwise screw, and the same negative screw holds for chromophores 2 and 3, because of C2 -symmetry. However, chromophores 1 and 3 have no helicity because the long axes of naphthalene moieties 1 and 3 are almost parallel to each other. Thus the AC of (+)-40 was determined to be (aS,aS ) by the exciton CD method and confirmed by X-ray crystallography of the CSDP ester [76]. The dihedral angle between naphthalene groups ranges over −84.1◦ , −87.2◦ ,

138

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HOH2C

+300

+200 CH2OH

Δe

+100 Δe 0

CD

(S) HOH2C

CH2OH

–100 224.4 (132,700)

e × 10–4

+200 (S)-[CD(+)231.3]-39

0 CH2OH

chrom 2 –200

15

CH2OH

–200 10 224.3 (–329.0)

–300

5

–400

chrom 3

–400 chrom 1

UV 200

300 λ (nm) 400

0

obsd CD 285.4 ( +21.9) A = –559.3 231.7 (–333.9) 223.4 (+225.4)

e × 10–4

+400 231.3 (+342.4)

(aS,aS)-(+)-40

200

20

(aS,aS)-(+)-40 obsd UV 292.8 ( 24,000) 224.0 (186,500) 250

10

0 300 λ (nm) 350

Figure 4.18. CD and UV spectra of (S)-1,1 -binaphthyl-2,2 -dimethanol 39 (EtOH) and (aS,aS)-(+)-1,1 :4 ,1 -ternaphthalene-2,2 -dimethanol 40 (95% aq. EtOH). (Redrawn from references 1 and 75, respectively, with permission.)

−89.1◦ , and −89.4◦ as determined by the X-ray analysis. That is, three naphthalene chromophores are almost perpendicular to one another. Figure 4.19a shows the CD and UV spectra of (1R,1 S ,2S )-2,2 -spirobi[benz[e] indan]-1,1 -diol diacetate 41 [77]. The exo/endo configuration of two acetate groups was determined by 1 H NMR data showing nonequivalent four methylene protons. Racemic diacetate was enantioseparated by chiral HPLC to give an enantiomer [CD(–)230.2]-41 whose CD and UV spectra are shown in Figure 4.19. In the region of the naphthalene 1 Bb transition (λmax 224.8 nm, ε 172,700), the intense negative first and positive second CEs (λext 230.2 nm, ε −961.5 : λext 221.6 nm, ε +567.1; A = −1528.6) led to the (1R,1 S ,2S ) AC [77]. [6,6]-Vespirene 42 is a unique member of chiral 9,9 -spirofluorene compounds, and its chiroptical activity arises from the 9,9 -spirobifluorene system twisted by the side-chain bridge (Figure 4.19b) [78]. The AC of (−)-42 was determined to be R by the coupled oscillator analysis of its CD spectrum [78, 79]. However, the CD spectrum showed a complex pattern deviated from the ideal exciton bisignate CEs, implying that the 9,9 spirobifluorene chromophores of 42 may be strongly strained. Thus, it was questionable whether the CD ECM is applicable to such a strained system in a straightforward manner. To confirm the AC of compound 42, enantiopure compounds 43 and 44 with two anthracene and naphthacene chromophores, respectively, were prepared (Figures 4.20) [80]. These spiroaromatics are more suited than [6,6]-vespirene 42 for the ECM, because of their less strained structures and clear definition of exciton chirality between two polyacene chromophores. Enantiopure [6,6]-vespirene (−)-42 and compound (R)-(+)-43 were synthesized starting from diester (R)-(+)-45 (Figure 4.20). The UV spectrum of (+)-43 shows an intense 1 Bb transition (λmax 288.6 nm, ε 152,000), which is polarized along the chromophoric long axis. In the 1 Bb transition region, the CD shows intense positive first and negative second CEs (λext 300.5 nm, ε +551.0 and λext 278.5 nm, ε −560.7; A value = +1111.7). The present results unambiguously indicate that the long-axis electric transition moments of the two anthracene chromophores constitute a clockwise screw sense (i.e., positive exciton chirality), leading to the (R) AC [80]. In such spiroaromatics, there is the so-called spiro-conjugation—that is homoconjugation between p orbitals surrounding the spiro quaternary center. It is known

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(a)

(b)

41

λ (nm)

Figure 4.19. (a) CD and UV spectra of (1R,1 S,2S)-2,2 -spirobi[benz[e]indan]-1,1 -diol diacetate 41 in EtOH. (b) Chiral spiroaromatics. (Redrawn from reference 77, with permission.)

(a)

(b)

λ (nm)

Figure 4.20. (a) Synthesis of [6,6]-vespirene (R)-(−)-42 and chiral spiroaromatics (R)-(+)-43 and (R)-(+)-44 starting from diester (R)-(+)-45. (b) CD and UV spectra of (R)-(+)-43 in EtOH. (Redrawn from reference 80, with permission.)

139

140

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(5R,12R)-(–)

(5R,12R)-(–)-46

λ (nm)

Figure 4.21. CD and UV spectra of (5R,12R)-(−)-1,15-diethynyl-5,12-dihydro-5,12[1 ,2 ]benzenonaphthacene 46 in EtOH. (Redrawn from reference 84, with permission.)

that in some cases, such spiro-conjugation makes a dominant contribution to the CD spectra [81]. However, in the case of (+)-43, the spiro-conjugation effect is negligibly small, because the observed CD spectrum shows intense conservative bisignate CEs due to the exciton coupling. Bis(naphthacene) compound (R)-(+)-44 also showed intense positive first and negative second CEs leading to the (R)-AC, although the CD curve was complex due to vibronic structures. The AC determination of (R)-(+)-43 and (R)-(+)-44 confirmed the (R)-ACs of dimethyl ester (+)-45 and [6,6]-vespirene (−)-42. The (R)-AC of (−)-42 was later confirmed by X-ray crystallography [83]. (5R,12R)-(−)-1,15-Diethynyl-5,12-dihydro-5,12[1 2 ]benzenonaphthacene (46) is a triptycene derivative having one naphthalene and two phenylacetylene chromophores, with no conformational flexibility due to its cage structure and linear acetylene groups (Figure 4.21). The CD spectrum of (5R,12R)-(−)-46 shows strong exciton CEs (λext 245.5 nm, ε −138.2 : λext 215.0 nm, ε +113.6; A = −251.8) in the UV absorption region (λmax 241 nm, ε 75,000). Clearly the CEs originate from the coupling between the longaxis polarized 1 Bb transition (λmax 220.2 nm, ε 107,300) of naphthalene and the long-axis polarized 1 La transition (λmax 234.2 nm, ε 15,000) of phenylacetylene [84]. The long axes of the naphthalene and one phenylacetylene moiety constitute a negative exciton chirality. Similarly, a negative helicity is found between the naphthalene and the other phenylacetylene. On the other hand, the long axes of two phenylacetylenes are parallel to each other, indicating nil exciton chirality. Thus, the total exciton chirality is negative, leading to the (5R,12R) AC of (−)-46 [84, 85]. This absolute configurational assignment is in line with the chemical correlation results [84].

4.8.2. Application of the CD ECM to Acyclic 1,2-Glycols and Polyols The CD exciton chirality method has been applied to dibenzoates or bis(2-anthroates) of acyclic 1,2-glycols, which show typical bisignate CEs as exemplified in Figures 4.22

141

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

Δε R

ε x 10–4

273 (+97) +90 OChrom CD +60 Chrom = p-BrBz H OChrom A = +190 or 2-anthroyl +30 47 first CD, (+) 0 second CD, (–) Exciton Chirality: –30 H3C OChrom zero –60 OChrom OChrom H OChrom 258 –90 253 OChrom H R OChrom ChromO 15 (147,000) Chrom = (–93) R H 2-anthroyl H H H H H 10 (S)-48 R H H OChrom [47A] [47B] UV 5 in CH3CN J (trans) = 6.8 ~ 8.4 Hz, [47C] J (gauche) = 3.6 Hz 0 200 250 300 350 λ (nm)

Figure 4.22. Applications of the CD exciton method to acyclic terminal 1,2-glycols. (Redrawn from reference 87, with permission.)

50

λ (nm)

51

λ (nm)

Figure 4.23. Application of the CD ECM to acyclic internal 1,2-glycols with threo-configuration. (Redrawn from reference 86.)

and 4.23 [86, 87]. Acyclic dibenzoates or bis(2-anthroates) can rotate around the bond connecting two benzoates or 2-anthroates, and hence the CD sign depends on the conformational equilibrium. Based on the exciton CD and the conformational analysis by 1 H NMR, the AC of acyclic 1,2-glycols can be determined (Figures 4.22 and 4.23).

142

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

In terminal acyclic 1,2-glycols, diester 47 {bis(p-Br-benzoate) [86] or bis(2anthroate) [87]} can adopt three rotational conformers 47A, 47B, and 47C (Figure 4.22); here conformer 47B is unstable due to two gauche relations among three bulky groups. In contrast, because conformers 47A and 47C have only one gauche relationship between two bulky groups, they are more stable and dominate the equilibrium. The stable conformer 47A has a positive exciton chirality between two chromophores, while in conformer 47C the two chromophores are trans, and hence no exciton chirality is generated. Thus the CD of diester 47 reflects the positive exciton chirality of 47A. The 1 H NMR coupling constants {J (trans) = 6.8–8.4 Hz, J (gauche) = 3.6Hz} support this conclusion. The CD of (S )-1,2-propanediol bis(2-anthroate) 48 shows intense exciton CEs, from which the AC was assigned (Figure 4.22) [87]. Similarly, the CD ECM is applicable to internal 1,2-glycols with threo-configuration (Figure 4.23) [86, 87]. For example, diester 49 {bis(p-Br-benzoate) or bis(2-anthroate)} adopts three rotational conformers 49A, 49B, and 49C, in which conformers 49B and 49C are unstable because of three gauche relationships. Conformer 49A with two gauche relations between bulky groups dominates the equilibrium. Conformers 49A and 49B have positive and negative exciton twists between two chromophores, respectively, while in conformer 49C two chromophores are in the transrelationship, and therefore no exciton chirality is generated. The preference of 49A is supported by the large 1 H NMR coupling constant {J (trans) = 6.1–8.7 Hz}. After all, the CD spectrum of diester 49 reflects a positive exciton chirality of conformer 49A. For example, (2S,3S )-2,3-butanediol bis(p-Br-benzoate) 50 exhibits bisignate CEs of positive exciton chirality and an 1 H NMR coupling constant {J (trans) = ∼6.1 Hz}, from which the AC can be determined (Figure 4.23) [86]. The above relationship between the AC and exciton CD CEs holds for most internal 1,2-glycols. However, if a glycol has polar or extremely bulky groups (R1 and R2 ), the equilibrium is changed; groups R1 and R2 adopt a trans-relation to diminish the electric repulsive force or steric repulsion, and conformer 49B becomes dominant. The preference of 49B is supported by 1 H NMR coupling constant {J (gauche) = 2.9–4.1 Hz}. The CD spectrum of (2R,3R)-diethyl tartrate bis(p-Br-benzoate) 51 shows a negative exciton couplet reflecting the preference of 49B [86], where the ethyl ester groups are trans as confirmed by 1 H NMR coupling constant {J (gauche) = 2.9 Hz}. If the groups R1 and R2 are identical, the 1 H NMR vicinal coupling constant between two methine protons cannot be measured because of the same chemical shifts. In such cases, the 13 C satellite band method is useful to determine the Jvic value [86, 88]. In erythro-1,2-glycols, the determination of AC is more difficult. When the two groups R1 and R2 are identical, the glycol is a meso-isomer and hence achiral. If they are different, the glycol is chiral. The exciton CD CEs of erythro-diester are weak, and they depend on the equilibrium of the rotational conformations. Thus assignment of ACs requires further analysis by other methods. Noteworthy, the AC of 1,3-acyclic polyols can be determined much easier and in more straightforward manner due to a better discrimination between CD of 1,3syn/1,3-anti diols bis(p-Br-benzoates) or bis(p-methoxycinnamates). An acyclic anti 1,3-bis(acylate) adopts a planar zigzag form in its most stable conformer and exhibits a typical CD exciton couplet corresponding to the sign of the screw sense between the two gauche-oriented chromophores, while the syn-analogue also in most stable zigzag conformation has almost parallel p-Br-benzoate or p-methoxycinnamate chromophores that exhibit negligible coupling [89–91].

143

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

4.8.3. Application of the ECM to Compounds Containing a Single Suitable Chromophore: Nondegenerate System The exciton coupled CD is observable even in a nondegenerate system composed of two different chromophores, which undergo different electronic excitations. The first example of the nondegenerate exciton CD used for determining the AC of a chiral compound is the case of chromomycin A3 (52), (Figure 4.24). The AC of 52 was determined in the very early stage of the development of the CD ECM [92]. The first strategy for determining the AC of 52 was to apply the dibenzoate chirality method to glycol derivative (53). However, benzoylation of glycol 53 yielded monobenzoate 54 due to the steric hindrance. This unexpected result led to a general protocol to apply the CD ECM to nondegenerate systems [92]. The CD spectrum of monobenzoate 54 showed intense bisignate CEs (λext 270 nm, ε −19.9: λext 230 nm, ε +16.8; A = −36.7), which are stronger than those of glycol 53 (λext 270 nm, negative: λext 220 nm, positive) as seen in Figure 4.24 [1, 2]. The CD data of 54 imply that the bisignate CEs originate from the interaction between the long-axis transition of benzoate (λmax 230 nm, ε 14,000) and the long-axis 1 Bb transition of naphthalenoid chromophore (λmax 270 nm, ε 57,200), where the naphthalenoid 1 Bb transition is red-shifted due to the conjugation with a carbonyl group. If this mechanism is true, the following can be expected. If a proper benzoate group is introduced whose long-axis transition is close to the naphthalenoid 270 nm, the exciton CD would be enhanced, because the exciton theory tells us that the exciton coupling is more effective when two transitions are close to each other in energy. Hence the p-methoxybenzoate chromophore (λmax 256 nm, ε 18,000) was chosen to yield monobenzoate 55, the CD of which showed much intense bisignate CEs as expected (λext 271 nm, ε −70.6: λext 250 nm, ε +34.0; A = −104.6) (Figure 4.24). This result not only verified the exciton mechanism of nondegenerate systems, but also established the absolute configurational assignment. Since the first CE is negative and the second positive, the long axes of two chromophores constitute a counterclockwise screw, leading to the AC as shown in Figure 4.24 [2, 92]. This assignment was later confirmed by chemical correlation [93].

+40

250 (+34.0) 55

R = p-MeO-C6H4CO- HO

H

OMe OH O OH

+20 Δe 0

OH OH O 54

R = C6H5CO-

–20

H

MeO OH 52

MeO

MeO

R1

OMeO

H

OR OH O

O O

53, R = H 54, R = Bz 55, R = p-MeOBz

Me

–40

53

R=H in EtOH

–60

271 (–70.6)

–80 200

300

λ (nm)

MeO MeO H

H O

O H H O O OH O Me Me O Me H

54, R1 = H 55, R1 = OMe

400

Figure 4.24. Preparation of benzoate 54 and p-methoxybenzoate 55 of chromomycin A3 derivative: CD spectra of p-methoxybenzoate, benzoate, and alcohol derivatives. (Redrawn from reference 1, with permission.)

144

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OR

+40 R=H λext 226 nm, Δe +14.8

+30

H R = Bz λext 235 nm, Δe +34.0 220 nm, Δe –14.3

+20 +10 Δe 0

MeO 56, R = H 57, R = Bz

in EtOH

Figure 4.25. CD spectra of

Me O O

–10

17-dihydroequilenin 3-methyl ether (56) and 17-benzoate (57). (Redrawn from reference 92, with

17

–20

H 300

λ (nm)

MeO

400

57

permission.)

To confirm the nondegenerate exciton method further, the exciton method was applied to equilenin derivative 17-benzoate (57) (Figure 4.25) [2, 92]. The CD of benzoate 57 shows intense bisignate CEs around 230 nm (λext 235 nm, ε +34.0: λext 220 nm, ε −14.3; A = +48.3), while alcohol 56 exhibits a positive CE (λext 226 nm, ε +14.8). It is thus clear that the exciton CD of benzoate 57 originates from the interaction between long-axis 1 Bb transition of naphthalenoid chromophore (λmax 230 nm) and the long-axis transition of benzoate (λmax 230 nm). The positive exciton couplet agrees with the AC as shown in Figure 4.25. These results corroborated the methodology of nondegenerate exciton coupling. The nondegenerate ECM is applicable also to the conjugated diene–benzoate system as exemplified in Figure 4.26, where the CD and UV of 3-methylenecholest-4-ene-6β-ol benzoate (58) is shown (λext 242.0 nm, ε −30.8: λext 225.0 nm, ε +39.1; A = −69.9) [1]. The long-axis polarized π –π ∗ transition of the s-trans diene appears at 234 nm (ε 20,000), which matches the long-axis polarized benzoate transition at 230 nm (ε 15,300). These ETDMs constitute a counterclockwise screw generating negative first and positive

(a)

(b)

225.0 (+39.1) +40

58 H

H +20 Δε 0

CD

in EtOH

OBz

H

58

e × 10–4

200

–20

242.0 (–30.8) 235.0 (33,800)

–40

300

OBz-p-Cl 59 Me

O O

2

UV 200

4

λ (nm)

0

H O

H

CD in MeOH λext 247 nm, Δe –24.4 λext 224 nm, Δe +23.0 A = –47.4

H

Figure 4.26. (a) CD and UV spectra of 3-methylenecholest-4-en-6β-ol benzoate (58). (b) CD data of 3-oxocholest-4-en-6β-ol p-chlorobenzoate (59). (Redrawn from reference 1, with permission.)

145

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

second CEs. The exciton coupling between conjugated diene and benzoate chromophores determine the AC of diene–alcohol compounds. The nondegenerate ECM was applied to the conjugated enone–benzoate system, 3-oxocholest-4-en-6β-ol p-chlorobenzoate (59) (λext 247 nm, ε −24.4 : λext 224 nm, ε +23.0; A = −47.4) (Figure 4.26) [59]. The π –π ∗ transition of conjugated enone (λmax 241 nm, ε 16,600) couples with the p-chlorobenzoate transition (λmax 240 nm, ε 21,400) to generate negative first and positive second CEs, stemming from the counterclockwise screw sense.

4.8.4. UV λmax Separation of Two Different Chromophores versus Exciton CD As described above, the exciton coupling is the most effective in a degenerate system having two identical chromophores. On the other hand, it is interesting to know how the exciton CD decreases, when the difference between UV λmax values of two chromophores increases. To clarify the effect of UV λmax separation on the exciton CD, we carried out the CD calculation by the exciton theory and also synthesized steroidal model compounds having two different p-substituted benzoate chromophores (Figures 4.27). Figure 4.27a shows the CD calculation result how the exciton CD decreases in intensity, when the separation between two λmax values increases [1]. The important aspect is that even when two chromophores undergo excitations at 230 nm and 310 nm, respectively, their interaction provides observable bisignate CEs, whose signs are governed by the exciton chirality between two ETDMs. The observed CD spectra of steroidal model compounds agree well with the calculated curves. Thus the ECM is applicable to nondegenerate systems, in which two λmax values are much separated. Based on this

223.0 (+15.7)

R=H

230 nm–230 nm

(1)

300 λ (nm) 350

200

238.5 (–15.2) 300 λ (nm)

200

200

250

300

R = Cl 350

230 nm–250 nm

(3) 200

300 230 nm–260 nm

(4) 200

300 230 nm–280 nm

(5) 200

250

200

300

350

229.5 (+14.5)

R = OMe

350 200

230 nm–257 nm 300

250 249.5 (–10.4)

350

350

O O

350

250

230 nm–240 nm

242.3 (–18.2)

O O

232.5 (+11.2)

230 nm–310 nm

(6)

350

227.0 (+16.8)

230 nm–240 nm

(2)

230 nm–230 nm

R = NMe2 R 230 nm–310 nm 350

(a)

200

250

312.0(–3.7)

350

(b)

Figure 4.27. (a) Calculated exciton CD curves, when changing λmax of two chromophores. (b) Observed CD spectra of cholest-5-ene-3β,4β-diol 3-benzoate 4-p-substituted benzoates in EtOH. (Redrawn from reference 1, with permission.)

146

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

mechanism, the CD allylic benzoate method was developed, as will be discussed in the next section.

4.8.5. Allylic Benzoate Exciton Method for Determining the AC of Allylic Alcohols

e × 10–4

0 CD

Δε –5

229.5 (–8.72) –10

(a)

229.0 (+11.45) +10

2 BzO

228.9 (16,400)

H 60

(b)

OBz

BzO CD

+5 Δε 0

61

1

60

200

250 λ (nm) 300 (a)

BzO 0

2

228.5 (13,100)

1 UV

e × 10–4

The chiroptical method for determining the AC of allylic alcohols was first developed as the Mills’ rule [94] and the Brewster’s benzoate rule [95], in which the AC was correlated with optical rotation. Later the benzoate sector rule was proposed by Harada and Nakanishi to correlate the AC with CD CEs at 230 nm, where the position of the C=C double bond against the benzoate chromophore was a key factor to govern the CD CE [96]. As an extension of this concept and based on the nondegenerate exciton CD mechanism, the allylic benzoate exciton method was developed for determining the AC of allylic alcohols. Figure 4.28a shows the CD and UV spectra of cholest-4-en-3β-ol benzoate 60, where a negative CE is seen at the long-axis polarized UV band (230 nm) [1, 97]. The mechanism of this CE is interpreted as follows. The C=C double bond undergoes an allowed π –π ∗ transition around 190 nm polarized along its long axis. This transition couples with the long-axis transition of benzoate chromophore at 230 nm; and by this nondegenerate exciton coupling, the negative first CE is observed around 230 nm. On the other hand, it is difficult to observe the expected positive second CE, because the π –π ∗ band of C=C double bond locates below 200 nm. Since the long axes of two chromophores constitute a counterclockwise screw sense, the benzoate CE becomes negative as the first CE of the non-degenerate exciton coupling. Another stereoisomer, cholest-4-en-3α-ol benzoate 61, exhibits a positive CE around 230 nm, reflecting the positive exciton chirality between benzoate and C=C double bond chromophores (Figure 4.28b). Note that the CD intensity of axial allylic benzoate is generally larger than that of equatorial allylic benzoate, as exemplified in Figure 4.28 [1, 97]. Allylic benzoate 60 undergoes a negative CD below 210 nm, which is opposite to the expected positive second CE around 190 nm. A similar phenomenon is observed in the case of allylic benzoate 61. These results may originate from the participations of the

H 61

UV 200

250 λ (nm) 300

0

(b)

Figure 4.28. (a) CD and UV spectra of cholest-4-en-3β-ol benzoate 60 in EtOH. (b) CD and UV spectra of cholest-4-en-3α-ol benzoate 61 in EtOH. (Redrawn from references 97 and 1, respectively, with permission.)

147

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

benzenoid 1 B transition (around 200 nm) of benzene chromophore and the weak π –σ ∗ transition of the C=C double bond chromophore. Despite such complexity, the 230-nm CE reflects solely the allylic benzoate chirality, which is useful for determining the AC of allylic alcohols [1, 97]. More recent studies led to a significant advance in application of allylic benzoate method to homoallylic alcohols, amines, and other enes by the use of microscale cross metathesis [98, 99]. By applying the microscale cross-metathesis protocol the double bond can be transformed into more suitable for coupling chromophores, such as styrene or p-substituted analogues, so the entire CD couplet can be observed and used for an AC determination. The introduction of fluorescent styrene provides an additional benefit, since the CD can be measured in emission by the more sensitive fluorescent detected CD (FDCD) method [98b, 99].

4.9. RECENT APPLICATIONS OF THE CD EXCITON CHIRALITY METHOD The following are the recent interesting application examples of the CD ECM for determining the ACs of chiral natural and synthetic compounds. (1) ACs of Ciguatoxin and Related Compounds Intense exciton coupled CDs are useful for comparison of CD data with those of a reference compound as exemplified below. The AC of C5 in CTX4A (62), a ciguatoxin precursor, was determined by the use of the exciton coupled CD spectra as summarized in Figure 4.29 [100]. The AB ring fragment (64) of CTX4A was stereoselectively synthesized, and its p-bromobenzoate 65 exhibits the exciton split CEs. It is obvious that these CEs originate from the exciton coupling between conjugated diene in the C5 side chain and C11 p-bromobenzoate, and these two chromophores constitute a clockwise screw sense (Figure 4.29). The CD spectrum of CTX4A tris(p-bromobenzoate) 63 also shows the exciton split CEs, which was interpreted as follows. Although there are six possible interactions among four chromophores in 63, the coupling between 1,3-diene and C11 p-bromobenzoate makes a dominant contribution, because the remaining ones are weak due to the remote distances. This interpretation was confirmed by the CD data of 65. The ACs of CTX4A and ciguatoxin (CTX) were thus determined by comparison of the exciton CD CEs of 63 and 65. The ACs of the CTXs were later confirmed by total syntheses [101].

Me CTX4A (62), R = H 63, R = OBz-Br-p

H 5

O

H

H

A 11 O O H H ORH H

H

Me H OR O H 47 O O O M O H H H H O H Me Me H Data of 63, CD (MeOH) λext 246 nm (Δe +32) 230 nm (Δe –28)

H Me RO O H HO

O O H H

32

H

H

A 5

O H H

O

H O

11

H OR

O

64, R = H 65, R = OBz-Br-p Data of 65, CD (MeOH) λext 242 nm (Δe +25) 225 nm (Δe –14)

Figure 4.29. ACs of ciguatoxin precursor CTX4A (62) and related compounds, along with their CD data.

148

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(2) AC of Urothion The AC of acyclic terminal 1,2-glycol was determined by the ECM as follows. To determine the AC of urothion (66), a yellow pteridine pigment isolated from human urine, the compound was subjected to desulfurization yielding 67, which was converted to tris(p-chlorobenzoate) 68 (Figure 4.30) [102]. An authentic sample of (R)-67 was synthesized starting from d-glucose. Since the [α]D values of diols (S )-67 and (R)-67 were too small to assign their ACs by comparison, tris(p-chlorobenzoates) (S )-68 and (R)-68 were prepared and their CD spectra were compared. The CD spectrum of (R)-68 was opposite in sign to that of (S )-68, thus leading to the (R)-AC of urothion 66. The bisignate CEs at 247 and 228 nm are mainly caused by the exciton coupling between the two benzoate groups in the side chain. According to the ECM applied to acyclic terminal 1,2-glycols (Section 4.8.2), the positive first CE leads to the S configuration, which agrees with that obtained by comparison of CD spectra. It should be noted that the intense exciton CEs are thus useful for determining the AC by comparison of chiroptical data. (3) AC of Cephalocyclidin A, a Five-Memberd Ring cis-α-Glycol The ECM was applied to the five-membered ring 1,2-cis-glycol system. The unprecedented pentacyclic structure of cephalocylidin A (69) was elucidated by X-ray crystallographic, 1 H-NMR, and CD methods [103]. To determine the AC, 2,3-bis(p-methoxycinnamate) (70) was synthesized, which showed bisignate CEs of negative exciton chirality (Figure 4.31). The observed exciton CD is weak, reflecting the small dihedral angle of cis-glycol moiety in a five-membered ring. The AC of compound 69 was determined as shown. It should be advised that in cases with small dihedral angle between two hydroxyl groups, the use of more symmetrical chromophore such as p-dimethylaminobenzoate or p-bromobenzoate would be suitable for obtaining unambiguous results.

O HN H2N

N

H OR

O SCH3

N N

HN

OH

S H

RHN

OR

N N

N

OH

(S)-68, CD (EtOH), λext 247 nm (Δe +8.0) 228 nm (Δe –3.0) UV (EtOH) λmax 241 nm (e 43,600)

(S)-67, R = H (S)-68, R = Bz-p-Cl

urothion (R)-66

Figure 4.30. Urothion and exciton CD data.

O

O

H H 3 OO 2 OHO

OH

HO

2 3

HO O 69

MeO N MeO

70

N O OO

Compound 70, CD (CH3OH) λext 325 nm (Δε –6.6), 283 nm (Δε +5.8)

O O 9 O 1 H

OAcO

HO H

NOEs

Compound 71, CD (CH3CN) λext 270.7 nm (Δe +20.3) 227.8 nm (Δe –18.1)

71

Figure 4.31. ACs of cephalocyclidin A 69 and dihydro-β-agarofuran sequiterpene 71 as determined by the CD ECM.

149

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(4) AC of Dihydro-β-Agarofuran Sequiterpene Dihydro-β-agarofuran sequiterpene 71 is unique because the natural product itself contains cinnamate and benzoate chromophores at C1 and C9 in an ideal 1–3 transrelationship for exhibiting exciton CEs (Figure 4.31) [104]. The CD spectrum of 71 shows positive first and negative second CEs, from which the AC was unambiguously determined as shown. (5) AC of Strevertenes This is an interesting example of the exciton CD due to the interaction between conjugated pentaene and p-dimethylaminobenzoate chromophores [105]. The relative stereostructures of strevertenes A (72) and G (73), antifungal macrolides, were determined by X-ray crystallography of compound 73, where the pentaene moiety adopts all transconfiguration and almost planar conformation as shown in Figure 4.32. To determine the AC of compound 72, the CD ECM was applied as follows. As the UV λmax of the pentaene chromophore is located around 330 nm, p-dimethylaminobenzoate (λmax = 311 nm) was selected as an exciton coupling partner with a red-shifted absorption suitable for coupling with the pentaene moiety. Thus, methyl ester 74 was benzoylated to give a mixture of mono-benzoates, which was separated by reverse-phase HPLC. All possible six mono-benzoates were isolated, and the esterification positions were assigned by 1 H NMR spectra. Of these benzoates, 15-p-dimethylaminobenzoate (75) was suitable for application of the ECM, because the allylic position generates a clear exciton chirality (Figure 4.32). The difference CD spectrum {CD(75) − CD(74)} between 15-benzoate 75 and alcohol 74 is shown in Figure 4.32, where the exciton coupling between pentaene and p-dimethylaminobenzoate chromophores generates intense positive first and negative

336 (+37.6)

O

OH

HO

O

R1 OR2

Strevertene A (72): R1 = COOH, R2 = H methyl ester (74): R1 = COOMe, R2 = H derivative (75): R1 = COOMe, R2 = Bz-p-NMe2

C6

C4

C3

C5

Δε

H16 O

C14 H15

C7

+145°

+20 N

0 Diff. CD = CD(75) – CD(74)

in MeOH

C10

C9

354 (+18.8)

O

C11 C8

C12 O7 C29 C13 C1 C14 O9 O2 C16 O1 C20 C18 C22 C15 C24 C27 C17 C19 O8 C25 C23 C21 C31 C26 C28

H3CO2C

O6

O5

O4

O3 C2

+40

–20

303 (–21.4) UV(75) UV(74)

–40

e × 10–5

OH OH OH

320

336 353

1

in MeOH

C30

Strevertene G (73) : R1 = CH2OH, R2 = H 200

250

300 λ (nm)

350

400

Figure 4.32. ACs of strevertenes A (72) and G (73), X-ray crystallographic stereoview of (73), difference CD spectrum {CD(75) − CD(74)}, and UV spectra of (75) and (74). (Redrawn from reference 105, with permission.)

150

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

second CEs. The positive A value indicates that the long axes of both chromophores constitute a clockwise screw sense as shown in Figure 4.32, where the trans-antiperiplanar relation between H-15 and H-16 was assigned by the 1 H NMR coupling constant (J15,16 = 9.2 Hz). The ACs of 15-benzoate 75 and natural products 72 and 73 were thus determined [105]. (6) AC of Phomopsidin The example shows an exciton CD due to the interaction between remote diene-ester and p-nitrobenzoate chromophores. The CD spectrum of phomopsidin 76, a marinederived fungal metabolite, shows only one very weak CE at 266 nm associated with the diene-carboxylic acid chromophore at 6-C with an intense UV absorption at 266 nm (Figure 4.33a) [106]. To determine the AC, alcohol 76 was converted to p-nitrobenzoate 77, which exhibited bisignate CEs due to the exciton coupling between diene-ester and p-nitrobenzoate. From the positive first CE, the AC having a clockwise helicity was determined. (7) AC of Spiroxin A, a Bis-Acetophenone Fungal Metabolite In this application, red-shifted chromophores were used to prevent the overlap of CEs. Spiroxin A (78) is a bis-acetophenone with a spiroketal moiety that locks the two conjugated chromophores (Figure 4.33b) [107]. The relative stereochemistry had previously been established by NMR. Spiroxin A 78 exhibits a complex CD in the 200 to 280 nm region, which was difficult to interpret. To determine the AC, two phenolic hydroxyl groups were esterified with retinoic acid because the UV band of all-trans retinoic acid ester (λmax 356 nm, ε 39,500) is well red-shifted from the absorption of the existing chromophore. However, while spiroxin A bis(retinoate) (79) did not exhibit useful exciton couplet in the retinoic ester region, the difference CD between bis(retinoate) 79 and spiroxin A 78 provided a clear-cut negative exciton couplet that allowed for the AC assignment [107]. (8) AC of Pinellic Acid The CD allylic benzoate method was applied to pinellic acid 80, a long-chain allylic alcohol, to determine the AC as shown in Figure 4.34 [108]. The relative configuration COOR1

(a)

O

(b) 76, R1 = R2 = H 77, R1 = CH3, R2 = p-NO2Bz

ax

OR

spiroxin A 78, R = H CI spiroxin A bis(retinoate) 79, R =

O

O

77

H

CD (MeOH), λext 271 nm (Δε +22.4) λext 244 nm (Δε –7.9) UV (MeOH), λmax 264 nm (ε 35,000)

6 11

eq H R2O

O

O O OR O

Difference CD = CD (79) – CD (78), λext 385 nm (Δε –17.3), 331 (+17.4)

Figure 4.33. (a) CD and UV data of phomopsidin methyl ester p-nitrobenzoate 77. (b) Spiroxin A 78 and CD data.

OR4 12 9 13 R1O 10 OR3 pinellic acid 80, R1 = R2 = R3 = R4 = H O

OR2

81, R1 = CH3, R2 = p-BrBz, R3 & R4 = acetonide

Figure 4.34. Pinellic acid 80 and exciton CD.

81, J9,10 = 7.0 Hz CD (CH3OH), λext 245 nm (Δe +6.9), 221 (+2.13), 209 (+5.7)

151

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

of three chirality centers was determined by NOE experiments of methyl ester-acetonide, from which the syn configuration of the 12-C/13-C vicinal diol was assigned. The acetonide was then converted to p-bromobenzoate 81, the 1 H NMR spectrum of which indicated an antiperiplanar relationship between the 9 and 10 protons (J9,10 = 7.0 Hz). The CD of the allylic p-bromobenzoate showed a positive CE at λext 245 nm (ε +6.97), indicative of the S configuration at 9-C. This predicted the AC of pinellic acid to be either (9S,12S,13S ) or (9S,12R,13R). The remaining question was solved by a stereospecific synthesis of both diastereomers. The comparison of spectral data with those of the authentic samples indicated that the AC of the natural product was (9S,12S,13S) as shown. (9) AC of Phorboxazole The CD allylic benzoate method was applied to determine the AC of allylic alcohol moiety as follows. The AC of phorboxazole A (82) was determined as shown by total synthesis except for the configuration of 38-C (Figure 4.35) [109]. Although the AC at the 38-C allylic alcohol had originally been assigned as R by application of the Mosher MTPA method, there was an anomaly in the NMR δ data. The threo and erythro model compounds (83 and 84) were synthesized from (S )-malic acid, and hence the ACs of 33C, 35-C, and 37-C were assigned as shown. To determine the AC of 38-C by the allylic exciton method, these alcohols were converted to 2-naphthoate esters (85 and 86). The NMR coupling constants J38,39 = 9.6 Hz for 85 and J38,39 = 9.2 Hz for 86 indicate that these two protons are in trans-relationship in their stable conformations. Threo ester 85 exhibited a negative CE at λext 234 nm (ε −9.2), indicating a negative twist between the naphthoate and the allylic double bond. In contrast, erythro product 86 showed a positive CE at λext 234 nm (ε +15.1) indicating a positive helicity. Thus the ACs of 83 and 84 were determined as shown. The comparison of the NMR coupling constant J37,38 of 82 with those of 85 and 86 led to the AC of natural product 82 as shown in Figure 4.35 [109]. (10) AC of Gymnocin-B The experiments for the AC determination of gymnocin-B 87, a cytotoxic marine natural product with the largest 15-rings polyether skeleton isolated so far, are very revealing. They demonstrate the effectiveness of porphyrin chromophore to serve as chiroptical probe for AC of remote stereogenic centers residing in a very flexible substrate, available only in a few hundred micrograms (Figure 4.36) [99, 110b]. The compound has two hydroxyl groups at 10-C and 37-C, but it was difficult to clearly assign the helicity between the two remote and sterically hindered OH groups, because of the conformational flexibility arising from the presence of seven-membered rings.

OH

O

phorboxazole A 82, J37,38 = 7.9 Hz

N O

OCH3 OCH3 Br

35

39 38

37

H OH

O

OH

O

O

N O

O O

OCH3

83, R = H n-C4H9

39 38

84, R = H

35 37

H OR

O

n-C4H9 OMe

39

38

OCH3 35 37

H O OMe OR

threo-85, R = 2-naphthoyl J37,38 = 7.0 Hz, J38,39 = 9.6 Hz

erythro-86, R = 2-naphthoyl J37,38 = 3.7 Hz, J38,39 = 9.2 Hz

CD (CH3CN) λext 234 nm (Δe –9.18)

CD (CH3CN) λext 235 nm (Δe +15.1)

Figure 4.35. Phorboxazole A 82 and allylic benzoate method.

152

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OR H H 37 RO H O Me H H H H H O H 10 O O O H I H J H K C D E F G B O O H O L O H A H H O O H O Me Me 5 Me H H H H Me H O M Me N O H O H O O 55 Gymnocin B 87, R = H H 54 O bis(TPP-cinnamate) 88 NHN O NHN R= CD (MeOH)

29°

15 10

10′

20 5

15′ 20′

5′

C5/Ph rotation

10-TPPcin K (ax) J

λext 419 nm, Δe +11 λext 414 nm, Δe –15

37-TPPcin (eq)

28.7 Å

Figure 4.36. Gymnocin B 87, the lowest-energy conformation of its 10,37-bis(TPP-cinnamate) derivative 88 obtained by Monte Carlo/MMFF94s, and CD data of 88. (Redrawn from reference 110b, with permission.)

Triphenylporphyrin-cinnamate chromophores were introduced into 10-OH and 37-OH by acryloylation/cross-metathesis under microscale conditions. The obtained diester 88 showed a clear exciton split CD, positive first and negative ˚ apart (Figure 4.36). After extensecond CEs, even though the two porphyrins are ∼30 A sive conformational analysis of the derivative by the MMFF94s/Monte Carlo calculation, the (10S,37S) AC was assigned to the derivative 88. The CD curve of 88 calculated by the DeVoe’s coupled oscillator method for the Boltzmann-weighted conformers agreed well with the observed CD spectrum. The AC of gymnocin B 87 was thus determined as shown [99, 110b]. (11) AC of Axially Chiral Binaphthoquinones The exciton CEs of 1,2 -binaphthyl derivative are well suited for AC determination than the CEs of binaphthoquinone as shown below. The AC of (–)-8 -hydroxyisodiospyrin 89, a naturally occurring bi(naphthoquinone), was determined by the synthesis of the opposite enantiomer (S )-(+)-89 (Figure 4.37) [111]. The AC of a synthetic intermediate was determined by X-ray analysis, and the intermediate was then converted to binaphthalene 90, with its CD showing intense exciton coupled CEs. From the positive sign of the couplet, an S configuration was assigned to 90, which was further converted to (S )-(+)89. The AC of the natural product was determined to be (R)-(−)-89. Thus the ACs of these compounds were confirmed by the CD ECM [111]. The CD spectrum of (S )-90 shows intense exciton CEs, while (S )-(+)-89 exhibits two positive and one negative CEs around 360–260 nm, but their ε values are much smaller than those of 90, and the CD curve deviates from the ideal pattern of the exciton coupling. Thus, to determine the ACs by the exciton method, it is important to select the

O

OMe OMe

OH

(S)-(+)-89 (S)-90 MeO MeO MeO

OMe OMe

225 nm (Δε –60.0)

CD (1,4-dioxane)

O O

CD (CH3CN) λext = 239 nm (Δε +46.7), HO

λext = 356.9 nm (Δε +6.9), O

298.5 nm (Δε +11.2), 263.8 nm (Δε –23.2)

OH

Figure 4.37. ACs and CD data of axially chiral binaphthalene and binaphthoquinone. Numerical CD data were obtained from the spectra reported in reference 111.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

most appropriate chromophores—that is binaphthalene rather than binaphthoquinone as exemplified in this case. (12) AC of Pre-anthraquinones The example shows an intense exciton coupling between naphthalene–ketone and anthraquinone. The ACs of atropisomeric pigments 91 and 92 were determined by spectroscopic methods including ECM (Figure 4.38) [112]. Atropisomer 91 exhibits intense negative first and positive second CEs at 272 nm and 251 nm, respectively, thus leading to an AC with negative helicity between long axes of two aromatic chromophores. The AC at the 3 position was deduced by chemical correlation; the reductive cleavage of 93 yielded (R)-torosachrysone methyl ether with known AC. Pigment 92 exhibits weak bisignate CEs compared to 91, because the two aromatic chromophores are connected by two σ bonds, and hence the dihedral angle between two long axes is close to 180◦ . (13) AC of Spiroleptosphol Figure 4.39 shows an interesting example of triol tribenzoate (97) where the AC has been successfully determined in the presence of three exciton interactions. Spiroleptosphol (93), a γ-methylidene-spirobutanolide, exhibited cytotoxicity against P388 murine leukemia and HeLa human cervix carcinoma (Figure 4.39) [113]. The relative stereostructures of compound 93 and spiroleptosphol C (94) were determined by NMR spectroscopy and/or by X-ray crystallography [114]. Compound 93 was converted to dibenzoate 96, which exhibited typical exciton CEs of negative exciton chirality leading to a counterclockwise helicity between two benzoate chromophores at the 6- and 7-positions (Figure 4.39). The AC of 93 was thus unambiguously determined as shown [114]. Triol 93 was also converted to tribenzoate 97, the CD of which showed intense positive first and negative second CEs [114]. These CD data are interpreted as follows: The exciton chirality between 4- and 6-benzoates is clockwise, and that between 4- and 7-benzoates is also clockwise, while that between 4- and 6-benzoates is anticlockwise. The total exciton chirality becomes thus positive and leads to positive first and negative second CEs, confirming the AC of triol 93. (14) AC of Leucettamol A, α,ω-Bifunctionalized Sphingolipid The intense exciton CEs are useful for amplifying the chiroptical properties of natural products with very weak optical rotation as exemplified below, where the deconvolution ECM was also useful for AC determination. Leucettamol A (98) was isolated as a marine natural product exhibiting a variety of biological activities. The relative configuration of 2-amino-3-hydroxy end groups was determined to be erythro by NOE experiments of a

OMe OH O

OMe OH O

(–)-92

(+)-91 3'

3'

MeO

CD (MeOH), λext 272 nm (Δε –159.3) 251 nm (Δε +158.8) UV (EtOH) λmax 274 nm (log ε 4.50)

OH O HO

OH O

OMe

MeO

OH O HO

OH O

OMe

Figure 4.38. Pre-anthraquinones and CD data.

CD (CHCl3), λext 286 nm (Δε +14.4) 257 nm (Δε –8.2) UV (EtOH) λmax 278 nm (log ε 4.36)

153

154

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O

O 4

O HO

12

OH 17 R

6 HO 7

spiroleptosphol (93), R = H spiroleptosphol C (94), R = OH

O RO

4

O

OH

O O

6

RO 7

O (95), R = H (96), R = Bz

4 6

(97)

BzO 7

O

H

O O

6

OBz

O

7 H O

OH R O

CD (MeOH) of (96) λext 238 nm, Δe –18.9 222 nm, Δe +10.9

O O

H

(96), R = side chain

(93)

O BzO

H 6

H4O R O

7 H O

CD (MeOH) of (97) λext 238 nm, Δe +38.4 220 nm, Δe −18.1 204 nm, Δe −6.4

(97), R = side chain

Figure 4.39. Determination of the absolute stereochemistry of spiroleptosphol (93) by the CD ECM. The CD data of (97) were obtained from the reported figure.

bis-oxazolone derivative. Compound 98 was originally assigned to be racemic because it did not exhibit any measureable optical rotation [115]. To reinvestigate the AC of leucettamol A 98, it was catalytically reduced to give perhydro-derivative 99, which was then converted to N,N ,O,O  -tetrabenzoyl derivative 100. The CD spectrum of compound 100 showed bisignate CEs due to the exciton coupling between benzoate/benzamide chromophores (Figure 4.40) [116]. Since the 2,3-N,Odibenzoyl chromophores are remote from 28,29-N,O-dibenzoyl chromophores, the simple additivity of exciton coupled CD is applicable. Erythro-101, selected as a model and prepared from (2S )-alanine, exhibited a negative exciton couplet (Figure 4.40) [117]. The simulated CD curve of (ent-erythro-101 + ent-erythro-101) agreed well with the observed CD of (erythro / erythro)-100 (Figure 4.40), but other combinations—for example (threo-102 + threo-102) and (ent-erythro-101 + threo-102)—disagreed with the observed CD of 100. The ACs of 100 and hence 98 were unequivocally determined to be (2R,3S,28S,29R). Thus leucettamol A is optically active; redetermination of the specific rotation of 98, averaged over 10 measurements, gave [α]D = −3.8±0.1 (c 4.4,

OH

NH2

NHBz 3

OH leucettamol A (2R,3S,28S,29R)-(–)-98 NHR

OR

99: R = H

100: R = Bz

NH2 OR

OBz erythro-101

CD (MeOH) λext = 235 nm, Δe = –5.6 λext = 220 nm, Δe = +1.6

NHR NHBz

erythro l erythro-100: obsd CD (MeOH), λext = 238 nm, Δe = +10.3, λext = 222 nm, Δe = –2.8 simulated CD based on (ent-erythro-101 + ent-erythro-101): λext = 235 nm, Δe = +11.3, λext = 220 nm, Δe = –3.2

OBz threo-102

CD (MeOH) λext = 237 nm, Δe = +3.0 λext = 221 nm, Δe = –3.5

Figure 4.40. ACs and CD data of leucettamol A (98) and related compounds.

155

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

MeOH). The deconvolution exciton CD method is thus useful for determining the AC of dimeric sphingolipids. (15) AC of 11-Deoxydiaporthein A by the Exciton Allylic Benzoate Method The AC of allylic alcohol was determined by ECM and X-ray analysis as shown below. 11-Deoxydiaporthein A (103) was isolated from a marine fungus [118], and its relative stereostructure was determined by NMR. To determine its AC, compound 103 was converted to benzoate (104), the CD spectrum of which showed a positive CE (λext = 230 nm, ε +5.8), reflecting a clockwise helicity between the long axes of benzoate and olefin chromophores (Figure 4.41a). The R configuration at 7-C was determined by the CD allylic benzoate method. The AC determined by CD was confirmed by X-ray crystallography of a single crystal of compound 103 obtained by recrystallization from chloroform. Interestingly, the crystal contained chloroform molecules as crystal solvent, and based on the strong anomalous scattering effect of chlorine atoms, the AC of 103 was established [118]. (16) AC of Cortistatin A This is an example of a natural product whose AC was determined by the exciton coupling of its two preexisting trans-diene and isoquinoline chromophores. Cortistatin A (105), an anti-angiogenic steroidal alkaloid, was isolated from a marine sponge. The relative stereostructure of 105 was elucidated by 2D-NMR (COSY and NOESY) and confirmed by X-ray as shown in Figure 4.41b. Interestingly, compound 105 itself exhibited bisignate CEs of negative exciton chirality, leading to the AC as shown [119]. Compound 105 has a conjugated diene (UV λmax = ∼234 nm) and a C17 isoquinoline (UV λmax = ∼220 nm), which couple to generate exciton CEs. The data indicate that the long axes of the two chromophores constitute a counterclockwise screw in the conformation shown in Figure 4.41, leading to the AC shown. (17) AC of trans-Acenaphthene-1,2-diol The ECM is applicable to glycols with 120◦ dihedral angle as follows. Chiral transacenaphthene-1,2-diol (106) and related compounds were prepared by baker’s yeastmediated reduction, and the AC of (−)-trans-106 was determined to be (1S,2S) by the CD ECM [120]. Diol (−)-106 (97% ee) was converted to bis(p-dimethylaminobenzoate) (+)-107, which exhibited positive first and negative second CEs. The negative exciton

OH

H

O 7 OR

OH OH

11-deoxydiaporthein A (103), R = H AC of (103) by X-ray of crystal (103)/CHCI3 (a)

(104), R = Bz CD λext = 230 nm, Δe = +5.8

OH

11 17

HO O N

N

H H N

H

cortistatin A (105) relative structure of (105) by X-ray

H CD of (105) λext 237 nm, Δe = –17 217 nm, Δe = +35

17S

(b)

Figure 4.41. (a) AC of 11-deoxydiaporthein A (103) as determined by the CD allylic benzoate and X-ray methods. (b) AC of cortistatin A (105) as determined by X-ray and ECM methods.

156

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CD couplet led to the (1S,2S) AC as shown in Figure 4.42a. When reporting the CD data, the intensity should be provided in ε units, not in raw θ /m◦ units, otherwise it is impossible to judge the intensities of the observed CEs. (18) AC of Kolokoside A Aglycone The ECM is useful for the AC determination of cyclic 1,2-glycols as exemplified below. The AC of kolokoside A (108), a triterpenoid glycoside, was determined by the CD ECM (Figure 4.42b) [121]. The aglycone was esterified to give 2,3-bis(pdimethylaminobenzoate) (109), whose CD spectrum showed negative first and positive second. Thus the (2R,3R) AC of the aglycone was determined. (19) AC of Dinemasone B This is an example of a cyclic trans-1,2-glycol. The AC of dinemasone B (110), a bioactive metabolite from fungi, was determined by the CD ECM [122]. Glycol 110 was converted to bis(p-bromobenzoate) (111), which exhibited a negative exciton CD. The AC of 110 was thus assigned as shown in Figure 4.42c. (20) AC of Oligonaphthalenes The following are unique chiral oligonaphthalenes whose ACs were determined by porphyrin–porphyrin coupling at very long distances (Figure 4.43) [123]. These chiral oligomers were synthesized from chiral (S )-binaphthyl derivative by repeating the oxidative coupling giving the diastereomeric products—that is, 4-mers, 8-mers, and 16-mers. For example, the coupling reaction of (S,S,S )-4-mer yielded diastereomeric (S,S,S,S,S,S,S )-8-mer and (S,S,S,R,S,S,S )-8-mer. To determine their ACs, the CD ECM using tetraphenylporphyrin (TPP) carboxylic acid was applied, because the Soret band (λmax = 420 nm) of TPP-carboxylic ester is far separated from the naphthalenoid UV bands. It has been previously reported that the exciton CD method using TPP-esters was applicable to remote hydroxyl groups, where the interchromophoric distance R ranges ˚ [47, 48]. around 50 A Two terminal phenol groups of diastereomers were esterified to yield bis(TPP-ester) 112 and 113. When applying the exciton method to these systems, it is important to elucidate the dihedral angle between naphthalene moieties. The X-ray crystallographic

N

N

N

H

HOOC O

O

O

H

O

O O

O

O O

(1S,2S)-(+)-trans-107, 97%ee CD (MeOH)

N

H

CD (CH3CN)

(a)

OH O

O O O

109

Kolokoside A aglycone bis(p-dimethylaminobenzoate)

λext 329 nm, postive λext 308 nm, negative

O

O

Br 111

Br Dinemasone B, bis(p-bromobenzoate) CD (CH3CN)

λext 320 nm, Δe = –22 λext 296 nm, Δe = +13

λext 252 nm, Δe = –45.3 λext 235 nm, Δe = +22.9

(b)

(c)

Figure 4.42. Application of ECM and ACs: (a) (1S,2S)-(+)-trans-acenaphthene-1,2-diol; (b) kolokoside A aglycone; (c) dinemasone B.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(S,S,S,S,S,S,S,S,S,S,S,S,S,S,S)-114,

(S,S,S,S,S,S,S,S,S,S,S,S,S,S,S)-114, UV (CHCI3) λmax 420.0 nm, e = 756,000 CD (CHCI3) λext 434 nm, Δe = –2.8 418 nm, Δe = +2.7

n = 7, R = 66 Å HN N

H N

N O

O

OO

OO

OO O

O

(S,S,S,S,S,S,S)-112, n = 3 UV (CHCI3) λmax 420.0 nm, e = 725,000 CD (CHCI3) λext 428 nm, Δe = –10.1 418 nm, Δe = +12.6

n (S,S,S,S,S,S,S)-112, n = 3 O

O

O

O O

OO

O

OO

O

O O

OO

O

O

O

(S,S,S,R,S,S,S)-113, n = 3 UV (CHCI3) λmax 420.0 nm, e = 755,000 CD (CHCI3) λext 428 nm, Δe = +10.7 418 nm, Δe = –11.1

Figure 4.43. Exciton CD data and the AC of oligonaphthalenes.

analyses of some derivatives and the CONFLEX-MM2 calculation indicated the average dihedral angle to be around 89.4◦ ∼90.0◦ , that is a right angle. Thus, it is predicted that if bis(TPP-ester) 8-mer takes an (S,S,S,S,S,S,S ) configuration, the dihedral angle between two terminal TPP-ester groups becomes −90◦ . That is, it is calculated 90◦ × 7 = (360◦ × 2) − 90◦ . Thus a negative exciton couplet is predicted. If bis(TPP-ester) 8-mer takes an (S,S,S,R,S,S,S ) configuration, the dihedral angle between two terminal TPP-ester groups becomes +90◦ , because 90◦ × 6 − 90◦ = (360◦ ) + 90◦ . That is a positive exciton couplet is predicted. As listed in Figure 4.43, bis(TPP-ester) 8-mer 112 exhibited negative first and positive second CEs, while bis(TPP-ester) 8-mer 113 exhibited an opposite couplet. Thus the observed CD curves are mirror images of each other. Thus, an (S,S,S,S,S,S,S ) configuration was assigned to 112 and an (S,S,S,R,S,S,S ) configuration was assigned to 113. The ECM method is applicable to a more remote diol diester—for example bis(TPP-ester) 16-mer 114, where two chromophores are separated at ∼66 ˚ but still gave an observable exciton CD. From the sign of the first CE, an A, (S,S,S,S,S,S,S,S,S,S,S,S,S,S,S ) AC was assigned to 16-mer 114 (Figure 4.43) [123]. (21) AC Assignment of Acetylene Alcohols by the CD ECM The combination of Sonogashira reaction and ECM enables the AC determination of chiral terminal acetylene alcohols as shown below. Enantiopure acyclic acetylene alcohols are employed as chiral synthons for bioactive compound syntheses. It is generally difficult to determine the ACs of acyclic acetylene alcohols, but the 1 H NMR anisotropy method using MαNP acid ester has been developed for enantioresolution of racemic alcohols and simultaneous determination of AC. The CD ECM is also useful for determining ACs of acetylene alcohols. A method for determining the ACs of terminal acetylene alcohols was developed from the exciton CD due to the interaction between p-methoxyphenylacetylene (λmax 252 nm) and p-methoxybenzoate (λmax 257 nm) [124]. For example, acetylene alcohol (+)115 was converted by a Sonogashira reaction to p-methoxyphenylacetylene alcohol (−)116, which was esterified giving benzoate (−)-117 (Figure 4.44). The CD spectrum of (−)-117 showed exciton bisignate CEs (Figure 4.44). The negative A-value led to an (R)-AC for (−)-117, consistent with the result from the 1 H NMR anisotropy method. This method is thus useful for determining the ACs of terminal acetylene alcohols [124].

157

158

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+20 CD

H OH

(R)-(–)-117

–10 –20

obsd CD 266.0 (–27.9) 246.2 (+21.0)

–30

A = –48.9 UV

e x 10–4

+10 Δe 0

300

(R)-(–)-116

CH3O

OCH3 252 nm

obsd UV 258.2 (41,500)

250

(R)-(+)-115

4

H O

O

(R)-(–)-117

2 CH3O

in EtOH 200

H OH

257 nm

350

Figure 4.44. A scheme for the AC determination of terminal acetylene alcohol by combination of the Sonogashira reaction and the CD ECM: CD and UV spectra of 1-(4-methoxyphenyl)1-dodecyn-3-ol 4-methoxybenzoate (R)-(−)-117 in EtOH. (Redrawn from reference 124, with permission.)

(22) ACs and Exciton CD of Unique 1,3-Diethynylallene Compounds The ACs of allene compounds can be determined by the ECM as exemplified below. Allene compounds are unique chiral synthons devoid of chirality centers, and they have been employed for the syntheses of various chiral compounds. Allene (M )-(−)-118 is one of such compounds, the AC of which was determined as follows. Compound (−)-118 was converted to 1,3-bis{(4-dimethylaminophenyl)ethynyl}allene (−)-119, with typical intense exciton split CEs (Figure 4.45) [37, 125]. The negative A value indicated that two identical chromophores of (4-dimethylaminophenyl)ethynyl olefin constitute a counterclockwise sense, and thus an (M )-AC was assigned to allene (−)-119. It should be noted that the dihedral angle between two (4-dimethylaminophenyl)ethynyl groups is 90◦ due to the allene skeleton, which satisfies the requirement of an ideal exciton CD mechanism. The CD spectral curves of (M )-(−)-119 as calculated by TDDFT and π -electron SCF-CI-DV MO methods were in good agreement with the observed spectrum confirming the above assignment by the ECM [37]. These absolute configurational assignments were consistent also with the X-ray analysis and chemical correlation results [125].

H i-Pr3Si (M)-(–)-118

N N

(M)-(–)-119 (e.r. > 91:9) CD (hexane) λext 328 nm, Δε = –77.9 287 nm, Δε = +37.2

(M)-(–)-119

Figure 4.45. Exciton CD of 1,3-diethynylallene derivative.

159

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

4.10. EXAMPLES REQUIRING CAUTION AND THEORETICAL ANALYSIS OF EXCITON CD The following are examples, where cautious considerations of CD were necessary. That is, the observed CD spectra could not be interpreted by the exciton coupling mechanism in a straightforward manner. (1) AC of Antitumor Antibiotic AT2433-A1 with a Secondary Amino Group To determine the AC of antitumor antibiotic AT2433-A1 (120), amino sugar bis(p-Brbenzoyl) derivative (121∗ ) was derived from the natural product (Figure 4.46). Its CD spectrum showed a negative couplet leading to the AC as shown [126]. However, it was later pointed out that this assignment was wrong as explained below [50]. To clarify the reasons for the wrong assignment, the authentic samples 121 and 122 were synthesized from a starting material with known AC. Surprisingly, the CD of 121 showed a weak positive exciton couplet, while that of 122 showed a strong negative one. The 1 H NMR of 121, a benzamide derivative of the secondary amine, indicated the presence of (Z ) and (E ) amide isomers, where the exciton chirality is negative and positive, respectively. Since the CD contributions cancel in some extent, the sign and intensity of observed CD are governed by those of the prevailing (E ) amide. On the other hand, the CD of the primary amine derivative 122 reflects straight the AC because of its (Z ) conformation. Therefore, when the ECM is applied to secondary amines, the analysis of (E ) and (Z ) conformations is critical. The AC of the secondary amine in natural product 120 was confirmed by the total synthesis of a related compound [50]. (2) Anomalous CD CEs of 1,1 -Biphenanthryl Compounds Enantiopure 2,2 -dimethoxy-1,1 -biphenanthryl (aR)-(+)-123 (Figure 4.47) was synthesized by oxidative coupling of 2-phenanthrol, followed by enantioresolution, where its AC was determined by the axial chirality recognition method [127]. It was expected that

CH3 N O

O

O

CI OH

O O HN CH3 OH

p-BrBzN CH3

N

N H O

OH OCH3

OCH3

O

p-BrBzO p-BrBzN

p-BrBzN Me

OBz-p-Br

R

121*(AC originally assigned) CD (CH3CN), λext 249 nm (negative CD) 219 nm (positive CD)

AT2433-A1 (120)

O H Ar N O Ar Me O H

Ar O H OCH3

121, (Z )-amide negative chirality

O Ar

H

N O Me O H

O

Ar H Ar

OCH3

121, (E)-amide positive chirality

O H N O H O H

OCH3

OBz-p-Br

121, R = Me CD (CH3CN), λext 251 nm (Δe +5.5) 219 nm (Δe –4.4)

O H OCH3

122, (Z )-amide negative chirality

122, R = H CD (CH3CN), λext 252 nm (Δe –33.4) 234 nm (Δe +12.6)

Figure 4.46. Application of the ECM to secondary amine. Numerical CD data were obtained from the spectra reported in reference 50.

160

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+200

(aR)-(+)-123 +100 CD Δe

(aR)-(+)-125

OCH3

OCH3

(aR)-(–)-124

OCH3

OCH3

e x 10–5

0

(aR)-(–)-124

(aR)-(+)-125

–100 2 OCH3 –200

200

OCH3

UV

250

1

300

350 λ (nm)

0

(aR)-(+)-123

(aR)-126

Figure 4.47. Chemical correlation of 1,1 -biphenanthryl derivative (aR)-(+)-123 and related compounds, along with their CD and UV spectra. (Redrawn from reference 127, with permission.)

the CD spectrum of (aR)-(+)-123 should show a negative exciton couplet around the 1 Bb transition of phenanthrene chromophore, because the two long axes of phenanthrene groups constitute an anticlockwise screw sense as illustrated in Figure 4.47. However, compound (aR)-(+)-123 showed positive and negative CEs at 274 nm 258 nm, respectively, disagreeing with the expectation. The anomalous CD results led to the reinvestigation the ACs of (+)-123 and related compounds. Starting from (aR)-(+)-1,1 -binaphthyl-2,2 -diol, compound (aR)-(−)-124 was synthesized, and it was easy to convert (aR)-(−)-124 to the target compounds (aR)-(+)-123 and (aR)-(+)-125 with a binaphthyl chromophore (Figure 4.47). The CD spectrum of (aR)-(+)-125 showed intense exciton CEs (λext 239.8 nm, ε −197.1; λext 227.8 nm, ε +133.3; A = −330.4) reflecting a negative helicity between two long axes of naphthalene moieties, which confirmed the (aR)-AC of (+)-125. Compound (aR)(−)-124 also exhibited exciton CEs (λext 249.6 nm, ε −112.6; λext 236.4 nm, ε +60.5; A = −173.1), but the A value decreased to about half. That is, the conjugation of the naphthalene chromophore with a double bond diminished the exciton CD intensity, because the corresponding ETDM deviated from the long axis of the naphthalene chromophore [127]. In contrast, compound (aR)-(+)-123 exhibited positive and negative CEs (λext 274.2 nm, ε +38.5; λext 258.0 nm, ε −52.5; A = +91.0) (Figure 4.47). These results confirmed that the CEs of 1,1 -biphenanthryl derivative (+)-123 are opposite in sign to those of 1,1 -binaphthyl derivative (+)-125 despite the same ACs. Furthermore, the CD shape of (+)-123 is complex, and the A value is about 1/4 compared to that of (aR)-(+)125. Additionally, more intense negative and positive CEs were observed around 220 nm. To gain insight into the anomalous CD behavior of compound (aR)-(+)-123, the CD and UV spectra of 1,1 -biphenanthryl (aR)-126 were calculated by the π -electron SCF-CI-DV MO method (for this MO method, see Chapter 5, this volume). The UV

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

calculation of phenanthrene chromophore revealed its complexity due to the presence of another electronic transition polarized along the short axis around 255 nm in addition to the intense long axis-polarized 1 Bb transition around 260 nm. Therefore, in (aR)-126, four ETDMs interact with one another, giving rise to positive first and negative second CEs. The simulated CD and UV spectra agreed well with the observed; the (aR)-AC of (+)-123 was confirmed also by the MO calculation [127]. It should be emphasized that the CD ECM itself is correct, but the electronic transitions of compounds (aR)-(+)-123 and (aR)-126 are complex, and therefore the simple and qualitative application is not valid for these compounds. In general, symmetrical chromophores (e.g., linear polyacenes such as naphthalene, anthracene, etc.) are more suitable for the CD ECM than the less symmetrical ones such as nonlinear condensed aromatics (e.g., phenanthrene).

4.11. CONCLUSION As discussed above, the CD ECM is useful for determining the ACs of various chiral compounds. The exciton coupling between two or more chromophores generates exciton split and intense bisignate CEs that reflect the helicity between ETDMs (positive or negative exciton couplet). The AC of the compound can be unambiguously determined from the sign of the couplet. In general, the exciton couplet CDs due to through-space chromophoric interaction are much stronger than the CEs of isolated chromophores, such as those due to ketone n → π ∗ , benzenoid π → π ∗ , and conjugated diene or enone π → π ∗ transitions. Their unique bisignate shapes facilitate the recognition of an exciton couplet. The CD ECM is readily proved by the quantum mechanical exciton theory as described, and therefore it is classified as a nonempirical method. Thus the ACs of chiral compounds can be determined by the exciton CD method without any reference compound. It was established that both X-ray Bijvoet and CD exciton chirality methods give the same AC, although assignments are based on totally different phenomena. However, the unambiguous determination of AC by ECM requires a very careful selection of the appropriate chromophores. It is critical that this selection takes into account not only the structural and conformational features of the chiral substrate, but also other basic requirements of this method. We hope that this chapter clarifies the main aspects of the CD exciton chirality method and provides useful guidelines for its application in stereochemical analysis.

ACKNOWLEDGMENTS The authors thank the co-workers in these studies, whose names are listed in references, and Dr. George A. Ellestad, Department of Chemistry, Columbia University, for his valuable suggestions. We are very grateful to JASCO Co., for their continuing instrumental and technical support.

REFERENCES 1. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy—Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, and Oxford University Press, Oxford, 1983. 2. N. Harada, K. Nakanishi, Acc. Chem. Res. 1972, 5 , 257–263.

161

162

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3. N. Berova, K. Nakanishi, Exciton chirality method: Principles and applications, in Circular Dichroism: Principles and Applications, K. Nakanishi, N. Berova, and R. Woody, eds., John Wiley & Sons, New York, 2000, pp. 337–382. 4. N. Berova, N. Harada, K. Nakanishi, Electronic spectroscopy: Exciton coupling, theory and applications, in Encyclopedia of Spectroscopy & Spectrometry, J. Lindon, G. Tranter, and J. Holmes, eds., Academic Press, London, 2000, pp. 470–488. 5. N. Berova, G. A. Ellestad, N. Harada, Modern methods in natural products chemistry: characterization by circular dichroism spectroscopy, in Comprehensive Natural Products Chemistry II, L. Mander and H. W. Liu, eds., Elsevier, Oxford, 2010, Chapter 9, pp. 91–146. 6. S. Superchi, E. Giorgio, C. Rosini, Chirality 2004, 16 , 422–451. 7. N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. 8. (a) L. M. Pasteur, C. R. Acad. Sci. Paris 1848, 26 , 535–538. (b) L. M. Pasteur, Anal. Chim. Phys. 1848, 24 , 442–4459. 9. See also: (a) H. D. Flack, Acta Cryst. 2009, A65 , 371–389. (b) Z. S. Derewenda, Acta Cryst. 2008, A64 , 246–258. 10. J. H. van’t Hoff, Arch. Neerl. Sci. Exactes Nat. 1874, 9 , 445–454. 11. J. A. Le Bel, Bull. Soc. Chim. Fr. 1874, [2], 22 , 337–347. 12. See also E. L. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds, John Wiley & Sons, New York, 1994, pp. 1–10. 13. (a) A. Cotton, Compt. Rend . 1895, 120 , 989–991. (b) A. Cotton, Ann. Chim. Phys. 1896, 8 , 347. 14. E. Fischer, Ber. Dtsch. Chem. Ges. 1891, 24 , 2683–2687. 15. V. L. Rosenfeld, Z. Physik 1928, 52 , 161–174. 16. (a) W. Kuhn, Z. Physik. Chem. 1929, B 4, 14–36. (b) W. Kuhn, Trans. Faraday Soc. 1930, 26 , 293–308. 17. (a) E. U. Condon, W. Alter, H. Eyring, J. Chem. Phys. 1937, 5 , 753–775. (b) E. U. Condon, Rev. Modern Phys. 1937, 9 , 432–457. (c) W. J. Kauzman, J. E. Walter, H. Eyring, Chem. Rev . 1940, 26 , 339–407. 18. (a) J. G. Kirkwood, J. Chem. Phys. 1937, 5 , 479–491. (b) W. W. Wood, W. Fickett, J. G. Kirkwood, J. Chem. Phys. 1952, 20 , 561–568. 19. (a) A. S. Davydov, Zhur. Eksptl. i Teoret. Fiz . 1948, 18, 210–218. (b) A. S. Davydov, Theory of Molecular Excitons, translated by M. Kasha, M. Oppenheimer, Jr., McGraw-Hill, New York, 1962. 20. (a) J. M. Bijvoet, A. F. Peerdeman, A. J. van Bommel, Nature 1951, 168 , 271–272. (b) J. M. Bijvoet, A. F. Peerdeman, A. J. van Bommel, Proc. K. Ned. Acad. Wet. B 1951, 54 , 16–19. (c) A. J. van Bommel, Proc. K. Ned. Acad. Wet. B 1953, 56 , 268–271. (d) J. Trommel, J. M. Bijvoet, Acta Crystallogr. 1954, 7 , 703–709. (e) J. M. Bijvoet, A. F. Peerdeman, Acta Crystallogr. 1956, 9 , 1012–1015. 21. (a) W. Moffitt, J. Chem. Phys. 1956, 25 , 467–478. (b) W. Moffitt, D. D. Fitts, J. G. Kirkwood, Proc. Natl. Acad. Sci. USA 1957, 43 , 723–730. 22. (a) I. Tinoco, Jr., Advan. Chem. Phys. 1962, 4 , 113–160. (b) I. Tinoco, Jr., R. W. Woody, D. F. Bradley, J. Chem. Phys. 1963, 38 , 1317–1325. (c) I. Tinoco, Jr., Radiat. Res. 1963, 20 , 133–139. (d) I. Tinoco, Jr., C. A. Bush, Biopolym. Symp. 1964, 1 , 235–250. (e) W. C. Johnson, I. Tinoco, Jr., Biopolymers 1969, 8 , 715–731. 23. (a) J. A. Schellman, P. Oriel, J. Chem. Phys. 1962, 37 , 2114–2124. (b) P. M. Bayley, E. B. Nielsen, J. A. Schellman, J. Phys. Chem. 1969, 73 , 228–243. 24. (a) S. F. Mason, Proc. Chem. Soc. 1962, 362–363. (b) S. F. Mason, G. W. Vane, J. Chem. Soc. B 1966, 370–374. 25. A. F. Beecham, A. C. Harley, A. McL. Mathieson, J. A. Lamberton, Nature 1973, 244 , 30–32.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

26. S. F. Mason, G. W. Vane, K. Schofield, R. J. Wells, J. S. Whitehurst, J. Chem. Soc. B 1967, 553–556. 27. S. H. Wilen, J. Z. Qi, P. G. Williard, J. Org. Chem. 1991, 56 , 485–487. 28. (a) N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1969, 91 , 3989–3991. (b) M. Koreeda, N. Harada, K. Nakanishi, Chem. Commun. 1969, 548–549. (c) N. Harada, K. Nakanishi, Chem. Commun. 1970, 310–311. (d) N. Harada, H. Sato, K. Nakanishi, Chem. Commun. 1970, 1691–1693. (e) N. Harada, S. Suzuki, H. Uda, K. Nakanishi, J. Am. Chem. Soc. 1971, 93 , 5577–5579. (f) S.-M. L. Chen, N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1974, 96 , 7352–7354. (g) N. Harada, S.-M. L. Chen, K. Nakanishi, J. Am. Chem. Soc. 1975, 97 , 5345–5352. 29. (a) J. Tanaka, F. Ogura, H. Kuritani, M. Nakagawa, Chimia 1972, 26 , 471–473. (b) J. Tanaka, K. Ozeki-Minakata, F. Ogura, M. Nakagawa, Nature (London), Phys. Sci . 1973, 241 , 22–23. (c) J. Tanaka, K. Ozeki-Minakata, F. Ogura, M. Nakagawa, Spectrochim. Acta, Part A 1973, 29 , 897–924. 30. Y. Saito, Kagaku-to-Kogyo (Japan) 1973, 26 , 128–134. 31. H. H. Brongersma, P. M. Mul, Chem. Phys. Lett. 1973, 19 , 217–220. 32. S. F. Mason, J. Chem. Soc., Chem. Commun. 1973, 239–241. 33. (a) N. Harada, Y. Takuma, H. Uda, J. Am. Chem. Soc. 1976, 98 , 5408–5409. (b) N. Harada, Y. Takuma, H. Uda, Bull. Chem. Soc. Jpn. 1977, 50 , 2033–2038. (c) N. Harada, Y. Takuma, H. Uda, Bull. Chem. Soc. Jpn. 1978, 51 , 265–270. (d) N. Harada, Y. Takuma, H. Uda, J. Am. Chem. Soc. 1978, 100 , 4029–4036. 34. P. J. Stephens, N. Harada, Chirality 2010, 22 , 229–233. 35. D. Bohm, Quantum Theory, Prentice-Hall, New York, 1951, p. 427. 36. E. Giorgio, K. Tanaka, W-D. Ding, G. Krishnamurthy, K. Pitts, G. A. Ellestad, C. Rosini, N. Berova, Bioorg. Med. Chem. 2005, 13 , 5072–5079. 37. J. L. Alonso-Gomez, A. G. Petrovic, N. Harada, P. Rivera-Fuentes, N. Berova, F. Diederich, Chem. Eur. J . 2009, 15 , 8396–8400. 38. (a) N. Sakabe, K. Sakabe, K. Ozeki-Minakata, J. Tanaka, Acta Crystallogr. Sect. B 1972, 28 , 3441–3446. (b) J. Tanaka, C. Katayama, F. Ogura, H. Tatemitsu, M. Nakagawa, Chem. Commun. 1973, 21–22. 39. D. H. R. Barton, H. T. Cheung, A. D. Cross, L. M. Jackman, M. Martin-Smith, J. Chem. Soc. 1961, 5061–5073. 40. I. C. Paul, G. A. Sim, T. A. Hamor, J. M. Robertson, J. Chem. Soc. 1962, 4133–4145. 41. S. Hosozawa, N. Kato, K. Munakata, Tetrahedron Lett. 1974, 15 , 3753–3756. 42. N. Harada, H. Uda, J. Am. Chem. Soc. 1978, 100 , 8022–8024. 43. D. Rogers, G. G. Unal, D. J. Williams, S. V. Ley, G. A. Sim, B. S. Joshi, K. R. Ravindranath, J . Chem. Soc., Chem. Commun. 1979, 97–99. 44. (a) N. Zhao, P. Zhou, N. Berova, K. Nakanishi, Chirality 1995, 7 , 636–651. (b) D. Rele, N. Zhao, K. Nakanishi, N. Berova, Tetrahedron 1996, 52 , 2759–2776. (c) N. Zhao, N. Berova, K. Nakanishi, M. Rohmer, P. Mougenot, U. J. Jurgens, Tetrahedron 1996, 52 , 2777–2788. 45. X. Huang, K. Nakanishi, N. Berova, Chirality 2000, 12 , 237–255. 46. G. Pescitelli, S. Gabriel, Y. Wang, J. Fleischhauer, R. W. Woody, N. Berova, J. Am. Chem. Soc. 2003, 125 , 7613–7628. 47. S. Matile, N. Berova, K. Nakanishi, J. Fleischhauer, R. W. Woody, J. Am. Chem. Soc. 1996, 118 , 5198–5206. 48. S. Matile, N. Berova, K. Nakanishi, Chemistry and Biology 1996, 3 , 379–392. 49. A. Mammana, G. Pescitelli, T. Asakawa, S. Jockusch, A. G. Petrovic, R. R. Monaco, R. Purrello, N. J. Turro, K. Nakanishi, G. Ellestad, M. Balaz, N. Berova, Chem. Eur. J . 2009, 15 , 11853–11866.

163

164

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

50. J. D. Chisholm, J. Golik, B. Krishnan, J. A. Matson, D. L. Van Vranken, J. Am. Chem. Soc. 1999, 121 , 3801–3802. 51. (a) F. Kazmierczak, K. Gawronska, U. Rychlewska, J. Gawronski, Tetrahedron: Asymmetry 1994, 5 , 527–530. (b) J. Gawronski, F. Kazmierczak, K. Gawronska, P. Skowronek, J. Waluk, J. Marczyk, Tetrahedron 1996, 52 , 13201–13214. (c) J. Gawronski, M. D. Rozwadowska, F. Kazmierczak, Polish J. Chem. 1994, 68 , 2279–2288. 52. M. Watanabe, T. Abe, Y. Fujii, N. Harada, Enantiomer 1998, 3 , 197–202. 53. N. Harada, H.-Y. Li, N. Koumura, T. Abe, M. Watanabe, M. Hagiwara, Enantiomer 1997, 2 , 349–352. 54. (a) A. Kawamura, N. Berova, V. Dirsch, A. Mangoni, K. Nakanishi, G. Schwartz, A. Bielawska, Y. Hannun, I. Kitagawa, Bioorg. Med. Chem. 1996, 4 , 1035–1043. (b) V. Dirsch, J. Frederico, N. Zhao, G. Cai, Y. Chen, S. Vunnam, J. Odingo, H. Pu, K. Nakanishi, N. Berova, D. Liotta, A. Bielawska, Y. Hannum, Tetrahedron Lett. 1995, 36 , 4959–4962. 55. K. Tanaka, G. Pescitelli, L. Di Bari, T. L. Xiao, K. Nakanishi, D. W. Armstrong, N. Berova, Org. Biomol. Chem. 2004, 2 , 48–58. 56. D. Gargiulo, N. Ikemoto, J. Odingo, N. Bozhkova, T. Iwashita, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1994, 116 , 3760–3767. 57. G. Cai, N. Bozhkova, J. Odingo, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1993, 115 , 7192–7198. 58. N. Harada, J. Am. Chem. Soc. 1973, 95 , 240–242. 59. M. Koreeda, N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1974, 96 , 266–268. 60. A. Guerriro, M. D’Ambrosio, V. Cuomo, F. Vanzanella, F. Pietra, Helvetica Chim. Acta, 1989, 72 , 438–446. 61. T. Ishikawa, M. Murota, T. Watanabe, T. Harayama, H. Ishii, Tetrahedron Lett. 1995, 36 , 4269–4272. 62. J. W. Moncrief, W. N. Lipscomb, J. Am. Chem. Soc. 1965, 87 , 4963–4964. 63. J. P. Kutney, D. E. Gregonis, R. Imhof, I. Itoh, E. Jahngen, A. I. Scott, W. K. Chan, J. Am. Chem. Soc. 1975, 97 , 5013–5015. 64. C. A. Parish, J-G. Dong, W. G. Bornmann, J. Chang, K. Nakanishi, N. Berova, Tetrahedron 1998, 54 , 15739–15758. 65. (a) X. Huang, B. H. Rickman, B. Bohran, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1998, 120 , 6185–6186. (b) T. Kurt´an, N. Nesnas, Y.-Q. Li, X. Huang, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2001, 123 , 5962–5973. (c) T. Kurt´an, N. Nesnas, F. E. Koehn, Y.-Q. Li, K. Nakanishi, N. Berova, J. Am. Chem. 2001, 123 , 5974–5982. 66. X. Huang, N. Fujioka, G. Pescitelli, F. E. Koehn, T. R. Williamson, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2002, 124 , 10320–10335. 67. N. Berova, G. Pescitelli, A. G. Petrovic, G. Proni, Chem. Commun. 2009, 5958–5980. 68. (a) A. G. Petrovic, Y. Chen. G. Pescitelli, N. Berova, G. Proni, Chirality, 2010, 22 , 129–139. (b) Y. Chen, A. G. Petrovic, M. Roje, G. Pescitelli, Y. Yang, M. Kayser, N. Berova, G. Proni, Chirality, 2010, 22 , 140–152. 69. J. W. Van Klink, S.-H. Baek, A. J. Barlow, H. Ishii, K. Nakanishi, N. Berova, N. B. Perry, R. T. Weavers, Chirality 2004, 16 , 549–558. 70. H. Ishii, S. Krane, Y. Itagaki, N. Berova, K. Nakanishi, P. J. Weldon, J. Natural Products 2004, 67, 1426–1430. 71. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Am. Chem. Soc. 2001, 123 , 2979–2989. 72. (a) Q. Yang, C. Olmsted, B. Borhan, Org. Lett. 2002, 4 , 3423–4326. (b) X. Li, M. Tanasova, C. Vasileiou, B. Borhan, J. Am. Chem. Soc. 2008, 130 , 1885–1893. (c) X. Li, B. Borhan, J. Am. Chem. Soc. 2008, 130 , 16126–16127. (d) M. Tanasova, C. Vasileiou, O. O. Olumolade, B. Borhan, Chirality 2009, 21 , 374–382. (e) M. Tanasova, Q. yang, C. C. Olmsted, C. Vasileiou, X. Li, M. Anyika, B. Borhan, Eur. J. Org. Chem. 2009, 4242–4253.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

73. (a) I. Hanazaki, H. Akimoto, J. Am. Chem. Soc. 1972, 94 , 4102–4106. (b) S. F. Mason, R. H. Seal, D. R. Roberts, Tetrahedron 1974, 30 , 1671–1682. (c) L. Di Bari, G. Pescitelli, P. Salvadori, J. Am. Chem. Soc. 1999, 121 , 7998–8004. 74. K. Harata, J. Tanaka, Bull. Chem. Soc. Jpn. 1973, 46 , 2747–2751. 75. N. Harada, N. Hiyoshi, V. P. Vassilev, T. Hayashi, Chirality 1997, 9 , 623–625. 76. N. Harada, V. P. Vassilev, N. Hiyoshi, Enantiomer 1997, 2 , 123–126. 77. (a) N. Harada, J. Iwabuchi, Y. Yokota, H. Uda, Y. Okamoto, H. Yuki, Y. Kawada, J. Chem. Soc., Perkin Trans. 1 1985, 1845–1848. (b) N. Harada, J. Iwabuchi, Y. Yokota, H. Uda, Croatica Chim. Acta 1989, 62 , 267–278. 78. (a) G. Hass, V. Prelog, Helv. Chim. Acta 1969, 52 , 1202–1218. (b) G. Hass, P. B. Hulbert, W. Klyne, V. Prelog, G. Snatzke, Helv. Chim. Acta 1971, 54 , 491–509. 79. J. Sagiv, A. Yogev, Y. Mazur, J. Am. Chem. Soc. 1977, 99 , 6861–6869. 80. N. Harada, H. Ono, T. Nishiwaki, H. Uda, J. Chem. Soc., Chem. Commun. 1991, 1753–1755. 81. K. Shingu, H. Kuritani, A. Kato, S. Imajo, Tetrahedron Lett. 1980, 21 , 3997–4000. 82. H. Ono, N. Harada, Enantiomer 1998, 3 , 245–250. 83. N. Harada, H. Ono, Enantiomer 1996, 1 , 115–118. 84. N. Harada, Y. Tamai, Y. Takuma, H. Uda, J. Am. Chem. Soc. 1980, 102 , 501–506. 85. N. Harada, Y. Tamai, H. Uda, J. Am. Chem. Soc. 1980, 102 , 506–511. 86. N. Harada, A. Saito, H. Ono, S. Murai, H.-Y Li, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, Enantiomer 1996, 1 , 119–138. 87. I. Akritopoulou-Zanze, K. Nakanishi, H. Stepowska, B. Grzeszczyk, A. Zamojski, N. Berova, Chirality 1997, 9 , 699–712. 88. N. Harada, H.-Y. Li, N. Koumura, T. Abe, M. Watanabe, M. Hagiwara, Enantiomer 1997, 2 , 349–352. 89. N. Harada, A. Saito, H. Ono, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, T. Kuriki, J. Am. Chem. Soc. 1991, 113 , 3842–3850. 90. P. Zhou, N. Zhao, D. Rele, N. Berova, K. Nakanishi, K. J. Am. Chem. Soc. 1993, 115 , 9313–9314. 91. (a) D. Rele, N. Zhao, K. Nakanishi, N. Berova, Tetrahedron 1996, 52 , 2759–2776. (b) N. Zhao, N. Berova, K. Nakanishi, M. Rohmer, P. Mougenot, U. J. J¨urgens, Tetrahedron 1996, 52 , 2777–2788. 92. N. Harada, K. Nakanishi, S. Tatsuoka, J. Am. Chem. Soc. 1969, 91 , 5896–5898. 93. Y. A. Berlin, M. N. Kolosov, L. A. Piotrovich, Tetrahedron Lett. 1970, 1329–1331. 94. J. A. Mills, J. Chem. Soc. 1952, 4976–4985. 95. J. H. Brewster, Tetrahedron 1961, 13 , 106–122. 96. (a) N. Harada, M. Ohashi, K. Nakanishi, J. Am. Chem. Soc. 1968, 90 , 7349–7351. (b) N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1968, 90 , 7351–7352. 97. (a) N. Harada, J. Iwabuchi, Y. Yokota, H. Uda, K. Nakanishi, J. Am. Chem. Soc. 1981, 103 , 5590–5591. (b) N. Harada, Y. Yokota, J. Iwabuchi, H. Uda, M. Ochi, J. Chem. Soc., Chem. Commun. 1984, 1220–1221. 98. (a) H.-U. Humpf, N. Berova, K. Nakanishi, M. B. Jarstfer, C. Dale Poulter, J. Org. Chem. 1995, 60 , 3539–3542. (b) K. Tanaka, G. Pescitelli, K. Nakanishi, N. Berova, Monatsh. Chem. 2005, 136 , 367–395. 99. K. Tanaka, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2003, 125 , 10802–10803. 100. M. Satake, A. Morohashi, H. Oguri, T. Oishi, M. Hirama, N. Harada, T. Yasumoto, J. Am. Chem. Soc. 1997, 119 , 11325–11326. 101. M. Hirama, T. Oishi, H. Uehara, M. Inoue, M. Maruyama, H. Oguri, M. Satake, Science, 2001, 294 , 1904–1907.

165

166

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

102. A. Sakurai, H. Horibe, N. Kuboyama, Y. Hashimoto, Y. Okumura, J. Biochem. 1995, 118 , 552–554. 103. J. Kobayashi, M. Yoshinaga, N. Yoshida, M. Shiro, H. Morita, J. Org. Chem. 2002, 67 , 2283–2286. 104. M. L. Kenney, F. Cort´es-Selva, J. M. Perez-Victoria, I. A. Jim´enez, A. G. Gonzalez, O, M. Mu˜noz, F. Gamarro, S. Castanys, A. G. Ravelo. J. Med. Chem. 2001, 44 , 4668–4676. 105. J. Guo, G. Schlingmann, G. T. Carter, K. Nakanishi, N. Berova, Chirality, 2000, 12 , 43–51. 106. H. Kobayashi, S. Meguro, T. Yoshimoto, M. Namikoshi, Tetrahedron 2003, 59 , 455–459. 107. T. Wang, O. Shirota, K. Nakanishi, N. Berova, L. A. McDonald, L. R. Barbieri, G. Carter, Can. J. Chem. 2001, 79 , 1786–1791. 108. T. Sunazuka, T. Shirahata, K. Yoshida, D. Yamamoto, Y. Harigaya, T. Nagai, H. Kiyohara, H. Yamada, I. Kuwajima, S, Omura, Tetrahedron Lett. 2002, 43 , 1265–1268. 109. T. F. Molinski, L. J. Brzezinski, J. W. Leahy, Tetrahedron: Asymmetry 2002, 13 , 1013–1016. 110. (a) K. Tanaka, G. Pescitelli, L. Dibari, T. L. Xiao, K. Nakanishi, D. W. Armstrong, N. Berova, Org. Biomol. Chem. 2004, 2 , 48–58. (b) K. Tanaka, Y. Itagaki, M. Satake, H. Naoki, T. Yasumoto, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2005, 127 , 9561–9570. 111. R. W. Baker, S. Liu, M. V. Sargent, Aust. J. Chem. 1998, 51 , 255–266. 112. M. S. Buchanan, M. Gill, P. Millar, S. Phonh-Axa, E. Raudies, J. Yu, J. Chem. Soc., Perkin Trans. 1 1999, 795–801. 113. M. Hashimoto, T. Tsushima, T. Murakami, M. Nomiya, N. Takada, K. Tanaka, Bioorg. Med. Chem. Lett. 2008, 18 , 4228–4231. 114. T. Murakami, T. Tsushima, N. Takada, K. Tanaka, K. Nihei, T. Miura, M. Hashimoto, Bioorg. Med. Chem. 2009, 17 , 492–495. 115. F. H. Kong, D. J. Faulkner, J. Org. Chem. 1993, 58 , 970–971. 116. D. S. Dalisay, S. Tsukamoto, T. Molinski, J. Natl. Prod . 2009, 72 , 353–359. 117. G. M. Nicholas, T. F. Molinski, J. Am. Chem. Soc. 2000, 122 , 4011–4019. 118. S. Yoshida, K. Kito, T. Ooi, K. Kanoh, Y. Shizuri, T. Kusumi, Chem. Lett. 2007, 36 , 1386–1387. 119. S. Aoki, Y. Watanabe, M. Sanagawa, A. Setiawan, N. Kotoku, M. Kobayashi, J. Am. Chem. Soc. 2006, 128 , 3148–3149. 120. L. Wang, X. Wang, J. Cui, W. Ren, N. Meng, J. Wang, X. Qian, Tetrahedron: Asymmetry 2010, 21 , 825–830. 121. S. T. Deyrup, J. B. Gloer, K. O’Donnell, D. T. Wicklow, J. Nat. Prod . 2007, 70 , 378–382. 122. K. Krohn, M. H. Sohrab, T. van Ree, S. Draeger, B. Schulz, S. Antus, T. Kurtan, Eur. J. Org. Chem. 2008, 5638–5646. 123. (a) K. Tsubaki, K. Takaishi, H. Tanaka, M. Miura, T. Kawabata, Org. Lett. 2006, 8 , 2587–2590. (b) K. Tsubaki, Org. Biomol. Chem. 2007, 5 , 2179–2188. 124. J. Naito, Y. Yamamoto, M. Akagi, S. Sekiguchi, M. Watanabe, N. Harada, Monatsh. Chem. 2005, 136 , 411–445. 125. J. L. Alonso-Gomez, P. Schanen, P. Rivera-Fuentes, P. Seiler, F. Diederich, Chem. Eur. J . 2008, 14 , 10564–10568. 126. J. Golik, T. W. Doyle, B. Krishnan, G. Dubay, J. A. Matson, J. Antibiot. 1989, 42 , 1784–1789. 127. T. Hattori, K. Sakurai, N. Koike, S. Miyano, H. Goto, F. Ishiya, N. Harada, J. Am. Chem. Soc. 1998, 120 , 9086–9087.

5 CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS: THEORETICAL DETERMINATION OF THE ABSOLUTE STEREOCHEMISTRY AND EXPERIMENTAL VERIFICATION Nobuyuki Harada and Shunsuke Kuwahara

5.1. INTRODUCTION Electronic circular dichroism (ECD) is very useful for characterization of chiral organic compounds with π -electron chromophores. That is, ECD enables one to determine the absolute configurations (ACs) of various natural products and chiral synthetic compounds by the use of appropriate CD methods, exemplified by the CD exciton chirality method [1–5]. In general, chiral compounds containing a chiral conjugated π -electron chormophore such as conjugated diene and enone exhibit medium CD Cotton effects. These chromophores contained in chiral compounds naturally adopt twisted conformations falling in the category of the inherently dissymmetrical chromophore. The observed CD Cotton effects are generally governed by the helicity of the diene or enone moiety, from which the absolute configuration can be determined. In some cases, however, the allylic group makes a contribution to the CD, and hence the observed CD does not agree with the chromophore helicity. On the other hand, compounds containing further extended and conjugated π -electron chromophores exhibit much more intense CD Cotton effects, as will be discussed in this chapter. In these cases, the CD Cotton effects are mostly governed by the helicity or twisted structure of the π -electron chromophore itself. Therefore, the CD spectra of these systems can be calculated by the π -electron approximation such as the π -electron self-consistent-field / configuration interaction / dipole velocity molecular orbital (SCFCI-DV MO) method [6–8]. In fact, we have determined the ACs of various natural products and chiral synthetic compounds by the use of the π -electron SCF-CI-DV MO method. In this chapter, the principle and applications of this method are explained in detail. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

167

168

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

To confirm the ACs as determined by the theoretical method, chiral model compounds and/or natural products themselves were synthesized in enantiopure forms. In such cases, the CSDP acid method and/or MαNP acid method are very useful for enantioresolving racemic compounds into enantiopure derivatives, and the methods simultaneously enable one to determine the ACs by X-ray crystallography and/or by 1 H NMR anisotropy [9–12]. The CD spectra of the synthesized model compounds were compared with those of the compounds in question. By these experimental studies, the theoretical method was established to lead to the correct absolute configurational assignments. Besides the theoretical method, the combination of CD spectroscopy and X-ray crystallographic analysis with an internal reference is also reliable for determining the ACs of various chiral compounds. This method has been applied to various natural products, chiral spiro compounds, light-powered chiral molecular motors, chiral C60 fullerene bis-adducts, and so on which led to the unambiguous assignment of ACs as described in this chapter. In the case of light-powered chiral molecular motors, the motor rotation mechanism and dynamics were also clarified by CD spectroscopy together with 1 H NMR spectroscopy. CD spectroscopy is thus useful not only for studying chiral stereochemistry, but also for the static and dynamic behavior of natural products and chiral synthetic functional compounds. In this chapter are described the research results carried out mostly by the authors’ group.

5.2. THEORETICAL CALCULATION OF CD AND UV SPECTRA BY THE π -ELECTRON SCF-CI-DV MO METHOD The CD and UV spectra of an extended π -electron system can be calculated by the π electron SCF-CI-DV MO method, where the rotational strength Rba and dipole strength Dba are expressed as follows [6]. Rba = 2(ψa |∇|ψb )(ψa |r × ∇|ψb )μB 2 /(π σba )

(5.1)

Dba = 2(ψa |∇|ψb ) μB /(π σba )

(5.2)

2

2

2

where ∇ is the del operator, r is a distance vector, μB is the Bohr magneton, and σba is the excitation wavenumber of the transition a → b. The z -axis components of the electric and magnetic transition moments are formulated, respectively, as [6] (ψa |∇|ψb )z =



(Cra Csb − Csa Crb )cos Zrs

bonds

(ψa |r × ∇|ψb )z =



(Cra Csb − Csa Crb )(Xrs cos Yrs − Yrs cos Xrs )

(5.3) (5.4)

bonds

cos Zrs = (Zr − Zs )/Rrs

(5.5)

Xrs = (Xr + Xs )/2

(5.6)

where Cra is the coefficient of atomic orbital r in the wavefunction ψa ; is the expectation value of a dipole velocity ∇rs , which is directed along the bond rs in the direction r → s; Xr , Yr , and Zr are the x , y, and z coordinates of an atom r, respectively; and Rrs is the interatomic distance between atoms r and s. The x and y components of the electric and magnetic transition moments can be similarly calculated.

169

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

In the π -electron SCF-CI-DV MO method, the following standard values of atomic orbital parameters are proposed: for sp2 carbon, Z (C) = 1.0, W (C) = −11.16 eV, ˚ = −2.32 eV, (C–C, 1.388 A) ˚ = 4.70 × (rr|rr)(C) = 11.13 eV, β(C–C, 1.388 A) 107 cm−1 ; for ether oxygen, Z (O) = 2.0, W (O) = −33.00 eV, (rr|rr)(O) = 21.53 eV, β(C–O) = −2.00 eV, (C–O) = 6.00 × 107 cm−1 ; for sp 2 nitrogen, Z (N) = 1.0, W (N) = −14.12 eV, (rr|rr)(N) = 12.34 eV, β(C–N) = −2.32 or −2.55 eV, (C– N) = 4.70 or 5.17 × 107 cm−1 [1]. The electric repulsion integral (rr|ss) was approximated by the Nishimoto–Mataga equation. The resonance integral β and del value were calculated by the use of following equations, respectively [1]: ˚ β(1.388 A) ˚ cos θ β = [S /S (1.388 A)]

(5.7)

˚ ˚ = [(empirical, 1.388 A)/(theoretical, 1.388 A)] × (theoretical) cos θ

(5.8)

where θ is a dihedral angle. The overlap integral S and (theoretical) were calculated on the basis of the Slater orbitals. The configuration interactions between all singly excited states were included. The curves of the component CD and UV bands were approximated by the Gaussian distribution [13], ε(σ ) =



εk exp[−{(σ − σk )/σ }2 ]

(5.9)

k

ε(σ ) =



εk exp[−{(σ − σk )/σ }2 ]

(5.10)

k

where σ is half the bandwidth at 1/e peak height. The σ value of 2500 cm−1 was used as a standard value [1].

5.3. SOME ESTABLISHED EXAMPLES OF THE π -ELECTRON SCF-CI-DV MO METHOD The following are some examples of the application of the π -electron SCF-CI-DV MO method applied to various chiral natural products and synthetic chiral compounds with extended π -electron chromophores. These examples were already explained in reference 3, and hence the summary of the results (i.e., comparison of observed and calculated CD and UV–Vis spectra, absolute stereostructures, and experimental verification by Xray crystallography and/or by synthesis) are briefly described. The exciton CD studies of chiral spiroaromatics of 9,9 -spirobifluorene skeleton are described in Chapter 4 of this volume.

5.3.1. Absolute Configuration of (+)-1,8a-Dihydro-3,8-dimethylazulene Chiroptically active 1,8a-dihydro-3,8-dimethylazulene (+)-(1) was isolated from the cell culture of the liverwort Calypogeia granulata Inoue (Figure 5.1) [14]. The labile intermediate 1 with a unique 1,8a-dihydroazulene skeleton shows very intense chiroptical activity, [α]D + 1164 and intense CD Cotton effects as shown in Figure 5.1, suggesting

170

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Δe 314.0 (+19.7) +20

CD

0

H e × 10–4

–20

–40 235.2 (–47.4)

(8aS)-(+)

3

227.5 (25,600) –60 Obsd in hexane

2

UV 1 308.5 (5,400)

200

Figure 5.1. CD and UV spectra of naturally occurring (8aS)-(+)-1,8a-dihydro-3,8dimethylazulene (1) in hexane. (Redrawn from

0

300 λ (nm)

reference 15, with permission.)

Δe 219 (+46.2) H +40

(8aR)

CD

e × 10–4

+20

0

–20

219 (27,300) 313 (–13.9)

Calcd

UV

3

2

313 (9,900) 1

Figure 5.2. CD and UV curves of (8aR)-1,8a-dihydroazulene 2 calculated by the 200

300 λ (nm)

0

π -electron SCF-CI-DV MO method. (Redrawn from reference 15, with permission.)

171

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

a strongly twisted conjugated tetraene system. Therefore, it is reasonable to consider that the chiroptical activity of 1 is mainly due to the twist of the π -electron chromophore. To determine the AC of 1 theoretically, we carried out the calculation of the CD curve of 1,8a-dihydroazulene (2) on the basis of the π -electron framework approximation, using the SCF-CI-DV MO method, where its AC was arbitrarily chosen to be (8aR) (Figure 5.2) [15]. The theoretically calculated CD and UV curves agree well with the observed CD and UV spectra except for the sign of the CD ε values (compare Figures 5.1 and 5.2). That is, the observed CD curve of compound 1 is almost a mirror image of the curve calculated for the model compound (8aR)-2. Accordingly, the AC of the labile biosynthetic

Br

OCH3

O

OCH3

O O

O

H

(8aS)-(+)-4

(1S,8aS)-(+)-3

CH3O (1S,3aR,4S,7R,8aS)-(+)-5 X-ray

Scheme 5.1. A synthesis of the model compound (1S,8aS)-(+)-3.

+20

321.0 (+5.7) CD

0 Δe –20

–40

221.3 (–24.5)

3

e × 10–4

OCH3

(1S,8aS)-(+) –60

223.2 (23,700) 2 Obsd in EtOH

UV 324.3 (6,000)

1

Figure 5.3. CD and UV spectra of (1S,8aS)-(+)200

300 λ (nm)

0

1,8a-dihydro-1-methoxy-8a-methylazulene (3) in EtOH. (Redrawn from reference 15, with permission.)

172

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

intermediate (+)-1 was theoretically determined to be (8aS ). This conclusion was proved experimentally by the synthesis of model compounds, as described in the following. As a chiral model compound, (1S,8aS)-(+)-(3) was selected because the angular position 8a is blocked by a methyl group, and hence it resists the oxidation to azulene. The model compound was synthesized (Scheme 5.1) [15, 16], starting from the enantiopure Wieland–Miescher ketone (S )-(+)-(4) [17] via an intermediate bromide (+)-5, the AC of which was confirmed by the Bijvoet method in X-ray crystallography [15, 16]. The CD and UV spectra of (1S,8aS)-(+)-3 are shown in Figure 5.3; the CD curve of (1S,8aS )-(+)-3 is quite similar, in both sign and shape of Cotton effects, to that of dihydroazulene (+)-1. Therefore, it was proved experimentally that the natural product (+)-1 has 8aS absolute configuration. Thus the present results verify the theoretical determination of the absolute configuration of (+)-1 discussed above.

5.3.2. Circular Dichroism and Absolute Stereochemistry of Chiral Troponoid Spiro Compounds The SCF-CI-DV MO method has been successfully applied to a chiral troponoid spiro compound (6) as follows (Figure 5.4) [18]. Racemic spiroacetal (±)-6 could be enantioseparated by chiral HPLC of (+)poly(triphenylmethylmethacrylate). In the HPLC, the first-eluted fraction gave an enantiomer (−)-6, [α]D −4700, which shows the CD and UV spectra as illustrated in Figure 5.4. To determine the absolute stereochemistry of (−)-6, we calculated the

Δe

Obsd OO

+100

287 (+80.4)

N

+50

N

(S)-(–) e × 10–4

CD

0

–50

285 (23,500)

3 398 (–45.3)

–100

2

MeOH UV

378 (7,900) 1

Figure 5.4. CD and UV spectra of troponoid 300

400 λ (nm)

500

spiro compound (S)-(−)-6 in MeOH. (Redrawn from reference 18, with permission.)

173

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

CD spectra of (S )-6 by applying the π -electron SCF-CI-DV MO method, where the absolute configuration was arbitrarily chosen as S . The calculated (calcd) CD and UV spectra shown in Figure 5.5 are in a good agreement with the observed (obsd) spectra. Accordingly, the absolute stereochemistry of (−)-6 was theoretically determined to be S . The present conclusion is in line with the X-ray crystallographic results of a related compound [19].

5.3.3. Absolute Stereochemistry of the Halenaquinol Family Marine Natural Products The π -electron SCF-CI-DV MO method for calculation of CD spectra was next applied to the determination of the AC of the compounds of the halenaquinol family (Chart 5.1) [20–23]. To determine the absolute configuration of halenaquinol (+)-7, we first planned to apply the CD exciton chirality method [21] to the interaction between the naphthalene and benzoate chromophores. During the synthetic studies of a pertinent benzoate derivative 8, we obtained a naphthalene–diene compound (−)-9 (Chart 5.1) [24], which surprisingly exhibited much stronger CD Cotton effects than other halenaquinol derivatives did (Figure 5.6). The result clearly indicates that the major part of the CD Cotton effects originates from the π -electron chromophore composed of the naphthalene–diene moiety, which is twisted by the angular methyl group at the 12b position. Therefore, the twisted

Δe

Calcd OO

+100

289 (+76.3)

N

N

+50

(S)

e × 10–4

CD

0

–50

3

284 (23,500)

–100

2 UV

394 (–106.9) 1

Figure 5.5. CD and UV curves of spiro 300

400 λ (nm)

500

troponoid compound (S)-6 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 18, with permission.)

174

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HO

1

2

O

12

10

OR

CH3O

OSi

CH3O

4 7

HO

4

OCH3 O

O O

3

CH3O

OBz

CH3O

(12bS)-(+)-7

O CH3O

CH3O

8

(–)-9

7

6

O

(12bS)-10

Chart 5.1. Halenaquinol (7) and related compounds.

naphthalene–diene moiety is an ideal system for the determination of the absolute stereochemistry by applying the π -electron SCF-CI-DV MO method. As a model compound for the theoretical calculation of CD spectra, we adopted the molecule (12bS )-10, which has the essential part of the π -electron system of naphthalenediene compound 9. The absolute configuration of 10 was arbitrarily chosen as 12bS for the calculation, and the molecular geometry of the model compound was calculated by molecular mechanics [25]. The theoretical calculation of the CD and UV spectra of (12bS )-10 by the π electron SCF-CI-DV MO method afforded the curves illustrated in Figure 5.7 [24]. The

229 (+40.9) +40

OSi

CH3O

OCH3 O CH3O

+20 Δe

(12bS)-(–) 338 (+6.4)

CD

e × 10–4

0

Obsd in MeOH –20

218 (42,000) 301 (–23.3)

UV

4

324 (27,000)

2

Figure 5.6. CD and UV spectra of 0 200

300

400 λ (nm)

helanaquinol trans-methoxy diene derivative (3R,4R,12bS)-(−)-(9) in MeOH. (Redrawn from reference 24b, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

+40

223 (+35.5)

175

CH3O

O CH3O

+20

(12bS )

CD

Δe

378 (+3.3)

248 (–5.7)

–20

e × 10–4

0

Calcd

219 (40,300)

322 (–22.4)

4

349 (29,900) UV 2

Figure 5.7. CD and UV curves of the model compound 200

300

(12bS)-10 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 24b, with permission.)

0

400 λ (nm)

+100

CH3O

+50

CH3O

O (12bS )

D × 1036 cgs unit

CD

0

Calcd

R × 1040 cgs unit

–50

30 UV 20 10

200

300

400 λ (nm)

0

Figure 5.8. Calculated rotational and dipole strengths of the model compound (12bS)-10. (Redrawn from reference 24b, with permission.)

176

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

theoretically simulated CD curve is in good agreement with the observed curve of (−)-9. It is thus evident that the basic pattern of the CD and UV spectral curves, including the sign, position, intensity, and shape of the bands, was well reproduced by the calculation. Since the absolute configuration of the model compound 10 is set as 12bS, the comparison of the present calculated and observed CD data leads to the unambiguous determination that the naphthalene–diene compound 9 has the 12bS absolute configuration. Accordingly, the absolute stereochemistry of halenaquinol (+)-7 was theoretically determined to be 12bS . To clarify the applicability of the present theoretical method to such a complex system, we analyzed the composition of the apparent CD and UV bands [24b]. As shown in Figure 5.8, there are nine major electronic transitions that contribute to the CD and UV bands. The first and second transitions with weak positive rotational strength at 374.5 and 351.6 nm, respectively, generate the weak positive Cotton effect at 378 nm (Figure 5.8). Furthermore, the third transition with an intense negative rotational strength at 324.4 nm results in the negative Cotton effect at 322 nm, and the sixth transition with a strong positive rotational strength contributes mainly to the intense positive Cotton effect at 223 nm. The correspondence between the component transitions and the apparent CD is thus clear. Therefore, the present analysis makes the absolute configurational determination of the halenaquinol compounds more reliable. Theories and theoretically obtained results should be proved experimentally. We succeeded in the first total synthesis of (+)-halenaquinol 7 and related natural products starting from (8aR)-(−)-Wieland–Miescher ketone [26–29]. The CD and UV spectra of the compounds synthesized were, of course, identical to those of the natural products. By these total syntheses of halenaquinol family compounds, we have proved, in an excellent way, that their absolute configurations theoretically determined were correct [26].

5.3.4. Atropisomerism of Natural Products: CD and Absolute Stereochemistry of the Biflavone, 4,4  ,7,7 -Tetra-O-methylcupressflavone Atropisomers are chiral compounds devoid of a chirality center. Those compounds are unique because the rotation about a single bond connecting two bulky moieties is sterically hindered, and hence their rotational conformers are sufficiently stable to be resolved into enantiomers. A natural product of biflavone, 4,4  ,7,7 tetra-O-methylcupressuflavone (11), is one of such atropisomers (Figure 5.9). The CD spectrum of biflavone (−)-11 shows strong bisignate Cotton effects of positive first and negative second signs at 400–300 nm (Figure 5.9). These Cotton effects look like an exciton split CD; therefore, one may assign the positive exciton chirality—specifically, clockwise screw sense of P -helicity—to this atropisomer. In fact, during the stereochemical studies of this flavone, the aS absolute stereochemistry (or P -helicity) had initially been assigned to (−)-11. However, such a careless application of the exciton chirality method leads to the erroneous assignment of the AC as

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

+50

177

Obsd 362.0 (+25.6)

267.5 (+21.3) Δe

CD

e × 10–4

0

OH O

CH3O

O

–50

OCH3 10 OCH3

326.2 (–54.4) O

CH3O

225.8 (51,800)

UV OH O (aR)-(–)

273.0 (41,400)

5 324.2 (40,900)

in EtOH

200

300

0

400 λ (nm)

+50

Calcd

Figure 5.9. CD and UV spectra of 4,4 ,7,7 -tetra-O-methylcupressuflavone (aR)-(−)-11 in EtOH. (Redrawn from reference 30, with permission.)

359.7 (+28.6)

263.2 (+21.7) Δe

CD

OH O

CH3O –50

e × 10–4

0

O

317.5 (–45.0)

OCH3 10 OCH3

226.8 (78,300)

CH3O 322.6 (66,200)

O

OH O (aR)

5

UV

Figure 5.10. CD and UV curves of 0 200

300

400 λ (nm)

(aR)-4,4 ,7,7 -tetra-O-methylcupressuflavone 11 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 30, with permission.)

178

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

described below [30]. We calculated the CD and UV spectra of biflavone (−)-11 by the π -electron SCF-CI-DV MO method and came to the conclusion that the correct absolute stereochemistry of (−)-11 is aR (or M -helicity). The absolute stereochemistry of a model compound for the calculation was arbitrarily chosen as (aR). The structure of atropisomer (aR)-11 was calculated by molecular mechanics (MMP2) to generate the stable conformation, where the calculated dihedral angle between two flavone moieties was 91◦ . The CD and UV spectra of (aR)-11 were calculated by the π -electron SCF-CI-DV MO method (Figure 5.10). The calculated CD and UV curves are in excellent agreement with the observed curves, including sign, intensity, and position of bands. Based on these results, the absolute stereochemistry of biflavone (−)-11 was unambiguously determined as aR (or M helicity) [30]. There are two nonempirical methods for determining the AC of chiral compounds; one is the X-ray Bijvoet method, and the other is the theoretical CD method including the exciton chirality method. These two methods are based on totally different phenomena, but they should come to the same AC for the same compound. In the case of the biflavone (−)-11, there had been considerable confusion in the stereochemical studies. After publication of our assignment of (aR) absolute configuration based on theoretical CD studies [30], the opposite AC, namely (aS ), by X-ray crystallography was reported [31]. Which determination is more reliable? Most people may support the determination by X-ray analysis. However, we were confident about our assignment by the theoretical CD calculation, because of the nonempirical nature of the method. To solve such a problem, we proved the absolute stereostructure by the total synthesis of the natural enantiomer. We designed a synthetic route, where the absolute stereochemistry of an intermediate was determined by X-ray crystallography, and we succeeded in the total synthesis of the natural atropisomer (−)-11 [32, 33]. The CD and UV spectra of the synthetic sample (aR)-(−)-11 were identical to those of the natural sample. Therefore, it was concluded that the absolute stereochemistry of natural biflavone (−)-11 is (aR). Thus our theoretical determination of the absolute stereochemistry of biflavone 11 by the π -electron SCF-CI-DV MO method was thus proved experimentally by the total synthesis of natural atropisomer 11.

5.4. CD SPECTRA AND ABSOLUTE STEREOSTRUCTURES OF UNIQUE CHIRAL OLEFINS: DISCOVERY AND DEVELOPMENT OF LIGHT-POWERED MOLECULAR MOTORS There are various kinds of chiral compounds devoid of chirality centers. Those compounds cannot take a planar structure because of strong steric hindrance. Some examples of these compounds are chiral olefins, (E )-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 biphenanthrylidene (12) and its (Z )-isomer (13) as shown in Chart 5.2. Olefins 12 and 13 can exist as chiral compounds, and they have been actually resolved into enantiomers by chiral HPLC using a chiral stationary phase [34]. The CD spectra of these olefins showed intense Cotton effects reflecting their twisted π -electron chromophores. However, their absolute configurations have remained undetermined. To solve this problem, we carried out the theoretical calculation of their CD spectra by the π -electron SCF-CI-DV MO method, synthesis of enantiopure compounds, and the experimental determination of their absolute configurations as follows [35].

179

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

8

8 8

8

10

5

10

5

10

1

1 3

5

1

1

3

[CD(+)239.0](M,M)-(E)-12

10

5

[CD(–)239.0](P,P)-(E)-12

3

3

[CD(+)238.1](M,M)-(Z)-13

[CD(–)238.1](P,P)-(Z)-13

Chart 5.2. Absolute stereochemistry of unique chiral olefins, (E)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 biphenanthrylidene (12) and its (Z)-isomer (13).

5.4.1. Synthesis of Enantiopure Chiral Olefins 12 and 13, and Their CD Spectra During our calculation of CD and UV spectra by the π -electron SCF-CI-DV MO method, we realized that the reported CD ε values [34] are too small compared to the calculated values. To obtain reliable CD and UV data, we synthesized the enantiopure target compounds. According to the reported procedure [34], the racemic olefins, trans-(±)-12 and cis-(±)-13 were synthesized as shown in Scheme 5.2. The relative stereostructures of trans-olefin 12 and cis-olefin 13 were determined by 1 H NMR spectroscopy and then confirmed by the X-ray crystallographic analysis of trans-olefin 12 [35] as illustrated in Figure 5.11. The molecular framework of this compound is thus nonplanar and strongly twisted, supporting that this molecule can take a chiral form. Next the enantioseparation of racemic trans-olefin (±)-12 by chiral HPLC was attempted. We found that hydrocarbon 12 could be completely resolved into enantiomers using a chiral stationary phase of (+)-poly(triphenylmethylmetacrylate) under the reverse phase condition using MeOH as eluent and a column temperature of 3◦ C. To remove a small amount of the polymer of the chiral stationary phase, which was present as a contaminant, the fraction of each enantiomer was purified by HPLC (ODS-C18 , MeOH). From the first-eluted fraction, enantiopure olefin [CD(+)239.0]-(E )-12 was obtained, and its 1 H NMR spectrum was identical to of racemate (±)-12. The CD and UV spectra of the first-eluted trans-olefin [CD(+)239.0]-(E )-12 are shown in Figure 5.12, where the UV spectrum shows a broad band at 329.8 nm, which

8 8 10

5

1 O 14

10

5

+ 1

3

3 (±)-(Z)-13 (±)-(E)-12

Scheme 5.2. Preparation of racemic olefins, (E)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 -biphenanthrylidene (12) and its (Z)-isomer (13).

180

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 5.11. ORTEP drawing of racemic trans-olefin, (E)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 -biphenanthrylidene (±)-12. (Redrawn from reference 35, with permission.)

+100 A = +211.5

239.0 (+58.2)

Δe

e × 10–4

0

CD –100 214.2 (−153.3) 216.2 (82,800)

10 (M,M)-(E)

–200 Obsd in MeOH

5

UV

200

250

300 λ (nm)

350

400

0

Figure 5.12. CD and UV spectra of the first-eluted trans-olefin [CD(+)239.0]-(E)-12 in MeOH. (Redrawn from reference 35, with permission.)

181

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

may be assigned to the 1 La transition of naphthalene chromophore. In the corresponding region, the CD spectrum shows a broad positive Cotton effect of medium intensity (λext 331.8 nm, ε +26.0). On the other hand, in the 1 Bb transition region, the UV spectrum shows an intense broad band (λext 232.2 nm, ε 61,800 and λext 216.2 nm, ε 82,800), and the CD spectrum shows intense positive and negative Cotton effects (λext 239.0 nm, ε +58.2 and λext 214.2 nm, ε −153.3): The amplitude A value between the peak and trough is +211.5. Such intense CD Cotton effects clearly indicate that the π -electron system of trans-12 is strongly twisted. The enantioseparation of cis-olefin (±)-13 was next examined, and we found that the reverse-phase HPLC used for trans-olefin (±)-12 was not useful. Instead, the chiral HPLC using (+)-poly(triphenylmethylmetacrylate) and hexane as eluent was effective; cis-olefin (±)-13 was partially separated into enantiomers at 3◦ C. To obtain the enantiopure olefin, the first-eluted fraction was recycled five times. Since we found the unexpected thermal racemization of cis-olefin 13 at room temperature, as will be discussed later, the CD and UV spectra were immediately measured after HPLC separation. Figure 5.13 shows the CD and UV spectra of the first-eluted cis-olefin [CD(+)238.1]-(Z )-13 in hexane. The first-eluted cis-olefin [CD(+)238.1]-(Z )-13 exhibits a broad UV band at 301.9 nm (ε 11,300) at the 1 La transition of naphthalene chromophore. In the corresponding region, the CD spectrum shows a broad weak negative Cotton effect

238.1 (+189.7)

+200

A = +429.0

CD

+100 Δe

e × 10–4

0

–100

–200 223.5 (–239.3)

10

(M,M)-(Z)

223.0 (73,700) –300 Obsd in hexane

5

UV

Figure 5.13. CD and UV spectra of the 200

250

300 λ (nm)

350

400

0

first-eluted cis-olefin [CD(+)238.1]-(Z)-13 in hexane. (Redrawn from reference 35, with permission.)

182

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(λext 339.0 nm, ε −12.3). In the 1 Bb transition region, the UV spectrum shows an intense broad band (λext 223.0 nm, ε 73,700), while the CD spectrum shows very intense positive and negative Cotton effects (λext 238.1 nm, ε +189.7 and λext 223.5 nm, ε −239.3): The amplitude A value between the peak and trough is +429.0. The intense CD data thus indicate that the π -electron system of cis-13 is also strongly twisted.

5.4.2. Absolute Stereochemistry of Chiral Olefins trans-12 and cis-13 as Determined by the Calculation of CD and UV Spectra Using the SCF-CI-DV MO Method To determine the absolute stereochemistry of chiral olefins 12 and 13, we next calculated the CD and UV spectra by the π -electron SCF-CI-DV MO method. As a model compound for the calculation, the (M,M)-(E )-enantiomer 12 was arbitrarily chosen, and the atomic coordinates were obtained by the MOPAC AM1 calculation. The calculated CD and UV spectra of trans-olefin (M,M)-(E )-12 are shown in Figure 5.14 [35]. As seen in figures 5.12 and 5.14, the CD and UV spectra of trans-olefin 12 were reproduced well by the calculation. In the 1 La transition around 300 nm, the positive CD band was obtained by calculation, which agreed with the observed CD, although its intensity was smaller than the observed one. In the 1 Bb transition around 200–250 nm, +100 240.4 (+87.9)

0 Δe –100

e × 10–4

CD

–200 219.3 (–256.0)

10 223.2 (94,400) –300

(M,M)-(E)

Calcd 5

UV A = +343.9

Figure 5.14. CD and UV spectral curves of 200

250

300 λ (nm)

350

400

0

trans-olefin (M,M)-(E)-12 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 35 with permission.)

183

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

the intense positive and negative CD Cotton effects were obtained by calculation. The sign and shape of Cotton effects agreed well with those of observed spectrum, although the calculated intensity was stronger. Therefore, the absolute stereochemistry of the firsteluted trans-olefin [CD(+)239.0]-(E )-12 was clearly determined to be (M,M) by the theoretical calculation. The first-eluted enantiomer is designated as [CD(+)239.0]-(M,M)(E )-12 [35]. The CD and UV spectra of cis-olefin (M,M)-(Z )-13 were similarly calculated by the π -electron SCF-CI-DV MO method as shown in Figure 5.15. When comparing with Figure 5.13, it was clear that the CD and UV spectra of cis-olefin 13 were also wellreproduced by the calculation. In the 1 La transition around 300–370 nm, the negative CD band was obtained by calculation, which agreed with the observed CD, although its intensity was again larger. In the 1 Bb transition around 200–250 nm, the intense positive and negative CD Cotton effects were obtained by calculation. The sign and shape of the Cotton effects agreed well with those of the observed spectrum, although the calculated intensity in this case was weaker. Therefore, the absolute stereochemistry of the first-eluted cis-olefin [CD(+)238.1]-(Z )-13 was clearly determined to be (M,M) by theoretical calculation. Thus the absolute stereochemistry of the first-eluted enantiomer is designated as [CD(+)238.1]-(M,M)-(Z )-13 [35].

+100 232.6 (+76.7)

0

Δe

10 CD e × 10–4

–100

215.5 (–158.0) 211.9 (57.600) –200

(M,M)-(Z)

5

Calcd A = +234.7

UV

Figure 5.15. CD and UV spectral curves of 200

250

300 λ (nm)

350

400

0

cis-olefin (M,M)-(Z)-13 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 35, with permission.)

184

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

5.4.3. Unexpected Thermal Racemization of cis-Olefin 13 During the studies discussed above, we observed that chiral cis-olefin 13 underwent an unexpected thermal racemization at room temperature. We had first considered that if one of these chiral olefins undergoes the racemization, it must be trans-olefin 12, because of less steric hindrance. In cis-olefin 13, two naphthalene moieties overlap with each other as seen in the X-ray stereostructure (Figure 5.16), which generates a severe steric hindrance, and therefore it is difficult to image the racemization. However, it was confirmed that cis-olefin 13 really undergoes thermal racemization, which was monitored by CD spectrum as illustrated in Figure 5.17. The thermal racemization of cis-olefin 13 was also measured by the magnetization transfer experiment of 1 H NMR spectroscopy. On the other hand, it was clarified that trans-olefin 12 does not undergo racemization at room temperature. To obtain the enantiopure cis-olefin 13 and to measure its CD spectrum, we carried out the HPLC separation at lower temperature. That is, the chiral HPLC column was cooled at −30◦ C during the enantioseparation, and the CD spectrum was measured at −50◦ C as illustrated in Figure 5.18, where the observed CD intensity was corrected for volume contraction. The CD intensity of cis-13 in Figure 5.18 is larger than that in Figure 5.13. It is thus important to measure the CD spectrum of the enantiopure sample. The reaction mechanism of the thermal racemization of cis-olefin 13 was later clarified by theoretical calculation [37], where the (M, P)-(Z ) isomer was included as an intermediate. This mechanism indicates that two naphthalene moieties slip by each other. This is the critical reaction step in the light-powered molecular motor discussed below.

Figure 5.16. ORTEP drawing of racemic cis-olefin, (Z)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 biphenanthrylidene (±)-13. (Redrawn from reference 36, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

185

+200 0h

Racemization of

1h

+100

2h Δe

(M,M)-cis

3h 4h 5h

in hexane at room temp.

0 % 100 UV λmax 222.8 nm –100

CD CD λext 238.1 nm

50

t1/2 = 1.2 h –200

0 200

250

Figure 5.17. Decrease of CD intensity of cis-olefin 0

300 λ (nm)

2

4

6 h 350

[CD(+)238.1]-(M,M)-(Z)-13 in hexane due to the thermal racemization at room temperature. (Redrawn from reference 36 with permission.)

5.4.4. Experimental Determination of Absolute Stereochemistry of trans-Olefin 12 and cis-Olefin 13: Use of the Internal Reference of Absolute Configuration in X-Ray Analysis The absolute stereostructures of trans- and cis-olefins were theoretically determined by the calculation of CD and UV spectra using the π -electron SCF-CI-DV MO method as described above. But the authors believe that the theoretically determined results have to be proved experimentally. That is, the problem is now how to prove the absolute stereochemistry of 12 and 13 experimentally. To solve this problem, we adopted the following strategy. At first we thought to synthesize derivatives containing a heavy atom like Br or S and to carry out X-ray crystallography for determining the absolute configuration using the X-ray Bijvoet method. However, all attempts of the synthesis were unsuccessful. It was then decided to introduce an internal reference of the absolute configuration—that is a methyl group in a chiral position, as shown in compounds 15 and 16 (Chart 5.3). As the starting material for the synthesis, racemic cis-alcohol (±)-17 was selected, and it was enantioresolved by the CSDP acid (camphor-sulfonyl-dichloro-phthalic acid) method [9–12] as shown in Scheme 5.3. Cis-alcohol (±)-17 was esterified with CSDP

186

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

239.6 (+222.2)

+200 Δe

CD

+100

282.5 (+11.9)

e × 10–4

0 338.0 (–14.0) –100 256.8 (–80.1) 10 –200 (M,M)-cis

224.0 (–281.3)

–50.0°C

–300

222.8 (71.900) in hexane 5 UV 301.9 (11,300)

300

200

400

Figure 5.18. CD and UV spectra of cis-olefin

0

[CD(+)238.1]-(M,M)-(Z)-13 in hexane at −50◦ C. (Redrawn from reference 36, with permission.)

λ (nm)

8 8

H CH3

10

5

10

5

1 1

3

H3C

[CD(–)237.2]-(3R,3'R)(P,P)-(E)-15

H

3

H CH3

H3C

H

[CD(–)238.0]-(3R,3′R)(P,P)-(Z)-16

Chart 5.3. Dimethyl-substituted chiral olefins 15 and 16 useful for determining absolute configurations.

acid (1S,2R,4R)-(−)-18 to yield a diastereomeric mixtures of esters, which was easily separated by HPLC on silica gel. The second-eluted ester (−)-19b was obtained as a solid, which was recrystallized from EtOAc giving large prisms suitable for X-ray crystallography. The AC of the second-eluted ester (−)-19b was unambiguously determined to be (3S, 4S ) by the heavy atom effect of Cl and S atoms and also by the use of the camphorsultam moiety as an internal reference of the AC (Figure 5.19).

187

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Cl

Cl Cl

Cl Cl N S O

+ OH CH3

O

N S O O

COOH

(1S,2R,4R)-(–)-18

+ O

O

H3C

O

(±)-17

OO

Cl N S O

O

O H3C

O

(3S,4S)-(–)-19b, X-ray

(3R,4R)-(+)-19a

Scheme 5.3. Enantioresolution of alcohol (±)-17 by the CSDP acid method.

Figure 5.19. ORTEP drawing of the second-eluted CSDP ester (3S,4S)(−)-19b. (Redrawn from reference 38, with permission.)

The AC of the first-eluted CSDP ester (+)-19a was, therefore, assigned as (3R,4R). The LiAlH4 reduction of the first-eluted CSDP ester (3R,4R)-(+)-19a yielded enantiopure cis-alcohol (3R,4R)-(+)-17, which was then oxidized to give enantiopure ketone (3R)-(−)-20 (Scheme 5.4). McMurry coupling reaction of ketone (3R)-(−)-20 yielded the desired dimethyl trans-olefin (−)-15, which was purified by repeated HPLC under normal-phase and reverse phase conditions affording enantiopure sample, [CD(−)237.2](3R,3 R)-(P, P)-(−)-15 ([α]D −446.2). The large negative optical rotation value of the product (−)-15 indicates that no racemization occurred during the McMurry reaction.

Cl Cl

S O

N O

OO H3C

O

H CH3

HO

O CH3

(3R,4R)-(+)-17 (3R,4R)-(+)-19a

CH3 (3R)-(–)-20

H3C H [CD(–)237.2]-(3R,3′R)(P,P)-(E)-15

Scheme 5.4. Synthesis of dimethyl trans-olefin [CD(–)237.2]-(3R,3 R)-(P, P)-(−)-15.

188

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The relative and absolute stereochemistry of dimethyl trans-olefin (−)-15 was determined by X-ray crystallography as follows. Recrystallization of the product (−)-15 from MeOH afforded prismatic crystals, one of which was subjected to X-ray crystallographic analysis, and the stereostructure was determined as shown in Figure 5.20 [38]. Since compound (−)-15 is just a hydrocarbon containing no heavy atom, the AC of (−)-15 could not be determined by the X-ray analysis. However, it has two methyl groups at chiral positions, and therefore the (3R,3 R) configuration was used as an internal reference of the AC. The absolute helicity of the chiral olefin part was thus determined as (P, P ) from the ORTEP drawing in Figure 5.20. The CD and UV spectra of dimethyl trans-olefin (3R,3 R)-(P, P)-(E )-(−)-15 were next measured as shown in Figure 5.21, where the CD spectrum shows a negative Cotton effect at 237.2 nm. Hence the enantiomer is designated as [CD(–)237.2]-(3R,3 R)-(P, P)(E )-(−)-15. The UV spectrum of (−)-15 is similar to that of trans-olefin [CD(+)239.0]-(M,M)(E )-12. On the other hand, the CD spectra of [CD(−)237.2]-(3R,3 R)-(P, P)-(E )-(−)-15 is also similar to that of [CD(+)239.0]-(M,M)-(E )-12 in position, shape, and absolute intensity, but opposite in sign (compare Figures 5.12 and 5.21). These results clearly indicate that the dimethyl groups in (−)-15 do not change the molecular conformation too much, and that the absolute helicity of the chiral olefin part in [CD(+)239.0]-(E )-12 is (M,M). The absolute stereochemistry of trans-olefin [CD(+)239.0]-(M,M)-(E )-12 as determined previously by the theoretical calculation of CD spectrum was thus confirmed in an experimental manner. To synthesize dimethyl cis-olefin (3R,3 R)-(Z )-16, enantiopure dimethyl trans-olefin [CD(–)237.2]-(3R,3 R)-(P, P)-(E )-(−)-15 was irradiated by a high-pressure mercury lamp using a Pyrex glass filter to yield dimethyl cis-olefin 16 (Scheme 5.5), the CD spectrum of which showed an intense negative Cotton effect at 238.0 nm. Therefore, the enantiomer was designated as [CD(−)238.0]-(3R,3 R)-(Z )-16. The helical sense of the naphthalene–double bond–naphthalene moiety of [CD(−)238.0]-(3R,3 R)-(Z )-16 was first studied by 1 H NMR spectroscopy and it was finally determined by X-ray crystallography as follows. We first obtained single crystals of racemate (±)-16, which were subjected to X-ray analysis, affording the ORTEP drawing as shown in Figure 5.22.

Figure 5.20. ORTEP drawing of dimethyl trans-olefin [CD(–)237.2](3R, 3 R)-(P, P)-(E)-(−)-15. (Redrawn from reference 38, with permission.)

189

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

217.8 (+148.5) +150 H CH3 +100 CD

H3 C H

+50 (3R,3′R)-(P,P)-(E)

Δe

e × 10–4

0

–50 10 237.2 (–92.5) A = –241.0 –100 218.4 (85,700) 5 UV

Obsd in MeOH

Figure 5.21. CD and UV spectra of dimethyl 200

250

300

350

400

0

trans-olefin [CD(–)237.2]-(3R,3 R)-(P, P)-(E)-(−)-15 in MeOH. (Redrawn from reference 38, with permission.)

λ (nm)

H CH3 hν

H3C H

[CD(–)237.2]-(3R,3′R)(P,P)-(E)-15

H

CH3

H 3C H

[CD(–)238.0]-(3R,3′R)(P,P)-(Z)-16

Scheme 5.5. Synthesis of dimethyl cis-olefin [CD(–)238.0]-(3R,3 R)-(P, P)-(Z)-16.

From the ORTEP drawing, the relative stereochemistry of (±)-16 was determined as (3R ∗ ,3 R ∗ )-(P ∗ , P ∗ )-(Z ). Since the 1 H NMR spectrum of chiral olefin [CD(−)238.0](3R,3 R)-(Z )-16 was identical with that of racemate (3R ∗ ,3 R ∗ )-(P ∗ , P ∗ )-(Z )-(±)-16, the absolute helicity of chiral olefin was determined as (P, P ). The absolute stereochemistry of dimethyl cis-olefin [CD(−)238.0]-(3R,3 R)-(P, P)-(Z )-16 was thus unambiguously determined. Later, we obtained single crystals of chiral dimethyl cis-olefin [CD(–)238.0](3R,3 R)-(P, P)-(Z )-16, and the same AC was determined from the X-ray crystallographic analysis [39].

190

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 5.22. ORTEP drawing of dimethyl cis-olefin (3R∗ ,3 R∗ )-(P ∗ , P ∗ )-(Z)-(±)-16. (Redrawn from reference 38, with permission.)

+400 223.4 (+334.0)

+200

CD

e × 10–4

H3 C H

H CH3

Δe

(3R,3′R)-(P,P)-(Z) 0

10

238.0 (–226.9)

–200

A = –560.9 5

222.4 (76,500)

UV

–400

Obsd in hexane

Figure 5.23. CD and UV spectra of dimethyl 200

250

300 λ (nm)

350

400

0

cis-olefin [CD(–)238.0]-(3R,3 R)-(P,P)-(Z)-16 in hexane. (Redrawn from reference 38, with permission.)

191

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

The CD and UV spectra of [CD(–)238.0]-(3R,3 R)-(P, P)-(Z )-16 is shown in Figure 5.23. As in the case of trans-olefin 12 and dimethyl trans-olefin 15, the CD and UV spectra of dimethyl cis-olefin [CD(−)238.0]-(3R,3 R)-(P, P)-(Z )-16 are very similar to those of cis-olefin [CD(+)238.1]-(M,M)-(Z )-13 in position, shape, and absolute intensity, but opposite in sign (compare Figures 5.13 and 5.23). Therefore, the absolute stereochemistry of cis-olefin [CD(+)238.1]-(M,M)-(Z )-13 previously determined by the theoretical calculation of CD was confirmed experimentally. The X-ray crystallographic method using an internal reference of absolute configuration is thus very useful for the absolute configurational assignment of various chiral compounds. It is interesting that the CD spectrum of dimethyl cis-olefin [CD(–)238.0]-(3R,3 R)(P, P)-(Z )-16 did not show a decrease in intensity at room temperature, unlike the case of cis-olefin [CD(+)238.1]-(M,M)-(Z )-13. That is, dimethyl cis-olefin 16 does not racemize at all, because the two methyl groups block the racemization process.

5.4.5. Unique Photo- and Thermo-chemical Behavior of Chiral Dimethyl Olefin: First Discovery and Development of a Light-Powered Chiral Molecular Motor Further studies of the photo- and thermo-chemistry of chiral dimethyl olefins 15 and 16 led to the first discovery of a light-powered chiral molecular motor as described below. During the photochemical studies of chiral dimethyl trans-olefin (−)-15, we found the formation of another product, yellow-colored dimethyl trans-olefin (+)-21 as shown in Scheme 5.6, although olefins (−)-15 and (+)-16 are colorless. Later it was clarified that the product (+)-21 was not directly formed from (−)-15 but from dimethyl cis-olefin (+)-16, and this photochemical step is reversible [40]. The structure of yellow-colored dimethyl trans-olefin (+)-21 was first studied by 1 H NMR spectrum, where two methyl groups appeared at δ 0.31 ppm, implying that the high field shift is due to the anisotropy effect by a neighboring naphthalene ring. These NMR data suggested the structure of (3R,3 R)-(M,M)-(E )-21, in which two methyl groups are placed at the equatorial position. The relative stereochemistry of olefin 21 was confirmed by X-ray crystallography of racemate (±)-21 as shown in Figure 5.24, where the two equatorial methyl groups are in contact with the naphthalene rings, causing a strong steric hindrance between methyl and naphthalene moieties. This effect makes olefin (+)-21 unstable, and also affects the π electron framework to change its color. This is the major reason why olefin 21 is yellow. Figure 5.25 shows the CD and UV spectra of the yellow dimethyl trans-olefin (3R,3 R)-(M,M)-(E )-(+)-21 together with those of colorless dimethyl trans-olefin

H CH3

H CH3

hn

hν

hn H 3C H (3R,3′R)-(P,P)-(E)-(–)-15 with two axial methyl groups

H CH3

H 3C H

(3R,3′R)-(P,P)-(Z)-(+)-16 with two axial methyl groups

H3C

H

(3R,3′R)-(M,M)-(E)-(+)-21 with two equatorial methyl groups

Scheme 5.6. Photochemical interconversion between chiral dimethyl olefins (−)-15, (+)-16, and (+)-21.

192

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 5.24. ORTEP drawing of yellow colored dimethyl trans-olefin (3R∗ ,3 R∗ )-(M∗,M∗ )-(E)-(±)-21. (Redrawn from reference 40, with permission.)

+200 H CH3

+100

CD

H3C

H

(3R,3′R)-(M,M)-(E)-(+)

e × 10–4

Δe 0

H CH3

–100

10

H3C H UV

5

–200

Figure 5.25. CD and UV spectra of

(3R,3′R)-(P,P)-(E)-(–)

200

300

400 λ (nm)

0

yellow-colored unstable dimethyl trans-olefin (3R,3 R)-(M,M)-(E)-(+)-21 in EtOH together with those of colorless stable (3R,3 R)-(P, P)-(E)-(−)-15 in EtOH. (Redrawn from reference 40, with permission.)

193

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

(3R,3 R)-(P, P)-(E )-(−)-15. It should be noted that in the UV–Vis spectrum of (3R,3 R)-(M,M)-(E )-(+)-21, the 1 La transition of the naphthalene chromophore is largely red-shifted at 320–420 nm, causing the yellow color. In the corresponding region, the CD spectrum shows a broad positive Cotton effect reflecting the inversed (M,M) helicity, while olefin (3R,3 R)-(P, P)-(E )-(−)-15 exhibits a negative Cotton effect in the 1 La transition region. In the 1 Bb transition region around 200–270 nm, the UV spectrum shows two absorption bands. On the other hand, the CD spectrum shows intense but complex Cotton effects as shown in Figure 5.25, reflecting the strongly twisted π -electron system. The absolute stereochemistry of the unique third isomer (3R,3 R)-(M,M)-(E )-(+)-21 was unambiguously determined. It should be noted that the internal reference method of absolute configuration is thus applicable to the X-ray analysis of racemic compounds. As shown in Scheme 5.6, the photochemical step between stable dimethyl cis-olefin (+)-16 and unstable dimethyl trans-olefin (+)-21 was reversible, as expected for olefin compounds. However, it was surprising that the step between stable dimethyl trans-olefin (−)-15 and stable dimethyl cis-olefin (+)-16 was irreversible. That is, the photochemical conversion from stable dimethyl trans-olefin (−)-15 to stable dimethyl cis-olefin (+)-16 occurred, but the reverse reaction did not proceed. Why? To solve this problem, we postulated the reaction scheme as shown in Scheme 5.7, where trans-olefin (−)-15, cis-olefin (+)-16, and trans-olefin (+)-21 were renamed trans-isomer (−)-22a, cis-isomer (+)-22c, and trans-isomer (+)-22d, respectively. As discussed above, it was observed that the photochemical isomerization between stable cis-isomer (+)-22c and unstable trans-isomer (+)-22d was reversible, but unstable transisomer (+)-22d underwent thermal isomerization to stable trans-isomer (−)-22a. It was assumed that the unstable cis-isomer (3R,3 R)-(M,M)-(Z )-22b must exist as a primary

hn

Δ hn

H

H

CH3

CH3

H3C

H

(3R,3′R)-(M,M)-(Z)-22b

H3C

H

H Δ

CH3

H3C H

(3R,3′R)-(P,P)-(Z)-(+)-22c

(3R,3′R)-(P,P)-(E)-(–)-22a

hn H

CH3 hn

CH3 H

(3R,3′R)-(M,M)-(E)-(+)-22d

Scheme 5.7. The cyclic reaction scheme of photo- and thermo-chemical conversions of unique dimethyl olefins.

194

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

product of the photochemical conversion of (−)-22a, and the reverse reaction from 22b to (−)-22a may occur. However, if the isomer 22b is extremely unstable, the formed isomer 22b immediately and irreversibly converts to stable cis-isomer (+)-22c. Therefore, the total reaction from (−)-22a to (+)-22c becomes irreversible in agreement with the observed results [41]. In fact, we succeeded in detecting the unstable isomer (±)-22b in the photochemical reaction of (±)-22a at −60◦ C by 1 H NMR spectroscopy (Figure 5.26) [41]. Later, we realized that this system makes a unidirectional molecular motor rotation around the central double bond [42]. That is, when looking at the molecule from the left side, the naphthalene moiety on the left rotates counterclockwise against the naphthalene moiety on the right-upper side. The photochemical step makes the rotation in both directions, but the thermal step rotates only counterclockwise. The reaction 22a → 22b → 22c → 22d thus makes the 360◦ rotation in the counterclockwise direction, and the motor returns to the starting place 22a. Therefore, the molecular motor can make a continuous rotation under photoirradiation and heating, where the direction of the motor rotation is governed by the molecular chirality. The photochemical energy is thus converted to the mechanical rotation of the molecule, and this was the discovery of the first light-powered molecular motor [42].

5.5. A NEW MODEL OF LIGHT-POWERED CHIRAL MOLECULAR MOTOR WITH HIGHER SPEED OF ROTATION The chiral olefins shown in Scheme 5.7 ideally satisfy the requirements of molecular motor, but its rotation was not fast, because the fourth rotation step (i.e., thermal reaction Aromatic Part

ppm

8.25

(3R*,3R*)-(P*,P*)-trans stable

8.00

7.75

7.50

7.25

7.00

6.75

7.75

7.50

7.25

7.00

6.75

irradiation at –60.0 °C after 1 day

ppm

8.25

8.00

NMR detection of unstable dimethyl cis-olefin in CD2 Cl2 : ∗, unstable cis-olefin (3R∗,3 R∗ )-(M∗ ,M∗ )-(Z)-22b; ◦, stable trans-olefin (3R∗ ,3 R∗ )-(P ∗ , P ∗ )-(E)-22a. (Redrawn from reference

Figure 5.26. 41.)

1H

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

195

22d → 22a), needs higher temperature and hence is slow, because of the severe steric hindrance between methyl group and naphthalene moiety. To make a faster molecular motor, we improved the structure as follows. A new model of chiral molecular motor 23 with higher speed of rotation was designed as shown in Chart 5.4, where the six-membered rings in molecular motor 22 were replaced by five-membered rings to diminish the steric hindrance between methyl group and naphthalene moiety [43, 44].

5.5.1. Synthesis, CD Spectra, X-ray Structure, and Absolute Stereochemistry To synthesize the new chiral molecular motor, we adopted the strategy shown in Scheme 5.8. Racemic cis-alcohol (±)-24 was esterified with CSDP acid, giving a diastereomeric mixture of esters, which was easily separated by HPLC on silica gel [43]. One of the advantages of the CSDP acid method is that CSDP esters tend to give single crystals suitable for X-ray crystallography with high probability [9–12]. In fact, the second-eluted CSDP ester (−)-cis-25b was obtained as single crystals by recrystallization from hexane/EtOAc. The single crystal was subjected to X-ray analysis, and the absolute stereostructure was determined by the use of CSDP acid moiety as an internal reference and also by the heavy atom effect (Figure 5.27). That is, the (1S,2S ) absolute configuration was assigned to (−)-cis-25b. The absolute configuration of the first-eluted ester was naturally determined as (1R,2R). To recover alcohol, the second-eluted CSDP ester (1S,2S )-(−)-cis-25b was hydrolyzed with KOH/MeOH yielding enantiopure cis-alcohol (1S,2S )-(+)-24. Cis-alcohol (1S,2S )-(+)-24 was next oxidized to yield enantiopure ketone (S )-(+)26. The product was then subjected to the McMurry coupling reaction with TiCl3 and LiAlH4 in THF giving dimethyl trans-olefin (2S,2 S)-(E )-(−)-23a (colorless prisms, yield 21%) and dimethyl cis-olefin (2S,2 S)-(Z )-(−)-23c (pale yellow prisms, yield 5%) as shown in Scheme 5.8. The relative stereochemistry of racemate (±)-23a was determined to be (2S ∗ ,2 S ∗ )∗ (M ,M ∗ )-(E ) by X-ray crystallography as shown in Figure 5.28. Since the 1 H NMR spectrum of chiral trans-olefin (−)-23a was identical to that of racemate (2S ∗ ,2 S ∗ )-(M ∗,M ∗ )(E )-(±)-23a, the absolute stereochemistry of (−)-23a was unambiguously determined to be (2S,2 S)-(M,M)-(E ).

8

CH3

H

5

2 1

1

2

H CH3

5 8

[CD(–)257.8]-(2S,2′S)(M,M)-(E)-23a

CH3

H

H CH3

[CD(–)270.0]-(2S,2′S)(M,M)-(Z)-23c

Chart 5.4. A new model of chiral molecular motor with higher speed of rotation.

196

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Cl

Cl

Cl

Cl

H3C

N

N

+

S

OH

O

S O

COOH

O

O

O (1S,2R,4R)-(–)-18

(±)-24

O

+ O

O

H3C (1R,2R)-(–)-cis-25a

Cl Cl N S

O

O

O

O

(1S,2S)-(–)-cis-25b

O

HO

H3C

CH3

(1S,2S)-(+)-cis-24 (1S,2S)-(–)-cis-25b, X-ray

H

CH3 +

O

CH3

H

H

H

CH3

CH3

CH3

(S)-(+)-26 [CD(–)257.8]-(2S,2′S)(M,M)-(E)-(–)-23a

[CD(–)270.0]-(2S,2′S)(M,M)-(Z)-(–)-23c

Scheme 5.8. Synthesis of chiral dimethyl olefins (−)-23a and (−)-23c.

Figure 5.27. ORTEP drawing of the second-eluted CSDP ester (1S,2S)-(−)-cis-25b. (Redrawn from reference 43, with permission.)

197

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Figure 5.28. ORTEP drawing of dimethyl trans-olefin (2S∗ ,2 S∗ )-(M∗ ,M∗ )-(E)-(±)-23a. (Redrawn from reference 43, with permission.)

It was surprising to see the CD spectrum of (−)-23a, which showed several complex negative Cotton effects in the 1 Bb transition region at 200–270 nm (Figure 5.29). The CD pattern is much different from that of previous chiral six-membered olefins. On the other hand, in the 1 La transition region at 300–400 nm, a broad positive Cotton effect was observed. Since the CD spectrum showed a negative Cotton effect at 257.8 nm, the enantiomer was designated as [CD(–)257.8]-(2S,2 S)-(M,M)-(E )-(−)-23a.

obsd CD 365.8 ( +16.7) 349.6 ( +18.2) 295.8 ( –10.5) 257.8 (–140.0) 247.8 ( –57.1) 226.6 (–100.2) 214.2 (–109.6)

200 CD Δe

H

CH3

e × 10–4

0

–100 CH3 –200

H

[CD(–)257.8]-(2S,2′S)(M,M)-(E) obsd UV 367.6 (25,600) 352.0 (25,900) 243.6 (39,200) 216.4 (87,700)

UV

200

10

in MeOH

300

400 λ (nm)

5

Figure 5.29. CD and UV spectra of stable dimethyl trans-olefin

0 500

[CD(–)257.8]-(2S,2 S)-(M,M)-(E)-(−)-23a in MeOH. (Redrawn from reference [43], with permission.)

198

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The stereochemistry of the other olefin (−)-23c was determined to be (2S,2 S)(M,M)-(Z )-(−)-23c by 1 H NMR analysis and also by the fact that the UV irradiation of trans-olefin (−)-23a yielded cis-olefin (−)-23c. The CD and UV spectra of cis-olefin (−)-23c are shown in Figure 5.30, where in the 1 La transition region at 350–400 nm a broad negative Cotton effect was observed. On the other hand, in the 1 Bb transition region at 200–290 nm, strong negative and positive Cotton effects were obtained. Since the CD Cotton at 270 nm was negative, this enantiomer was fully designated as [CD(–)270.0](2S,2 S)-(M,M)-(Z )-(−)-23c. For the cis-olefin [CD(–)269.8]-23c, the absolute stereochemistry had been previously assigned by the Feringa group as (2R,2 R)-(P, P)-(Z )-23c by comparison with the CD spectrum of the six-membered ring compound [CD(–)238.0]-(3R,3 R)-(P, P)-(Z )-16 in Figure 5.23. They considered that the pattern of negative (270.0 nm)/positive (232.0 nm) bands of [CD(–)269.8]-23c was similar to that of the negative (238.0 nm)/positive (223.4 nm) bands of six-membered cis-olefin 16 [45]. Since their assignment was thus opposite to ours, it was concluded that such comparison of CD spectra led to the erroneous assignment [43].

5.5.2. Light-Powered Chiral Molecular Motor with Higher Speed of Rotation: Isolation or 1 H NMR Detection of the Unstable Motor Rotation Isomers and Their CD Spectra The new chiral olefins with five-membered ring systems worked as a light-powered chiral molecular motor with higher speed of rotation as shown in Scheme 5.9. To clarify the

100

obsd CD 379.2 ( –6.5) 304.2 ( –21.5) 291.8 ( –23.9) 270.0 (–159.6) 232.0 (+116.2) 223.8 ( –21.0) 215.4 ( +42.7) 210.2 ( +30.5)

CD

Δe

e × 10–4

0

–100

CH3 –200

in MeOH

UV

200

H

H

CH3

[CD(–)270.0]-(2S,2′S)(M,M)-(Z) obsd UV 369.6 (15,300) 330.4 ( 7,000) 305.8 ( 8,200) 294.4 ( 6,500) 254.2 (28,200) 221.4 (72,100)

300

400 λ (nm)

10

5

Figure 5.30. CD and UV spectra of stable 0 500

dimethyl cis-olefin [CD(−) 270.0]-(2S,2 S)(M,M)-(Z)-(−)-23c in MeOH. (Redrawn from reference 43, with permission.)

199

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

hn

Δ hn

H

CH3

CH3

CH3

H

H

CH3

[CD(+)279.2]-(2S,2′S)(P,P)-(Z)-23b

H CH3

Δ hn

[CD(–)257.8]-(2S,2′S)(M,M)-(E)-(–)-23a

H

CH3

H

H CH3

[CD(–)270.0]-(2S,2′S)(M,M)-(Z)-(–)-23c hn

H

Scheme 5.9. Rotation scheme of a

CH3

light-powered chiral molecular motor of five-membered ring type rotating with a higher speed.

[CD(+)269.0]-(2S,2′S)(P,P)-(E)-23d

motor rotation mechanism, we first tried to isolate the unstable cis-olefin 23b as follows [44]. A solution of the stable trans-olefin 23a in CH2 Cl2 was irradiated with UV light at 312 nm at −78◦ C, and the reaction mixture was subjected to HPLC (ODS, MeOH) at −40◦ C. The desired unstable cis-olefin 23b was obtained as a yellow powder of the second-eluted fraction, and the structure of 23b was determined by 1 H NMR spectra measured at −30◦ C. To avoid the thermal isomerization, the CD spectrum of the unstable cis-olefin (2S,2 S)-(P, P)-(Z )-23b was measured in MeOH at −32.5◦ C as illustrated in Figure 5.31, where very complex and intense Cotton effects were observed in the 1 Bb transition region (200–300 nm), reflecting the strongly twisted π -electron structure. In addition, a weak

Δε +100 CD +50

H H CH3 CH3 [CD(+)279.2]-(2S,2′S)(P,P)-(Z)

0

–50 in MeOH at –32.5°C –100

–150 200

300

obsd CD 400.6 ( +8.8) 279.2 (+100.6) 261.0 ( –71.3) 233.0 (–144.1) 221.6 ( +70.9) 400

λ (nm)

Figure 5.31. CD spectrum of the unstable 500

cis-olefin [CD(+)279.2]-(2S,2 S)-(P, P)-(Z)-23b in MeOH at −32.5◦ C. (Redrawn from reference 44, with permission.)

200

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

positive broad Cotton effect appeared at 350–450 nm, and the strong red-shift of the electronic transition caused the yellow color of this unstable motor rotation isomer 23b. Since the unstable cis-olefin (2S,2 S)-(P, P)-(Z )-23b shows an intense positive CD band at 279.2 nm, its AC was fully designated as [CD(+)279.2]-(2S,2 S)-(P, P)-(Z )-23b. We also attempted to isolate the other unstable motor rotation isomer 23d in a similar manner. A solution of the stable cis-olefin (−)-23c in MeOH was irradiated with UV light at 330 nm at −78◦ C for 18 s. The photochemical reaction was monitored by CD spectroscopy performed at −62◦ C to find that the reaction had reached a photoequilibrium state between 23c and 23d. The equilibrium ratio was determined to be 23d/23c = 93 : 7 by 1 H NMR spectra measured at −60◦ C. The extremely unstable trans-olefin (2S,2 S)-(P, P)-(E )-23d shows intense and complex CD Cotton effects in the 1 Bb transition region (200–290 nm), reflecting the strongly twisted π -electron structure, while at 300–450 nm a negative broad CD band was observed (Figure 5.32). Since the unstable trans-olefin 23d shows a positive CD band at 269.0 nm, its AC was designated as [CD(+)269.0]-(2S,2 S)-(P, P)-(E )-23d.

5.5.3. Light-Powered Chiral Molecular Motor with Higher Speed of Rotation: Dynamics of Motor Rotation Studied by CD Spectroscopy To clarify the motor rotation mechanism, the photochemical and thermal reactions of olefin 23 were studied by CD and 1 H NMR spectroscopy as follows [44]. (1) The Photochemical First Motor Rotation Step. A solution of stable transolefin (2S,2 S)-(M,M)-(E )-(−)-23a in MeOH was irradiated with UV light at 312 nm at −78◦ C, and the change was monitored by CD measured at −25◦ C (Figure 5.33). The photochemical reaction was thus very fast and reached photoequilibrium after 20.5 s of irradiation. The 23b/23a ratio at the final stage was 94:6. The reverse reaction 23b → 23a was similarly studied by CD spectroscopy, where a solution of unstable cis-olefin (2S,2 S)-(P, P)-(Z )-23b in MeOH was irradiated with

+150 Δε H

+100 CD CH3

+50

H

[CD(+)269.0]-(2S,2′S)(P,P)-(E)

0

obsd CD 386.0 ( –35.0) 269.0 ( +52.4) 255.4 ( +41.4) 240.0 ( –7.9) 230.4 (+137.1) 220.2 ( –30.2) 214.8 ( +67.7)

–50 in MeOH at –62.0 °C –100

200

CH3

300

400 λ (nm)

500

Figure 5.32. CD spectrum of the unstable trans-olefin [CD(+)269.0]-(2S,2 S)-(P, P)-(E)-23d (93%) in MeOH at −62.0◦ C; the sample contained some of the stable cis-olefin (2S,2 S)-(M,M)-(Z)-23c (7%); the ratio was determined by 1 H NMR spectroscopy. (Redrawn from reference 44, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Δε 233.0 nm

201

in MeOH –25 °C

+100

7

+50 2

1

0

1 2

1, 0.00 sec 2, 2.62 sec 3, 5.09 sec 4, 7.06 sec 5, 10.37 sec 6, 20.50 sec 7, unstable

7

–50

–100 1

7 220

240

260

280

300

Figure 5.33. CD spectral change due to the photoisomerization of the stable trans-olefin (2S,2 S)-(M,M)-(E)-23a into the unstable cis-olefin (2S,2 S)-(P, P)-(Z)-23b on UV irradiation in MeOH at 312 nm at −78◦ C, as monitored by CD at

320

λ (nm)

−25◦ C. (Redrawn from reference 44, with permission.)

visible light (430 nm) at −78◦ C. After 65 sec irradiation, the reaction 23b → 23a was complete. (2) The Thermal Second Motor Rotation Step. The thermal reaction 23b → 23c was also monitored by CD spectroscopy, where a solution of unstable cis-olefin (2S,2 S)-(P, P)-(Z )-23b in MeOH was kept at 14.8◦ C and CD spectra were measured at intervals of 1 h (Figure 5.34). Based on the Arrhenius and Eyring plots, the dynamics data of the thermal molecular motor rotation step 23b → 23c were determined. The data obtained by CD spectroscopy agreed well with those by 1 H NMR spectroscopy.

Δε 13

+100

in MeOH 14.8 °C

1

+50 0 13

–50 –100

1, 2, 3, 4, 5, 6, 7,

2 1

–150

233.8 nm 220

240

260

280 λ (nm)

0.0 h 1.0 h 2.0 h 3.0 h 4.0 h 5.0 h 6.0 h 300

8, 7.0 h 9, 8.0 h 10, 9.0 h 11, 10.0 h 12, 11.0 h 13, stable cis 320

Figure 5.34. CD spectral change due to the thermal isomerization of the unstable cis-olefin (2S,2 S)-(P, P)-(Z)-23b into the stable cis-olefin (2S,2 S)-(M,M)-(Z)-23c at 14.8◦ C. (Redrawn from reference 44, with permission.)

202

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(3) The Photochemical Third Motor Rotation Step. A solution of stable cisolefin (2S,2 S)-(M,M)-(Z )-(−)-23c in MeOH was irradiated with UV light at 330 nm at −78◦ C giving unstable trans-olefin (2S,2 S)-(P, P)-(E )-23d, and the change was monitored by CD at −60◦ C (Figure 5.35). The photochemical reaction was also very fast and reached photoequilibrium after 23.0 s of irradiation. The 23d/23c ratio at the final stage was 93:7. The low-temperature CD spectra (−60◦ C) are very useful for studying the photodynamics of extremely unstable motor rotation isomer. (4) The Thermal Fourth Motor Rotation Step. The thermal reaction 23d → 23a was also monitored by CD spectroscopy as follows; CD spectra of the solution 23d/23c (93:7) in MeOH was measured at −19.1◦ C at intervals of 1 h (Figure 5.36). The thermal step 23d → 23a occurred even at subzero temperatures like −19.1◦ C, indicating that the fourth motor rotation is much faster than that of six-membered ring motor described above. This five-membered ring molecular motor thus rotates very fast as predicted. From the Arrhenius and Eyring plots, the dynamics data of the thermal fourth molecular motor rotation step 23d → 23a were determined. The molecular motor rotation dynamics data are summarized in Table 5.1. (i) The first photochemical rotation step 23a → 23b is much faster under the conditions of CD measurement than under the conditions of 1 H NMR measurement, because of the photoirradiation efficiency. That is, the speed of the photochemical rotation step essentially depends on the photoirradiation conditions. As the motor can rotate backward (e.g., 23b → 23a), the direction of motor rotation has to be controlled by choosing the optimal wavelength. Thus, the motor rotates forward on UV irradiation at 312 nm, reaching photoequilibrium in 20 s with the ratio 23b/23a = 94 : 6. (ii) The second step of motor rotation 23b → 23c is thermally controlled, and the dynamics data obtained by 1 H NMR and CD methods are similar. (iii) The third photochemical rotation step 23c → 23d was again faster under the conditions of CD measurement, where the motor rotates forward on UV irra-

Δε in MeOH

–60 °C

+100 1 5

+50 0 5

3

–50 –100 1

–150

1, 0.00 sec 2, 0.69 sec 3, 1.86 sec 4, 4.22 sec 5, 13.78 sec

270.0 nm 220

240

260

280 λ (nm)

300

320

Figure 5.35. CD spectral change due to the photoisomerization of the stable cis-olefin (2S,2 S)-(M,M)-(Z)-23c into the unstable trans-olefin (2S,2 S)-(P, P)-(E)-23d on UV irradiation in MeOH at 330 nm at −78◦ C, as monitored by CD at −60◦ C. (Redrawn from reference 44, with permission.)

203

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Δε in MeOH

–19.1 °C

+50

1

0

1, 2, 3, 4, 5, 6, 7, 8,

–50

8

–100

258.2 nm 220

240

260

280

0.0 h 1.0 h 2.0 h 3.0 h 4.0 h 5.0 h 6.0 h stable trans 300

320

λ (nm)

Figure 5.36. CD spectral change due to the thermal isomerization of the unstable trans-olefin (2S,2 S)-(P, P)-(E)-23d into the stable trans-olefin (2S,2 S)-(M,M)-(E)-23a at −19.1◦ C. (Redrawn from reference 44, with permission.)

TAB L E 5.1. The Molecular Motor Rotation Dynamics Data as Monitored by 1 H NMR and CD Methods Motor Rotation Forward rotationa 23a → 23b, hν, 312 nmb Backward rotationa 23b → 23a, hν, 430 nmb Forward rotationc 23b → 23c, thermal Forward rotationa 23c → 23d, hν, 330 nmb Backward rotationa 23d → 23c, hν, 430 nmb Forward rotationc 23d → 23a, thermal

a

1

H NMR 400 MHz in CD2 Cl2 15 min 23b/23a = 91 : 9 4 min 23b/23a = 0 : 100 Ea = 20.7 H = = 20.1 S = = −6.17 R = 0.999, 0.999 12 min 23d/23c = 96 : 4 60 min 23d/23c = 25 : 75 Ea = 17.1 H = = 16.5 S = = −9.23 R = 0.999, 0.999

CD in MeOH 20.5 s 23b/23a = 94 : 6 90 s 23b/23a = 0 : 100 Ea = 21.4 H = = 20.8 S = = −6.30 R = 0.999, 0.999 23 s 23d/23c = 93 : 7 90 s 23d/23c = 20 : 80 Ea = 16.8 H = = 16.3 S = = −12.6 R = 0.975, 0.974

Time to photoequilibrium. Ratio at photoequilibrium. c Activation energy Ea and activation enthalpy H = in kcal mol−1 unit: activation entropy S = in cal K−1 mol−1 unit; the value of correlation coefficient R for Arrhenius and Eyring plots, respectively. Source: Adapted from reference [44], with permission. b

204

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

diation at 330 nm, reaching the photoequilibrium state in 23 s with the ratio 23d/23c = 93 : 7. (iv) The fourth thermal rotation step 23d → 23a was similarly monitored by 1 H NMR and CD methods, but in this case the CD study was difficult because of the instability of compound 23d. However, it should be noted that both methods gave similar values of kinetic parameters as seen in Table 5.1. The activation energy of the fourth thermal step was much lower than that found for the motor of sixmembered ring type, and hence the new motor rotates much faster than the old one.

5.5.4. Continuous Rotation of the New Light-Powered Chiral Molecular Motor Studied by CD Spectroscopy Continuous rotation experiments were carried out as follows: (i) For the first rotation step, a solution of the enantiopure stable trans-olefin 23a in n-pentanol was irradiated with UV light at 312 nm at −78◦ C for 30 s; (ii) for the second rotation step, the solution was heated at 120◦ C for 20 s; (iii) for the third rotation step, the solution was irradiated with UV light at 330 nm at −78◦ C for 30 s; (iv) for the fourth rotation step, the solution was heated at 120◦ C for 10 s. After each operation, the CD spectrum of the solution was measured at −50◦ C, and the CD intensity at 275 nm was plotted. One cycle of operations (namely, 360◦ rotation of motor) took 90 s under these conditions. Figure 5.37 shows the CD plot for 10 cycles of rotation, indicating that this new molecular motor rotates continuously and is durable for such operations.

5.6. ABSOLUTE CONFIGURATION OF CHIRAL C60 -FULLERENE CIS-3 BISADDUCTS DETERMINED BY X-RAY CRYSTALLOGRAPHY AND CD SPECTROSCOPY The fullerene C60 is a symmetrical and achiral molecule. However, an addition reaction at two chiral positions of the C60 skeleton (e.g. a cis-3 addition) makes the π -electron 1

2

3

4

5

6

7

8

9

10 cycles

+60 +40 +20 Δε 0 –20 –40 –60 –80 –100

Figure 5.37. Continuous rotation of the 0

200

400

600

Time (sec)

800

new molecular motor 23 as monitored by CD. (Redrawn from reference 44, with permission.)

205

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

system in fullerene chiral. Synthetic and absolute configurational studies of various chiral fullerene cis-3 bisadducts have been carried out by many research groups as listed in Chart 5.5 [46]. Compound [CD(–)288]-27 was synthesized as the first chiral cis-3 bisadduct by using a chiral tether, and its AC was tentatively assigned as (R, R, f,s A) {≡ (R, R, f C ), old nomenclature} on the basis of MM2/Monte Carlo calculations, because it was assumed that the product formed should be the most energetically stable diastereomer. Therefore, the stereochemistry of the most stable diastereomer calculated was assigned as shown in Chart 5.5 [47]. In a similar manner, the AC of compound [CD(+)281]-28 was determined as shown by assuming that the product formed should have the most energetically stable stereostructure. Thus the diastereomeric structure (S,S,f,s A) was assigned as the most stable one by molecular mechanics calculation [48]. In addition, the CD spectrum of compound (S,S,f,s A)-28 was calculated by the π electron SCF-CI-DV MO method to compare with the observed CD spectrum [49]. By comparison of the data, the AC of bisadduct [CD(+)281]-28 was determined to be (S,S,f,s A). Later, the same AC was assigned by applying the CD exciton chirality method to a related compound to confirm the previous assignment of 28 [50].

O

O

O

H

H

H

O O

O

(R,R,f,sA)-[CD(–)288]-27* energy calculation (1996)

O

B

O

O

B

H

CH3O O

O

H3C CH3 H H

O

O

O

OCH3

O

O

O

EtO

OEt

(S,S,f,sA)-[CD(+)281]-28 energy calculation (1997) CD calculation (1998) Exciton CD (2000)

(S,S,f,sC)-[CD(+)284]-29* energy calculation (1997–8) revised (2003)

H3CO H H OCH3

H3CO H H OCH3

O

(f,sC)-[CD(−)287]-30 comparison of CD (1999)

O O

O

O O

O

O

O O

O

O

O

O

O

EtO

OEt

EtO

OEt

(S,S,f,sA)-[CD(+)281]-31a 1H NMR analysis (2002)

(S,S,f,sC)-[CD(−)281]-31b 1H NMR analysis (2002)

Chart 5.5. Previously reported chiral C60 -fullerene cis-3 adducts together with their CD data and absolute configurations, which are designated by the new systematic nomenclature (f,s C and f,s A). Our studies described here suggested that the ACs of compounds 27 and 29 should be revised.

206

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The AC of cis-3 bisadduct [CD(+)284]-29 was tentatively assigned as shown in Chart 5.5 by calculation of the stable diastereomeric structures and their energy levels, because the tether part was synthesized from (2S,3S )-butanediol, and the stereostructure (S,S,f,s C ) was calculated to be the most stable isomer [51]. Compound [CD(–)287]-30 has no chirality center in the tether moiety, and therefore its AC was determined to be (f,s C ) by comparison of CD spectrum with that of compound [CD(+)281]-(S,S,f,s A)-28 [52]. That is, the CD spectra of [CD(–)287]-30 and [CD(+)281]-(S,S,f,s A)-28 were opposite to each other. Regarding the ACs of these chiral fullerene derivatives, a serious problem was raised as follows. Among compounds 27–30, their tether moieties are different from each other, but the remaining C60 chiral chromophores with the cis-3 bisadduct pattern are the same or mirror images of each other. As listed in Table 5.2, however, the CD spectrum of (R,R,f,s A)-27 is opposite to that of (S,S,f,s A)-28, although they have the same (f,s A) AC of the fullerene moiety. On the other hand, compounds (S,S,f,s A)-28 and (S,S,f,s C )-29 have the opposite ACs in the fullerene part, but they exhibited similar CD spectra as seen in Table 5.2. These results strongly indicated that some absolute configurational assignments were wrong. To solve this problem, a different strategy was taken as follows. In the previous syntheses, only one diastereomer of two possible products was isolated and it was assumed to be the most stable diastereomer. This strategy brought some ambiguity in the determination of AC. To overcome these difficulties, we selected a tether with a more flexible conformation to yield two possible disatereomeric products in the synthesis. That is, we thought that it was relatively easy to determine the relative stereochemistry by comparing the 1 H NMR data of two diastereomers [53]. We were able to synthesize the two possible diastereomers [CD(+)281]-31a and [CD(–)281]-31b starting from (2S,3S )-(−)-2,3-butanediol, and these diastereomers were separable by HPLC on silica gel (Chart 5.5). Their CD spectra were almost mirror images of one another, reflecting the opposite chirality of the π -electron system in the two C60 skeletons. Careful analysis of the 1 H NMR data (i.e., chemical shift and coupling constant) led to the absolute configurational assignments, (S,S,f,s A)-[CD(+)281]-31a and (S,S,f,s C )-[CD(–)281]-31b [53]; these results thus confirmed the AC of (S,S,f,s A)[CD(+)281]-28 reported previously by the Diederich and our groups [48–50]. Although the new assignment based on the 1 H NMR analyses of two diastereomers was more reliable than the previous ones using only one diastereomer, we wanted to confirm our results by X-ray crystallography.

TAB L E 5.2. Reported CD Data and ACs of Chiral Fullerene cis-3 Bisadducts Compound

27

28

29

30

AC CD λ (nm) ε CD λ (nm) ε References

(R,R,f,s A)a 720.0 +4.69c 288.0 −75.0c 47

(S,S,f,s A) 706.0 −37.0 281.0 +89.0 48–50

(S,S,f,s C )b 737.0 −36.3 284.0 +60.6 51

(f,s C ) 726.0 +26.7 287.0 −73.2 52

a The

revision of AC was later suggested; see the discussion below. was later revised [51]. c Values are 1/1000 of the ε values reported in reference 47, which are too large. b AC

207

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Although it was known that most fullerene derivatives were obtained as amorphous solids, we attempted to get single crystals of various chiral cis-3 bisadducts, and we finally succeeded in obtaining single crystals of diethyl ester [CD(+)280]-32, which was previously synthesized by the Diederich group [50]. As seen in Scheme 5.10, bisadduct [CD(+)280]-32 was prepared starting from (2R,3R)-(−)-2,3-butanediol, the moiety of which would be useful as the internal reference of absolute configuration in X-ray crystallography. The product [CD(+)280]-32 was recrystallized from chloroform/hexane (1:1), giving extremely thin, red plate crystals with a thickness of 1–2 μm, which were too thin for conventional X-ray diffractometers [54]. Therefore, the X-ray diffraction experiment was carried out with extremely strong synchrotron radiation at the SPring-8 in Hyogo ˚ space group P 21 (#4), R = 0.180. Although (Japan): X ray, 22.00 keV, λ = 0.5633 A, the final R value remained large, the AC of bisadduct [CD(+)280]-32 was unambiguously determined as (f,s A) by using the (2R,3R) absolute configuration of the tether moiety as an internal reference (Figure 5.38). Thus the use of the internal reference method in X-ray crystallographic analysis is very useful for unambiguous determination of AC [54].

H3C H H CH3 HO

H3C H H CH3

O

O

EtO

O

O

H3C H H CH3 O O O O

O OEt

O

OH

(2R,3R)-(–)2,3-butanediol

O OEt EtO

O

(2R,3R)-tether

Scheme 5.10. Synthesis of chiral cis-3 bisadduct (R,R,f,sA)-[CD(+)280]-32

(R,R,f,s A)-[CD(+)280]-32.

a

Figure 5.38. Absolute stereo-structure of the C60 fullerene cis-3 bisadduct (R,R,f,s A)[CD(+)280]-32 (top) and projection along b-axis

o

c Projection along b-axis

(bottom). (Redrawn from reference 54, with permission.) (See insert for color representation of the figure.)

208

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The CD and UV–Vis spectra of bisadduct (R,R,f,s A)-32 are shown in Figure 5.39, where the CD spectrum shows an intense positive Cotton effect (ε +93.7) at 280 nm. In addition, compound (R,R,f,s A)-32 exhibited an anomalously large optical rotation [α]28 D +3950 (c = 0.0214, CHCl3 ). Therefore, its AC was designated as (R,R,f,s A)-[CD(+)280]-(+)-32 [54]. Based on these X-ray and CD results, the ACs of C60 fullerene cis-3 bisadducts were rationalized as follows: cis-3 derivatives showing a positive CD band around 280 nm should have the (f,s A) absolute configuration, while cis-3 compounds showing a negative CD around 280 nm should have the (f,s C ) absolute configuration. Therefore, our previous assignments of (S,S,f,s A)-[CD(+)281]-31a and (S,S,f,s C )-[CD(−)281]-31b were corroborated by this study. The assignment of (S,S,f,s A)-[CD(+)281]-28 by Diederich and co-workers was also confirmed. However, it was concluded that the AC of [CD(−)288]27 should be revised to be (R,R,f,s C ) [54]. In a similar manner, revision of the AC was suggested so that cis-3 bisadduct [CD(+)284]-29 should have the (S,S,f,s A) absolute configuration. We found unique phenomena in the CD and UV–Vis spectra as follows. Compound (R,R,f,s A)-[CD(+)280]-(+)-32 exhibits a very weak absorption band at 706.2 nm (ε = 338); this band is due to the forbidden π –π ∗ transition as a result of the small ε value. In the corresponding region, the CD spectrum shows an intense negative Cotton effect at 701.5 nm (ε −36.3). The curve of g value (g = ε/ε) was calculated as illustrated in Figure 5.39, where the maximum value found was g = −0.110 at 697.0 nm [54]. This g value is much larger than that of n–π ∗ forbidden transition of chiral ketones and that of

+0.10

+100

280.0 (+93.7) CD +0.05

+50

701.5 (–36.3) 0 Δe/e

Δe 0

–50

–100

–150

ε × 10–4

10

5

220

obsd CD g-value 701.5 (–36.3) 639.5 (–10.0) 697.0 (–0.110) 593.0 (+5.6) 486.5 (+21.5) g-value 424.5 (–20.1) 697.0 (–0.110) 412.0 (+3.0) 398.0 (–13.3) 635.5 (–0.0191) 378.5 (–32.4) 593.0 (+0.00752) 339.0 (+75.8) 490.0 (+0.0130) 313.0 (+37.5) 425.5 (–0.00717) 280.0 (+93.7) 380.5 (–0.00425) 263.0 (–151.2) 343.0 (–0.00298) 232.0 (–47.9) 706.2 (338) obsd UV UV 706.2 (338) 643.6 (517) × 100 316.5 (37100) 254.5 (106600) 300

400

500 λ (nm)

600

700

–0.05

–0.1

–0.15

Figure 5.39. CD and UV–Vis spectra and g-value curve of cis-3 bisadduct (R,R,f,s A)[CD(+)280]-(+)-32 in ClCH2 CH2 Cl: CD and

800

UV–Vis, top curve; g value (g = ε/ε), bottom curve. (Redrawn from reference 54, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

209

the π –π ∗ allowed transition of other compounds. For example, the n –π ∗ transition of (R)-(+)-3-methylcyclohexanone showed g = +0.03 at 298 nm; π –π ∗ , (+)-hexahelicene, g = +0.007 at 325 nm [55]; π –π ∗ , exciton coupling CD of (2S,3S )-butanediyl bis(4bromobenzoate), g = +0.00035 at 252 nm. A similar large g value was also observed for bisadduct (S,S,f,s A)-[CD(+)281]-31a: g = −0.126 at 696 nm [54]. Why does this transition of chiral cis-3 fullerene make such a large g value? The phenomenon could be interpreted as follows: The π –π ∗ transition is electronically forbidden as discussed above, but is magnetically allowed, because molecular orbitals (MOs) involved are spherical, reflecting the spherical shape of fullerene skeleton. Therefore, the transition contains a lot of angular momenta, thus yielding a large magnetic moment and generating an intense CD Cotton effect. The g-value thus becomes large in this transition [54].

5.7. ABSOLUTE STEREOCHEMISTRY AND CD SPECTRA OF AN ALLENO-ACETYLENIC MACROCYCLE The alleno-acetylenic tetrameric macrocycle (P, P, P, P)-(−)-33 is a unique chiral compound devoid of any chirality center, in which each allene moiety takes a P helicity, and hence the compound takes a D4 symmetric structure as illustrated in Figure 5.40 [56, 57]. Chiral macrocycle (P, P, P, P)-(−)-33 was first synthesized by Diederich and coworkers starting from tert-alcohol 34 as shown in Scheme 5.11 [56], where racemic alcohol (±)-34 was resolved as camphanate esters. The absolute configuration of alcohol (R)-(−)-34 was determined by X-ray crystallographic analysis of its camphanate ester, where the camphanate group was used as an internal reference of the absolute configuration [58]. The other enantiomer (S )-34 was converted to chiral allene derivative (P )-(+)-35 via a rearrangement reaction. The obtained chiral alleno-acetylene (P )-(+)35 was then dimerized, followed by deprotection, to yield a chiral dimer (P, P)-(+)-36, [α]D 20 +506 (c 1, CHCl3 ), the absolute configuration of which was determined by X-ray crystallographic analysis of bis[Si(iso-Pr)3 ] derivative 37 of the other (M,M)-enantiomer series, where the anomalous dispersion effect of silicon atoms was used [59]. The further dimerization–cyclization of bis(acetylene) (P, P)-(+)-36 furnished the target chiral macrocycle (P, P, P, P)-(−)-33, [α]D 20 −770 (c 1, CHCl3 ), the relative crown configuration of which was determined as shown by X-ray analysis of racemate (±)-33 (Scheme 5.11) [57]. The absolute stereochemistry of macrocycle (P, P, P, P)-(−)-33 was thus unambiguously determined by X-ray crystallography.

Figure 5.40. A novel unique chiral alleno-actylenic (P,P,P,P)-(–)

macrocycle (P, P, P, P)-(−)-33 and its crown stereostructure. (Redrawn from reference 56, with permission.)

210

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HO

OH

H H (S)-34

(P)-(+)-35

abs. config. by X-ray of camphanate ester of (R)-34

R

R

(P,P)-(+)-36: R = H abs. config. by X-ray of Si(iso-Pr)3 derivative of (M,M)-series

(P,P,P,P)-(–)-33 relat. config. by X-ray of (+)-33 –

Scheme 5.11. Synthesis of macrocycle (P,P,P,P)-(−)-33 and absolute configurational assignment by X-ray crystallographic analyses of related derivatives.

It is interesting that macrocycle (P, P, P, P)-(−)-33 shows extremely intense CD Cotton effects as illustrated in Figure 5.41; for example, the positive CD band at 253 nm has an intensity of ε = +790, which is about 100 times larger than that of monomer (P )-(+)-35 and ≈ 8 times larger than that of dimer (P, P)-(+)-36. The CD spectrum of macrocycle (P, P, P, P)-(−)-33 is thus unique and a good example for CD studies.

+800 CD

(P,P,P,P)-(–) e × 10–5

+400 Δe 0 –400

3

–800 2

UV

1

Figure 5.41. CD and UV–Vis spectra of enantiopure macrocycle (P, P, P, P)-(−)-33 (thick

200

300 λ (nm)

0 400

line) and CD spectra of (M, M, M, M)-(+)-33 (thin line) in hexane. (Redrawn from reference 56, with permission.)

211

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

The UV spectrum of (P,P,P,P)-(−)-33 shows three maxima of medium intensity around 275–375 nm, which were assigned to the vibronic structure of conjugated acetylene groups because of the vibronic interval of ≈ 2100 cm−1 (Figure 5.41). The three UV bands around 275–375 nm were thus assigned due to a single π –π ∗ transition of medium intensity, but not due to three different π –π ∗ transitions. On the other hand, a very intense UV band is seen at 245 nm (ε 210,000), which was assigned to an allowed π –π ∗ transition [56]. In the region of 270–350 nm, the CD spectrum of (P, P, P, P)-(−)-33 shows three negative extrema corresponding to the UV vibronic structure, while a very intense positive Cotton effect is observed in the region of 230–270 nm (Figure 5.41). The g values (g = ε/ε) of these Cotton effects were calculated to give interesting data as follows: g = approximately −0.007 to −0.008 for the Cotton effects in the region of 270–350 nm; g = approximately +0.004 for the intense Cotton effect at 253 nm. These data clearly indicate that the CD bands in the region of 270–350 nm have a large magnetic transition dipole moment contribution, and the transition is magnetically allowed. On the other hand, the CD band at 253 nm has a large electric transition dipole moment contribution, indicating the character of electric allowed transition [56]. To give insight into the CD mechanism of macrocycle (P, P, P, P)-(−)-33 and also to determine its absolute stereochemistry in a theoretical manner, the CD and UV spectra were calculated by the ZINDO MO method. As illustrated in Figure 5.42, the basic pattern of the CD curve including the sign, amplitude, and position of Cotton effects, but not the vibronic structures around 270–350 nm, was reproduced well by the ZINDO calculation. The (P,P,P,P ) absolute stereochemistry of macrocycle (−)-33 was thus determined by the MO calculation of the ZINDO level, and the theoretical absolute configurational assignment was consistent with the experimental one [56]. The calculation results show that there are three major electronic transitions as shown in Figure 5.42. The negative CD bands at 270–350 nm are due to the S1 transition, while the positive CD band at 253 nm is due to two degenerate S2 and S3 transitions. The S1 transition is very unique, since its large magnetic transition dipole moment (MTDM) is perpendicular to the ring plane of the macrocycle and is oriented upward, as seen in the

MTDM

+800 +600 +400

R × 1037 cgs unit

Δe

S2,S3 ETDM

+2.0

transition S1

+200

+1.0

0 –200

0

Figure 5.42. Comparison of the observed CD

–1.0

spectrum of (−)-33 (thick line) with the CD curve of (P, P, P, P)-33 (thin line) calculated by the ZINDO MO method, where the graphic inserted

obsd CD –400 S1

–600 200

–2.0

calcd CD

300 λ (nm)

–3.0 400

above shows the electric transition dipole moment (ETDM) and magnetic transition dipole moment (MTDM) of the transition S1 in the geometry of the compound. (Redrawn from reference 56, with permission.)

212

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

graphic inserted in Figure 5.42. This large MTDM is generated by the cyclic movement of an electron along the macrocycle ring during the electronic transition. At the same time, the S1 transition generates a small electric transition dipole moment (ETDM), which is also perpendicular to the ring plane, but is oriented downward. Therefore, MTDM and ETDM are antiparallel to each other and take nonzero values, generating an intense negative CD band at the S1 transition. The ETDM and MTDM of the S2 transition are placed in the macrocycle plane and oriented parallel to each other generating a positive CD band (Figure 5.42). Both ETDM and MTDM take nonzero values, which is the main reason for the intense CD at the S2 transition. The S3 transition has a similar character to that of the S2 transition, generating an intense positive CD. The ETDMs of S2 and S3 transitions are perpendicular to each other reflecting the D4 symmetric structure of macrocycle 33. This is the reason that the S2 and S3 transitions are degenerate. The mechanism of intense CD of macrocycle (P, P, P, P)-(−)-33 was thus clarified even by the MO calculation of ZINDO level [56]. A similar result was obtained by the more advanced MO method, that is, the Coulombattenuated hybrid exchange-correlation functional (CAM-B3LYP), supporting the CD mechanism discussed above [57].

5.8. CONCLUSION As discussed above, the ACs of various chiral compounds with extended π -electron system, including natural products and synthetic compounds, have been theoretically determined on the basis of the calculation of their CD spectra by the π -electron SCF-CIDV MO method. The ACs theoretically assigned were proved experimentally by X-ray crystallographic analyses using an internal reference of AC, and/or by the total synthesis of natural enantiomers. The combination of X-ray crystallography and CD spectroscopy is thus very reliable for determining ACs. In addition, in the case of the light-powered chiral molecular motors, CD spectroscopy is useful for clarifying the motor rotation mechanism and dynamics. Therefore, the CD methodology discussed here is a promising and powerful tool for determining the ACs of various chiral compounds with an extended and twisted π -electron chromophore.

ACKNOWLEDGMENTS The authors sincerely thank the co-workers of the studies described here for their contributions, whose names are listed in references, and Dr. George A. Ellestad, Department of Chemistry, Columbia University, for his valuable suggestions.

REFERENCES 1. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy—Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, and Oxford University Press, Oxford, 1983. 2. N. Harada, Circular dichroism spectroscopy and absolute stereochemistry of biologically active compounds, in The Biology—Chemistry Interface, R. Cooper and J. K. Snyder, eds., Marcel Dekker, Inc., New York, 1999, Chapter 6, pp. 139–190.

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

3. N. Harada, Circular dichroism of twisted π -electron systems: theoretical determination of the absolute stereochemistry of natural products and chiral synthetic organic compounds, in Circular Dichroism: Principles and Applications, K. Nakanishi, N. Berova, and R. Woody, eds., John Wiley & Sons, New York, 2000, pp. 431–457. 4. N. Berova, N. Harada, K. Nakanishi, Electronic Spectroscopy: Exciton coupling, theory and applications, in Encyclopedia of Spectroscopy & Spectrometry, J. Lindon, G. Tranter, and J. Holmes, eds., Academic Press, London, 2000, pp. 470–488. 5. N. Berova, G. A. Ellestad, N. Harada, Modern methods in natural products chemistry: characterization by circular dichroism spectroscopy, in Comprehensive Natural Products Chemistry II , L. Mander and H. W. Liu, eds., Elsevier, Oxford, 2010, Chapter 9, pp. 91–146. 6. (a) C. Kemp, S. F. Mason, Tetrahedron 1966, 22 , 629–635. (b) A. Brown, C. Kemp, S. F. Mason, J. Chem. Soc. A 1971, 751–755. c) W. S. Brickell, A. Brown, C. Kemp, S. F. Mason, J. Chem. Soc. A 1971, 756–760. 7. C. Rossini, C. Bertucci, P. Salvadori, M. Zandomeneghi, J. Am. Chem. Soc. 1985, 107 , 17–19. 8. J. Roschster, U. Berg, M. Pierrot, J. Sandstrom, J. Am. Chem. Soc. 1987, 109 , 492–507. 9. N. Harada, Chiral auxiliaries powerful for both enantiomer resolution and determination of absolute configuration by X-ray crystallography, in Topics in Stereochemistry, Vol. 25, S. E. Denmark, J. S. Siegel, eds., John Wiley & Sons, Hoboken, NJ, 2006, Chapter 6, pp. 177–203. 10. N. Harada, Powerful chiral molecular tools for preparation of enantiopure alcohols and simultaneous determination of their absolute configurations by X-ray crystallography and/or 1 H NMR anisotropy methods, in Chirality in Drug Research, Ed. E. Francotte, W. Lindner, WILEYVCH, Weinheim, Germany, 2006, Chapter 9, pp. 283–321. 11. N. Harada, M. Watanabe, S. Kuwahara, Novel chiral derivatizing agents powerful for enantioresolution and determination of absolute stereochemistry by X-ray crystallographic and 1 H NMR anisotropy methods, in Chiral Analysis, K. W. Busch and M. A. Busch, eds., Elsevier Science Publishers, Amsterdam, 2006, Chapter 18, pp. 661–691. 12. N. Harada, Chirality 2008, 20 , 691–723. 13. P. J. Stephens, N. Harada, Chirality 2010, 22 , 229–233. 14. R. Takeda, K. Katoh, J. Am. Chem. Soc. 1983, 105 , 4056–4058. 15. N. Harada, J. Kohori, H. Uda, K. Nakanishi, R. Takeda, J. Am. Chem. Soc. 1985, 107 , 423–428. 16. N. Harada, J. Kohori, H. Uda, K. Toriumi, J. Org. Chem. 1989, 54 , 1820–1826. 17. N. Harada, T. Sugioka, H. Uda, T. Kuriki, Synthesis 1990, 53–56. 18. N. Harada, H. Uda, T. Nozoe, Y. Okamoto, H. Wakabayashi, S. Ishikawa, J. Am. Chem. Soc. 1987, 109 , 1661–1665. 19. Y. Okamoto, S. Honda, H. Yuki, H. Nakamura, Y. Iitaka, T. Nozoe, Chemistry Lett. 1984, 1149–1152. 20. D. M. Roll, P. J. Scheuer, G. K. Matsumoto, J. Clardy, J. Am. Chem. Soc. 1983, 105 , 6177–6178. 21. M. Kobayashi, N. Shimizu, Y. Kyogoku, I. Kitagawa, Chem. Pharm. Bull . 1985, 33 , 1305–1308. 22. H. Nakamura, J. Kobayashi, M. Kobayashi, Y. Ohizumi, Y. Hirata, Chem. Lett. 1985, 713–716. 23. F. J. Schmitz, S. J. Bloor, J. Org. Chem. 1988, 53 , 3922–3925. 24. (a) M. Kobayashi, N. Shimizu, I. Kitagawa, Y. Kyogoku, N. Harada, H. Uda, Tetrahedron Lett. 1985, 26 , 3833–3836. (b) N. Harada, H. Uda, M. Kobayashi, N. Shimizu, I. Kitagawa, J. Am. Chem. Soc. 1989, 111 , 5668–5674. 25. N. L. Allinger, J. Am. Chem. Soc. 1977, 99 , 8127–8134; N. L. Allinger, Y. H. Yuh, QCPE Program No. 395, 1980. 26. N. Harada, T. Sugioka, Y. Ando, H. Uda, T. Kuriki, J. Am. Chem. Soc. 1988, 110 , 8483–8487.

213

214

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

27. N. Harada, T. Sugioka, H. Uda, T. Kuriki, J. Org. Chem. 1990, 55 , 3158–3163. 28. N. Harada, T. Sugioka, H. Uda, T. Kuriki, M. Kobayashi, I. Kitagawa, J. Org. Chem. 1994, 59 , 6606–6613. 29. N. Harada, T. Sugioka, T. Soutome, N. Hiyoshi, H. Uda, T. Kuriki, Tetrahedron Asym. 1995, 6 , 375–376. 30. N. Harada, H. Ono, H. Uda, M. Parveen, N. U-P. Khan, B. Achari, P. K. Dutta, J . Am. Chem. Soc. 1992, 114 , 7687–7692. 31. F.-J. Zhang, G.-Q. Lin, Q.-C. Huang, J. Org. Chem. 1995, 60 , 6427–6430; Additions and corrections in J. Org. Chem. 1996, 61 , 5700. Later, the authors changed to the (aR) absolute configuration; G.-Q. Lin, M. Zhong, Tetrahedron Lett. 1997, 38 , 1087–1990. 32. N. Harada, H.-Y. Li, T. Nehira, M. Hagiwara, Enantiomer 1997, 2 , 353–358. 33. H.-Y. Li, T. Nehira, M. Hagiwara, N. Harada, J. Org. Chem. 1997, 62 , 7222–7227. 34. B. Feringa, H. Wynberg, J. Am. Chem. Soc. 1977, 99 , 602–603. 35. N. Harada, A. Saito, N. Koumura, H. Uda, B. de Lange, W. F. Jager, H. Wynberg, B. L. Feringa, J. Am. Chem. Soc. 1997, 119 , 7241–7248. 36. N. Harada, A. Saito, N. Koumura, D. C. Roe, W. F. Jager, R. W. J. Zijlstra, B. de Lange, B. L. Feringa, J . Am. Chem. Soc. 1997, 119 , 7249–7255. 37. R. W. J. Zijlstra, W. F. Jager, B. de Lange, P. T. van Duijnen, B. L. Feringa, H. Goto, A. Saito, N. Koumura, N. Harada, J. Org. Chem. 1999, 64 , 1667–1674. 38. N. Harada, N. Koumura, and B. L. Feringa, J. Am. Chem. Soc. 1997, 119 , 7256–7264. 39. N. Koumura, N. Harada, Enantiomer 1998, 3 , 251–253. 40. N. Koumura, N. Harada, Chem Lett. 1998, 1151–1152. 41. N. Koumura, Ph.D. Thesis, Tohoku University, March 1999. 42. N. Koumura, R. W. J. Zijlstra, R. A. van Delden, N. Harada, B. L. Feringa, Nature 1999, 401 , 152–155. 43. T. Fujita, S. Kuwahara, N. Harada, Eur. J. Org. Chem. 2005, 4533–4543. 44. S. Kuwahara, T. Fujita, N. Harada, Eur. J. Org. Chem. 2005, 4544–4556. 45. M. K. J. ter Wiel, R. A. van delden, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2003, 125 , 15076–15086. 46. For the nomenclature of regioisomeric fullerene derivatives, see: (a) A. Hirsch, I. Lamparth, H. R. Karfunkel, Angew. Chem. 1994, 106 , 453–455; Angew. Chem. Int. Ed. Engl . 1994, 33 , 437–438; (b) for the nomenclature of chiral fullerene derivatives, see: C. Thilgen, A. Herrmann, F. Diederich, Helv. Chim. Acta 1997, 80 , 183–199; (c) for a new and systematic nomenclature for fullerenes (IUPAC Recommendations 2002), see W. H. Powell, F. Cozzi, G. P. Moss, C. Thilgen, R. J.-R. Hwu, A. Yerin, Pure Appl. Chem. 2002, 74 , 629–695. 47. E. Nakamura, H. Isobe, H. Tokuyama, M. Sawamura, Chem. Commun. 1996, 1747–1748. 48. J.-F. Nierengarten, T. Habicher, R. Kessinger, F. Cardullo, F. Diederich, V. Gramlich, J.-P. Gisselbrecht, C. Boudon, M. Gross, Helv. Chim. Acta 1997, 80 , 2238–2276. 49. H. Goto, N. Harada, J. Crassous, F. Diederich, J. Chem. Soc. Perkin Trans. 2 1998, 1719–1723. 50. R. Kessigner, C. Thilgen, T. Mordasini, F. Diederich, Helv. Chim. Acta 2000, 83 , 3069–3096. 51. (a) M. Taki, S. Sugita, Y. Nakamura, E. Kawashima, E. Yashima, Y. Okamoto, J. Nishimura, J. Am. Chem. Soc. 1997, 119 , 926–932; (b) N. Taki, Y. Nakamura, H. Uehara, M. Sato, J. Nishimura, Enantiomer 1998, 3 , 231–239; (c) revised assignment: Y. Nakamura, K. O-kawa, J. Nishimura, Bull. Chem. Soc. Jpn. 2003, 76 , 865–882. 52. T. Ishi-i, K. Nakashima, S. Shinkai, A. Ikeda, J. Org. Chem. 1999, 64 , 984–990. 53. K. Yoshida, S. Osawa, K. Monde, M. Watanabe, N. Harada, Enantiomer 2002, 7 , 23–32. 54. S. Kuwahara, K. Obata, K. Yoshida, T. Matsumoto, N. Harada, N. Yasuda, Y. Ozawa, K. Toriumi, Angew. Chem. Int. Ed ., 2005, 44 , 2262–2265.

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

55. S. F. Mason, Molecular Optical Activity and the Chiral Discrimination, Cambridge University Press, Cambridge, 1982. 56. J. L. Alonso-G´omez, P. Rivera-Fuentes, N. Harada, N. Berova, F. Diederich, Angew. Chem. Int. Ed . 2009, 48 , 5545–5548. 57. P. Rivera-Fuentes, J. L. Alonso-G´omez, A. G. Petrovic, P. Seiler, F. Santoro, N. Harada, N. Berova, H. S. Rzepa, F. Diederich, Chem. Eur. J . 2010, 16 , 9796–9807. 58. M. K. J. ter Wiel, S. Odermatt, P. Schanen, P. Seiler, F. Diederich, Eur. J. Org. Chem. 2007, 3449–3462. 59. J. L. Alonso-G´omez, P. Schanen, P. Rivera-Fuentes, P. Seiler, F. Diederich, Chem. Eur. J . 2008, 14 , 10564–10568.

215

6 ASSIGNMENT OF THE ABSOLUTE CONFIGURATIONS OF NATURAL PRODUCTS BY MEANS OF SOLID-STATE ELECTRONIC CIRCULAR DICHROISM AND QUANTUM MECHANICAL CALCULATIONS ´ and Karsten Krohn Gennaro Pescitelli, Tibor Kurtan,

6.1. INTRODUCTION Natural products represent a rich source of therapeutically useful compounds. About 70% of the drugs marketed in the 1982–2007 period was more or less directly derived from natural products, especially in the field of anticancer drugs [1]. At the same time, the portion of chiral drugs patented as a single enantiomer has increased sharply in recent years [2], mostly as a consequence of the regulatory prescriptions concerning the stereochemical characterization of new drugs [3]. Isolation and identification of new chemical compounds from natural sources, including elucidation of their absolute stereochemistry, has therefore important consequences in many disciplines. It is hard to believe that a rather fundamental aspect of chemistry—that is, the assignment of absolute configuration (AC)—was established only in 1951 with the Xray diffraction experiment of Bijvoet et al. [4] on NaRb (+)-tartrate. The observation of anomalous dispersion has remained unsurpassed as the most reliable means for assigning absolute configurations, though with its well-known limitations. In the last decade, a valid alternative has been offered by quantum mechanical (QM) calculations of chiroptical properties such as electronic circular dichroism (ECD), vibrational CD (VCD), and Raman optical activity (ROA). Thanks to the development of computer technology, QM calculations have become accessible at a higher level of sophistication and their reliability has increased significantly. The computational approach for assigning ACs is based on the comparison of experimental and calculated ECD, VCD, or ROA spectra and has certain advantages over X-ray structure elucidation, including its practicability and wider scope. Its main limitations are the appearance of significant CD or ROA bands and (at some extent) the molecular size. As far as electronic CD (ECD) is concerned, the Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

217

218

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

compound must show a significant ECD spectrum in the commonly accessible spectral range (above 185 nm). This is, however, not uncommon with natural products, most of which contain at least a conjugated moiety leading to distinct ECD spectra. Every molecular calculation is necessarily size-dependent, and demanding computational methods may be restricted to small systems. However, ECD spectra of increasingly larger and complex molecular assemblies have been treated by high-level methods in recent years [5]. The most serious drawback of the computational approach is related to the reliability of the input structure, an issue that will be extensively discussed in the following sections because it represents the main reason for the development of the solid-state ECD/TDDFT (time-dependent density functional theory) approach described in the present chapter.

6.2. ASSIGNING ABSOLUTE CONFIGURATIONS THROUGH ECD CALCULATIONS 6.2.1. Conformation and Configuration Whenever a molecular property is predicted by theory and used for comparison with the experiment, it is crucial to employ a correct input structure. “Correct” means it must represent as faithfully as possible the true structure (or structures) responsible for the observed property. In particular, chiroptical data such as ECD, VCD, and ROA spectra are extremely sensitive to the overall molecular geometry in terms of both conformation and absolute configuration. Configurational and conformational elements are often strictly intertwined, and chiroptical approaches usually determine only one of them in the knowledge of the other [6, 7]. There are spectacular examples in the literature where two slightly different conformations of a compound (with fixed AC) led to almost mirrorimage computed ECD spectra: In other words, regarding their chiroptical parameters, two conformers behaved as two enantiomers [8]. Generally speaking, any solution CD spectrum amounts to the sum of contributions from all populated conformations, and the set of input structures to be considered in the calculation must be representative of the whole conformational ensemble [6]. As a consequence, when ECD calculations are applied to deduce absolute configurations, they must rely on an independently established conformational picture, which is gained through the use of other spectroscopic and/or theoretical means. Modern NMR techniques play a major role in deducing solution conformations, in particular when they provide data sensitive to the three-dimensional structure such as NOE effects and scalar J -couplings [9, 10]. In challenging cases represented by very flexible compounds, assistance by molecular modelling may be indispensable. A conformational analysis is normally started with a rapid computational procedure, based on Monte Carlo or molecular dynamics approaches at a low level of theory, such as molecular mechanics (MM) [11]. These conformational search routines provide a set of structures that are further optimized at a higher level of theory, usually density functional theory (DFT) or other ab initio methods. Geometry optimizations provide structures with respective (internal) energies, thereafter used to estimate the population of each calculated conformer. Calculated geometries must be checked against NMR data by considering H–H distances versus observable NOEs, as well as H/H and C/H dihedrals versus measured J -couplings [10]; when necessary, 13 C chemical shifts may be calculated and compared with the experimental set [9]. After a reliable set of input structures has been generated with an initial arbitrary AC, ECD calculations must be run on all minima within a certain energy threshold (2–3

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

kcal/mol)—that is, with significant population (say, >3–5%) at the working temperature (normally, 298 or 300 K). Thereafter, a weighted average of all computed ECD spectra is estimated according to the Boltzmann weights at the working temperature. Reliable populations can be obtained by single-point calculations at a higher calculation level than that used for geometry optimizations, for example, using larger basis sets. Furthermore, internal energies should be corrected with zero-point vibrational and entropy terms to afford true free energies. Finally, the weighted average ECD is compared with the experimental ECD spectrum: If the agreement is good with either the calculated ECD (for the initially assumed AC) or its mirror image (i.e., for the opposite AC), the configuration may be assigned. The outlined procedure, first developed by Bringmann and co-workers [12], has been used frequently in recent years to assign the AC of natural products, in particular by employing QM semiempirical [7, 13] and TDDFT calculations [14, 15]. Several approximations are often used, the most drastic of which are the neglect of solvent and vibronic effects on geometries and calculated ECD. Treatment of solute–solvent interactions, in both geometry optimization and CD calculation steps, is possible by adopting a suitable solvent model (which increases the computational time) [16]. Inclusion of vibronic effects is a rather complicated process and remains still restricted to some models or small molecules [17]. Provided that the calculation method employed for ECD calculations is efficient, the crucial point of the above procedure lies in the generation of the input structure(s). In fact, the conformational analysis may be both computationally demanding and prone to inaccuracy. The major sources of error lie in the prediction of relative energies, as well as in the possible overlooking of one or more significant conformers. Finally, since the CD calculation must be run on each calculated conformer, flexible molecules may represent very difficult cases to handle.

6.2.2. The Origin of the Solid-State ECD/Computational Approach X-ray analysis of crystalline natural products offers a twofold advantage to surmount the difficulty just described in the generation of the input structure for ECD calculations. First, in the crystals, the molecular conformation is fixed and univocal (unless polymorphs occur); second, its structure can be determined with high accuracy by diffraction experiments. In the course of a screening for novel bioactive compounds from natural sources, Krohn and co-workers investigated a Phomopsis sp. and isolated (+)-phomoxanthone A (1, Figure 6.1) [18]. The relative configuration of the chirality centers and axial chirality element was established by X-ray single-crystal analysis (Figure 6.1), while the ECD spectrum served for the assignment of the absolute configuration. Atropisomeric biaryl compounds like 1 show intense ECD spectra dominated by the exciton coupling (see Section 6.2.3) between the two aromatic chromophores, and the exciton-coupled spectra are easily correlated with the absolute stereochemistry [19]. Following the computational procedure outlined in the previous section, the authors first performed a conformational analysis with MMFF and AM1 methods [20], which afforded eight low-energy conformers within 11 kJ/mol with arbitrarily assumed (aS , 5R, 6R, 10aR, 5 R, 6 R, 10a R)-(+)-1 absolute configuration. They were used as input for ECD calculations with a semiempirical method using the BDZDO/MCDSPD program package [21]. The calculated spectra were then weighted with their respective Boltzmann populations and summed to obtain a weighted-average ECD spectrum, which reproduced the experimental solution spectrum. Surprisingly, however, when the X-ray geometry was used as input for

219

220

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

01

02

OH

O

03

OH

C7 C6

OAc OAc H3C

O 4a

4

4’

C15

C2 C1

C33

C18

013 C32

C9

4a’ O

AcO

C5 C4

C8

07

C3

CH3

C16 C10

OAc

C19

C14 06 C11 04

C17

C12

C20

C31

016

C21

C34

C38

05

O

OH

C22

C24

C30 014

C26

C36

C29

C35 C28

C13

OH

C37

012

08

015

C27

C23

C25

Phomoxanthone A (1)

011 010 09

80 20

40

0 Δε (M–1cm–1)

Calculated CD spectra Weighted average using AM1-optimized structures On X-ray geometry

60

20 –20 Experimental CD spectra Solid-state (KCl disc)

–40

Solution (4:1 CH3OH/CH2CI2) Solution (CH2CI2)

Figure 6.1.

260

300

340

–20 –40

–60 220

0

380

–60

220

260

300

λ (nm)

λ (nm)

(a)

(b)

340

380

(a) Experimental ECD spectra of (aS, 5R, 6R, 10aR, 5 R, 6 R,10a R)-(+)-phomoxanthone

(1) in solution (CH2 Cl2 and methanol/CH2 Cl2 4:1) and in the solid state (KCl disc). (b) Calculated Boltzmann-weighted average ECD spectrum over eight low-energy AM1 structures, as well as calculated ECD spectrum using the X-ray geometry (shown on the top). All calculations run with the BDZDO method.

BDZDO calculations, the resultant spectrum showed a better agreement with the experimental one. Clearly, the comparison between calculated and experimental properties is more justified when they refer to the same geometry. Therefore, the ECD spectrum was also recorded in the solid state using the KCl pellet technique (described in Section 6.3.2) and, as expected, the match between this latter spectrum and that calculated using the X-ray geometry was improved. In this way the idea was born to use X-ray coordinates as input data for ECD calculations for future characterization of chiral nonracemic natural products. It was immediately clear that the major advantage of the method would lie in skipping the conformational analysis step and its related uncertainties, with consequent time-saving and increased reliability.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

6.2.3. The Choice of the Computational Method After its first application to phomoxanthone A, the major improvement of the approach based on solid-state ECD has consisted in a different choice of the standard computational method employed, which was switched from the semiempirical BDZDO method to TDDFT. Some specific applications to large molecules and/or molecules containing multiple aromatic chromophores have been dealt with alternative computational methods such as ZINDO or even the semiclassical coupled oscillator DeVoe’s one. In this section, we will briefly discuss the choice of the computational method for solid-state ECD calculations [22]. The reader is also referred to the many chapters of Volume 1 especially devoted to the topic of simulations of ECD spectra [23]. ECD calculations with high-level quantum mechanical (QM) methods are nowadays fully practicable for moderately large organic molecules and metal complexes [24–27]. Reliable ECD predictions require theoretical methods that take electron correlation into account and use of large basis sets. Among the many possible ab initio methods employed for excited states calculations, time-dependent density functional theory (TDDFT) has emerged in recent years as one of those leading to the best accuracy/cost compromise [26, 28] (it must be noted that many DFT functionals contain adjusted parameters and are not strictly “ab initio”). Although DFT functionals are usually designed to reproduce thermochemical data [11, 29], hybrid functionals such as B3LYP, BH&HLYP, and PBE0 [30, 31] predict with high accuracy transition energies and rotational strengths. Several commonly employed DFT functionals have well-recognized drawbacks, especially a poor description of some loosely localized states, such as charge-transfer, diffuse, and Rydberg states, which may be alleviated using some of the new long-range functionals [32–34]. TDDFT arises from a perturbative approach to DFT, and therefore it is intrinsically more accurate in predicting low-lying excited states [35]. Apart from these issues, the scope of TDDFT calculations is practically unlimited. Thus, when we had to choose a general and reliable method for the calculation of solid-state ECD spectra, the choice of TDDFT was almost obvious, and now we refer to this approach as the solid-state ECD/TDDFT approach [22]. Semiempirical quantum mechanical methods rely on strong simplifications, which include ignoring core electrons and neglecting differential overlap (NDO) [36]. They are much faster than ab initio methods and may be helpful when dealing with complex molecules and simple supramolecular aggregates. Various approximation schemes have been purposely developed for spectroscopic simulations, such as CNDO/S (complete NDO) and ZINDO/S (Zerner’s intermediate NDO). They have been parameterized in order to describe aromatic and heteroaromatic chromophores and, for ZINDO, some transition metals. Both methods reproduce well ECD spectra dominated by “strong” mechanisms of optical activity [37], such as the exciton coupling between aromatic rings, like in biaryls, and the inherent chirality of twisted π -electron systems [7, 38]. The accuracy of semiempirical methods in predicting high-lying electric-dipole forbidden transitions is comparatively much poorer [39]. Another problem with semiempirical methods is related to the two gauge formulations (length and velocity) used for calculating rotational strengths. Unless using gauge-independent methods, dipole-length (DL) and dipolevelocity (DV) values for rotational strengths are always different [40]. The difference is related to basis set completeness; therefore using, for example, TDDFT with double- or triple-ζ quality basis sets including polarization functions leads to substantially equivalent DL and DV rotational strengths. By contrast, semiempirical methods often employ minimal basis sets and very large and unpredictable discrepancies may be obtained. Finally, there are cases where the ECD pattern is dominated by the so-called exciton coupling mechanism [19, 41, 42]. It arises when two or more chromophores, allied

221

222

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

with strong electric-dipole-allowed transitions, are close in the space and arranged in a skewed fashion with respect to each other. In this condition the coupling between the two transition dipoles generates a strong bisignate ECD signal, called an exciton couplet. Analysis of exciton-coupled ECD spectra is often straightforward, as witnessed by the countless applications of the exciton chirality method in the AC assignments of organic compounds, including natural products [6, 19, 41, 42]. On a quantitative ground, excitoncoupled ECD spectra may be simulated by some techniques that may be grouped under the name of semiclassical methods, such as DeVoe’s approach. For a comprehensive description of this method, we refer the reader to the original papers [43] and to recent reviews and applications [27, 44]. Provided that the chromophores under considerations are known and characterized, DeVoe-type calculations are extremely fast and applicable to systems containing even dozens of chromophores. Therefore, it would be a method of choice to estimate the ECD of large supramolecular aggregates in the coupled-oscillator approximation.

6.3. THE SOLID STATE ECD/TDDFT APPROACH: METHODOLOGY AND SCOPE 6.3.1. Principle The essence of the procedure for assigning absolute configurations by means of the solid-state CD/TDDFT method consists of measuring the solid-state ECD spectrum of a microcrystalline sample and comparing it with the spectrum calculated using the Xray geometry as input structure [22]. The main advantage of the solid-state approach, with respect to similar procedures based on solution ECD calculations, is that it does not require a conformational analysis and therefore it is computationally fast and avoids the difficulties connected to conformational searches and geometry optimizations [6b]. Moreover, the experimental and calculated ECD spectra refer to the very same geometry, implying a good agreement between theory and experiment and a reliable assignment. It is of vital importance that an artifact-free experimental solid-state ECD spectrum has to be measured (see Section 6.3.2). Furthermore, the solid-state ECD spectrum must be devoid of bands intrinsic to the solid state—for example, arising from intermolecular interactions in the crystals [45, 46]. A step-by-step formulation of the solid-state CD/TDDFT approach is the following: 1. Isolation of the natural product and determination of its constitution and relative configuration [22]. 2. Growing crystals for X-ray analysis. 3. X-ray single-crystal diffraction analysis. 4. Measurement of microcrystalline solid-state ECD spectrum as KCl pellet, as well as of solution ECD spectra in one or more solvents. 5. Generation of the input geometry for ECD calculations (with initial arbitrary AC), by optimizing hydrogen atoms of the X-ray structure. 6. Calculation of rotational strengths with the TDDFT method, possibly employing various combinations of functionals and basis sets. 7. Generation of the ECD spectrum as a sum of Gaussians. 8. Comparison of the experimental solid-state (and solution) ECD spectrum with the TDDFT-calculated one.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

The reason for recording both solid-state and solution ECD spectra is that their comparison is useful to reveal the presence of unexpected bands in the solid-state ECD spectrum. Bands much stronger than anticipated, or bands appearing in unexpected positions of the spectrum, may be due to measurement artifacts (Section 6.3.2) or be allied to solid-state intrinsic ECD effects (Section 6.5). In point 5, it is stated that the input geometry for ECD calculations is generated from the X-ray one upon optimization of the hydrogen atoms. Hydrogen atoms are often not accurately located using X-ray data because of low scattering power, distorted electron density, and large librations. In the process of structure refinement, hydrogen atoms are generated in positions determined by distance and angle constraints relative to the heavy atoms they are attached to. The SHELXL routine for structure refinement [47] adopts a “riding” model based on default C–H distances that, for some reason, are much shorter than real ones and would introduce unwanted errors in the calculated ECD spectra. Thus a preliminary DFT optimization of hydrogen atoms is run with B3LYP/6-31G(d), leaving all other atoms frozen, which usually produces geometries whose Y–H bonds (Y = C,O,N) are longer by 10–15% than the input ones. Although librations of carbon, oxygen, and nitrogen atoms are usually much smaller than for hydrogen, their effect may also lead to a slight underestimation of, for example, C–C, C–O, and C–N bond lengths by X-ray diffraction measurements [48]. In our experience, the discrepancies between Xray geometries and DFT-optimized ones (using the X-ray geometries as input structures and keeping all dihedral angles frozen) are very small, because C–C, C–O, and C–N bond lengths may vary within ±2%. Possible exceptions are compounds with a compact polycyclic structure, for example, containing fused or spiro-linked five-membered rings, where many bond lengths are strictly correlated to each other (see compounds 4 and 15 in Chart 6.1) [49, 50]. In these cases, we observed that relieving the bond lengths by means of DFT geometry optimizations was beneficial for a better agreement between experimental and computed solid-state ECD spectra, but never decisive for assigning the absolute configuration. Concerning point 6, it is clear that the choice of a proper DFT functional and basis set is crucial for the success of the method. One of the advantages of the solid-state approach is that the time saved in the generation of the input structure may be devoted to testing different combinations of functionals and basis sets to look for the best agreement with the experimental spectrum. According to our experience on several compounds with diverse structures, the three hybrid functionals B3LYP [29], PBE0 [31], and BH&HLYP [30], with increasing fraction of exact (Hartree–Fock) exchange, can be used to handle most common situations and often result in calculated ECD spectra in very good agreement with experimental ones. Usually, the spectra calculated with B3LYP and PBE0 are quite similar to each other, and they differ from BH&HLYP for a systematic wavelength shift (the direction depends on the nature of transitions involved). To improve the predictions of energies and transitions dipoles allied with charge-transfer, diffuse, and Rydberg states, use of the Coulomb-attenuated version of B3LYP, known as CAMB3LYP, is advisable [34]. In our hands, however, BH&HLYP has demonstrated to be capable of solving the same kinds of problems efficiently. It is remarkable that this “halfand-half” functional is included in the most recent versions of computational packages “for backward-compatibility only,” but its usefulness is indisputable [51]. As for the basis sets, use of the largest possible basis set is recommended for many reasons, some of which have been outlined above [26, 40]. On a practical ground, a split-valence basis set including a reasonable set of polarization functions and a minimum set of diffuse functions would be sufficient in most cases. Our favorite in the field is Ahlrich’s triple-ζ

223

224

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

basis set TZVP [52], whose set of flexible polarization functions somehow counterbalances the lack of diffuse functions. When inclusion of these latter seemed necessary, we used the well-known Dunning’s correlation-consistent aug-cc-pVDZ set [53], or the less popular ADZP [54], or else aug-TZVP, obtained by augmenting TZVP with the set of most diffuse functions taken from aug-cc-pVDZ. In almost all cases analyzed thus far, however, use of, for example, aug-TZVP versus TVZP improved the general agreement between experimental and calculated spectra, without being decisive for the configurational assignment. In point 7, the set of calculated rotational strengths as a function of frequency (in other words, a stick plot) is converted in a more handy ECD spectrum by applying a broadening or band-shape. To each rotational strength is associated a band-shape function, usually of Gaussian or Lorentzian type, with intensity proportional to the absolute rotational strength value. A sum of all bands is then evaluated to generate a full ECD spectrum. Such a procedure requires the selection of a band-width parameter σ that is normally established on an empirical ground, selecting the value providing the best fit with the experimental spectrum in the more relevant spectral region(s). Expressing the rotational strengths in 10−40 cgs units and ε in the common M−1 · cm−1 units (where M is molarity, mol · L−1 ), the ECD spectrum calculated as sums of Gaussians in the wavenumbers (˜ν in cm−1 ) domain is [55]      Ri v˜i v˜ − v˜i 2 exp − . (6.1) ε(˜ν ) = 0.0247 σi σi i

If computed transition wavelengths are systematically shifted with respect to experimental ones, a wavelength correction may be applied to better compare computed and experimental spectra. In the so-called UV correction, one looks for the match between computed and experimental UV–vis spectra and then applies the same shift to ECD ones [7]. The comparison between the experimental solid-state ECD spectrum and the TDDFT-calculated one (point 8) is clearly the decisive step. The advantage of using full predictions of ECD spectra with respect to other kinds of treatments such as the exciton chirality approach is that the former provide a full ECD spectrum extending on a more or less wide wavelength range, rather than focusing on a single spectral feature. When the experimental ECD spectrum is well-structured—that is, when it contains several distinct bands—this offers the chance for a critical evaluation of the calculation results going beyond a mere comparison of band signs. This “quality check” may be very important to assess the reliability of the calculation method employed and may point, if necessary, to the need of testing further ones.

6.3.2. Solid-State CD Measurements Solid-state ECD spectra can be measured by different techniques described in several reviews [22, 45, 46, 56, 57]. The microcrystalline pellet or disc method is the most frequently used technique and was adopted as the standard in our solid-state ECD/TDDFT protocol. The crystalline sample is mixed and powdered with a suitable matrix such as KBr, KCl, or CsI, and the microcrystalline powder is pressed to produce a translucent glassy disc. KBr can be used for ECD measurements only above 220 nm, because of its UV cutoff. The use of KCl is less common in the literature, despite the fact that KCl can be used down to 180 nm and its handling is not different from KBr. The shortwavelength range extension may be crucial for the solid-state ECD analysis of natural

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

products containing only weak chromophores with high-lying transitions, such as alkene, ester, or anhydride groups (see compounds 7, 13, and 14 in Chart 6.1). Detailed procedures of pellet preparation are reported in several articles [58, 59]. In our solid-state ECD measurements of natural products, the disc is prepared by grinding and mixing ≈ 180–250 mg of KCl (≥99.999% Fluka, preheated at 100◦ C) and 30–250 μg of sample (depending on the chromophore) with the aid of a Perkin–Elmer vibrating mill equipped with a stainless steel ball for 5 min. The mixture is then pressed under vacuum at 10 tons with a Perkin–Elmer press for 5–10 min to provide a translucent disc. To decrease diffused reflections at grain boundaries, the sample and the matrix must be finely powdered and mixed to provide homogeneous distribution. Elaborated grinding is necessary because the intensity of the scattered light is proportional to 6th power of the particle diameter. The lowest possible sample amount providing acceptable ECD spectra is generally used to decrease the effect of absorption flattening [60] and assure linearity with sample concentration [58]. Because KCl is hygroscopic, measurements are carried out right after the preparation of the disk by mounting it on a rotatable holder placed as close as possible to the detector. Normally, solid-state ECD data are reported as ellipticity φ in mdeg units. However, they can be normalized to ε when necessary, provided that the approximate dimensions of the pellet are known. It is well established that solid-state ECD measurements are easily contaminated by artifacts due to linear dichroism (LD) and linear birefringence (LB) allied with macroscopic sample anisotropies in the sample [56, 61, 62]. These artifacts have both rotation-dependent and rotation-independent contributions, whose presence may be ascertained by, respectively, sample rotation around the incident-light axis (or z axis) and flip (180◦ rotation) around the vertical y axis [61, 62]. However, averaging the various spectra obtained by z -rotation and y-flip is not sufficient for obtaining artifact-free ECD spectra. As clearly pointed out by Kuroda and Shindo, the only proper way to extract true ECD data from a raw spectrum with LB and LD contributions consists of recording ECD, LB, and LD signals simultaneously and applying a specific protocol based on Mueller matrix formalism [56, 61]. Recording ECD, LB, and LD spectra is possible with a dedicated chiroptical spectrophotometer (Jasco J-800KCM) [61]. Using commercially available ECD spectrophotometers, at least LD data can be recorded, and simultaneous ECD/LD measurement should always be performed whenever possible on solid samples. In our protocol, various ECD spectra are usually recorded upon stepwise 90◦ rotations around the z axis as well as 180◦ flip around the y axis, to exclude the presence of both rotation-dependent and -independent contributions from macroscopic anisotropies. In most cases, the spectra showed negligible changes with rotation or flip [22]. When some influence of rotation around z axis was noticed, it consisted of a periodic baseline shift leading to minor differences between spectra (Figure 6.2). Solution spectra in various solvents are always recorded to be compared with solid-state ones to further exclude the presence of spectral artifacts in solid-state spectra (Figure 6.2).

6.3.3. Applicability The most important prerequisite of the solid-state ECD/TDDFT approach is the availability of X-ray geometries from single-crystal diffraction. One may question why to measure and calculate an ECD spectrum once the X-ray geometry is known. This is because the configurational assignment by X-ray anomalous scattering effect is subjected to important limitations that do not apply to the solid-state ECD method. When crystals belonging to noncentrosymmetric space groups exhibit anomalous scattering, the differences observed in Friedel pairs intensities may be used to extract

225

226

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60

60

40

20

0

–20

–40

0° 90° 180° 270°

40

φ (mdeg)

Δε (M–1cm–1)

Solid-state ECD spectra (KCI disc), disc rotation

Solution ECD spectrum (CH3CN)

20

0

–20

200

225

250 275 λ (nm)

300

325

350

–40

200

225

250

275

300

325

350

λ (nm)

Figure 6.2. Comparison between solution (left) and solid-state ECD spectra (right) measured on KCl pellets upon rotations around the incident light z axis. The sample is bis(4-bromobenzoate) of palmarumycin M1 (29b) discussed in Section 6.6.

phase information and assign absolute configurations. Such a difference is, however, rather small, and it is measurable only in the presence of one or more atoms whose absorption edge is close to the X-ray wavelength. The scattering factor responsible for the magnitude of the anomalous dispersion effect is in fact both atom- and wavelength-dependent, and it is almost negligible for light elements (up to O). Although a single F or even O atom may suffice for measuring the anomalous dispersion effect, excellent data collection should be performed with well-defined single crystals at low temperatures, recording at least a full set of Friedel pairs and using CuKα radiation. A statistical survey of Cambridge Structural Database (CSD) [63] clearly reveals the scope of the anomalous dispersion method to assign absolute configurations. The version of the CSD updated to March 2010 contains ≈ 8000 compounds in chiral space groups flagged with “absolute configuration.” However, only 586 consist of H, C, O, and N atoms, and 51 consist of H, C, O, N, F, amounting overall to less than 8% on the total. Even in the presence of a heavy scatterer, assignment of the absolute structure by X-ray analysis is not trivial. Two literature surveys that were reported around 25 years apart [64] revealed “many unsatisfactory features in the original publications,” including incorrect space groups and insufficient Friedel coverage leading to great uncertainty in some of reported configurations. It must also be stressed that a large majority of natural products contains only H, C, O, and N atoms. Of the around 226,500 entries comprised in the Dictionary of Natural Products as of June 2010 [65], only 15,500 (less than 7%) contain heteroatoms other than O and N. In summary, there is much space available for application of the solid-state ECD method for assigning absolute configurations of natural compounds. A major issue associated with solid-state ECD spectra is that intermolecular interactions between molecules closely packed in the crystals may give rise to non-negligible contributions to the spectrum [45, 46]. Any phenomenon of this kind cannot be predicted by a calculation run on a single molecule. The impact of crystalline intermolecular effects on the solid-state ECD/TDDFT method will be discussed in Section 6.5.

227

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

6.4. THE SOLID-STATE ECD/TDDFT APPROACH: EXAMPLES OF APPLICATIONS The solid-state ECD/TDDFT approach has been applied thus far to 17 natural products (Chart 6.1), secondary metabolites of fungi, often endowed with biological activities such as antibiotic, antifungal, and anti-inflammatory [15, 49, 50, 66–76]. These compounds exhibit a great structural variability and contain different chromophores, fully or partially unsaturated rings, flexible saturated rings or chains, and centers of chirality in diverse O

O

O HO

OH

OH COOCH3

H3C

HO

O O

Globosuxanthone A (2)

O OHO

OCH3 OH

O

H O

O

O O Microsphaeropsone A (8)

O

3,16-Diketoaphidicolan (6)

O H Sinularolide B (5)

OH

O

H

O

O HO

HO

O

O HOH2C

O

HO

O Massarilactone E (4)

OH O Ascochin (3)

O

OH O

Viburspiran (7)

O

OH

H

O

O

O

OH

O

HO

H

HO O

O

MeO

O

OH

O Curvulone A (9)

O

O

O

α,β-Dehydrocurvularin (10) O

OCH3

OH

O

OH

OH

Fusidilactone B (14)

O Tetrahydropyrenopherol (13)

H3C H3CO O RO

OH

O O

N Macropodumine B (12)

H

O

HO

O

OH

O

H

O

OH

Blennolide A (11)

H

H3C

O

OH

O

O

CH3

O

H O O

Papyracillic acid A (R = H, 15a) and methyl acetal (R = CH3, 15b)

H3CO

O

Hypothemycin (16)

O OHOH

O

O 1β,10β-Epoxydesacetoxymatricarin (17)

Chart 6.1. Natural products whose AC has been assigned by means of the solid-state ECD/TDDFT approach.

228

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

spatial relationships. In all cases, the absolute configuration could be established thanks to a good agreement between the experimental and calculated solid-state ECD spectra. In the following, some relevant cases will be discussed in detail to exemplify the adopted procedure and stress its scopes and limitations [22]. After phomoxanthone (1) described above, the first case investigated with the solidstate ECD/TDDFT approach was globosuxanthone A (2, Chart 6.1) extracted from Microdiplodia sp. and obtained in a crystalline form suitable for X-ray analysis [68]. Globosuxanthone A is a dihydroxanthenone with potent antitumor activity, also isolated from Chaetomium globosum [77]. ECD spectra of 2 were recorded in solution (methanol and a tertiary solvent mixture) and in the solid state as a KCl pellet (Figure 6.3). The three spectra were roughly consistent, but they displayed some differences probably related to the conformational freedom around the ester and diol moieties. Most importantly, however, the solid-state ECD spectrum did not show any extra band with respect to the solution ones and both the signs and relative intensities of the major bands were preserved. This observation pointed to the essentially intramolecular origin of the solidstate ECD spectrum, meaning that intermolecular couplings in the crystal lattice did not give rise to apparent contributions. Starting from the X-ray geometry of 2 and assuming (1R,2R) configuration, hydrogen atoms were re-optimized with DFT using B3LYP/631G(d). As a result, C–H and O–H bond lengths increased by 15% on the average, demonstrating the necessity of this preliminary optimization. TDDFT and ZINDO calculations were tested in ECD calculations, employing in the first case various hybrid functionals (B3LYP, PBE0, and BH&HLYP) and TZVP basis set. Both B3LYP/TZVP and PBE0/TZVP calculations led to a good agreement with the experimental spectrum above 250 nm in terms of position, sign, and (to some extent) relative intensity of bands, including a couple of shoulders in the 300- to 330-nm range (Figure 6.3). The BH&HLYP functional reproduced the sign sequence but with a less satisfying general agreement. On the contrary, ZINDO performed very poorly, which was not unexpected in view of the

Δε (M–1cm–1), φ (mdeg), and R (10–39 cgs)

8 6

Experimental CD in CH3OH solution

4

Experimental solid-state CD (KCl disc)

O OH COOCH3

OH O HO Globosuxanthone A (2)

2 0

C3

O2 C6

–2

C2 C1

C5

C7

Calculated CD on X-ray geometry Calculated R

–4

C4

C12 C10 C11

C8

–6

O5

C13 C14

C9

O1 07 C15

04

–8

O3

200

250

300 λ (nm)

350

O6

400

Figure 6.3. Experimental ECD spectra of (1R,2R)-(−)-globosuxanthone A (2) in methanol solution and in the solid state (KCl disc), as well as ECD spectrum calculated with B3LYP/TZVP on the solid state geometry (shown on the right). Vertical bars represent calculated rotational strengths.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

molecular and electronic structure. In fact, the most plausible mechanism of optical activity responsible for the moderately intense ECD spectrum of 2 is the second-sphere chirality [78] provided by the chiral ring on the conjugated chromophore (which includes the planar diene). Based on the results obtained with TDDFT calculations, the absolute configuration of globosuxanthone A was assigned as (1R, 2R)-(−)-2 [68]. The conclusion was reached that TDDFT, with a proper choice of the functional and a sufficiently flexible basis set, would grant in subsequent cases a wide applicability with a reasonable computational time. In particular, since the conformational analysis step is skipped and the ECD calculation needs to be run on a single structure, various functionals (and possibly basis sets) may be tested to look for the best agreement with the experiment in terms of overall spectral appearance, which can be taken as an indication of the reliability of the configurational assignment. Essentially the same procedure was followed for successive compounds 3–17 (Chart 6.1) isolated from natural sources, the absolute configurations of which were assigned by means of the solid-state ECD/TDDFT approach. These included (4S )-(+)ascochin (3), isolated from the fungus Ascochyta sp. from the plant Meliotus dentatus, exhibiting antifungal and algicidal activity [69]; (4S , 5R, 6S , 7R, 10R)(+)-massarilactone E (4), isolated from Coniothyrium sp. associated with Artimisia maritima [49]; (1R, 2R, 3R, 12S , 13R)-(−)-sinularolide B (5), a cytotoxic compound isolated from the soft coral Lobophytum crassum [71]; (4R, 5R, 8S , 9S , 10S , 12R)-(−)-3,16-diketoaphidicolan (6), extracted from the endophytic fungus Phoma sp. isolated from Aizoon canariense [76]; and (1R, 2S , 4R, 5R, 6R, 7R, 8S )-(+)-viburspiran (7), a new octadride member of maleic anhydride natural products, extracted from the endophytic fungus Cryptsporiopsis sp., isolated from Viburnum tinus [74]. A somewhat peculiar case is represented by (1R, 2R)-microsphaeropsone A (8), the first natural compound with dihydrooxepino[2,3-b]chromen-6-one skeleton, isolated from Microsphaeropsis sp. from the bush Lycium intricatum. Calculated ECD spectra using the solid-state geometry for 8 showed a great variation with the functional employed, a situation occurring sometimes with complicated chromophores which calls for special caution in the interpretation of results. Although at least BH&HLYP/TZVP calculations reproduced the experimental solid-state ECD spectrum satisfactorily, the assignment was substantiated by means of vibrational CD [79], following the recommendation to use diverse chiroptical techniques to assign the absolute stereochemistry in ambiguous situations [80]. The VCD spectrum of 8 was measured in solution (CDCl3 ) and compared with the Boltzmann-averaged calculated spectrum using B3LYP/6-31G(d) on three DFT-optimized low-energy structures, including SCRF-PCM solvent model for chloroform [73]. More recently, a series of curvularin derivatives with 12-membered lactone ring skeleton belonging to the polyketide family was extracted from the fungus Curvularia sp. 6540, isolated from the marine red algae Gracilaria folifera. All compounds showed antibacterial, antifungal, and antialgal activity, and they included two new compounds, (10S ,15R)-(−)-curvulone A (9, Chart 6.1) and (11R,15R)-(−)-curvulone B, and two known curvularins with unusual (15R) configuration, namely (11R,15R)11-hydroxycurvularin and (10E ,15R)-(+)-10,11-dehydrocurvularin (10) [75]. For compounds 9 and 10, crystals suitable for X-ray analysis were obtained and thus the solid-state ECD/TDDFT approach was applied. In both cases, the absolute configuration assignment was confirmed by independent methods (chemical correlation and X-ray analysis) [75, 81]. Compound 10 was especially interesting because it crystallized in two different forms, depending on the solvent employed (CDCl3 versus wet CH3 OH/CH2 Cl2 ). The

229

230

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

two X-ray geometries showed large discrepancies in the conformation of the flexible lactone ring, and, interestingly enough, the ECD spectra calculated with TDDFT (B3LYP/TZVP) on the two X-ray structures were also very different (Figure 6.4). The solid-state ECD spectrum measured on the crystals obtained from CDCl3 (type A) was well-reproduced by the TDDFT-calculated spectrum using the respective geometry. The above results suggested that the solid-state ECD/TDDFT approach may be well employed for crystalline materials exhibiting polymorphism, provided that the crystal form used for the solid-state ECD measurement is known and the corresponding molecular structure is determined [75]. Another large collection of natural compounds was obtained from Blennoria sp., an endophytic fungus from the succulent Carpobrotus edulis, which led to eight compounds related to the known secalonic acids [72, 82]. Besides secalonic acid B (18, Chart 6.2), a powerful fungicide and algaecide, a series of new monomeric and a mixed dimer derivatives were found, named blennolides (Chart 6.2). These compounds represent the long-awaited monomeric units of the dimeric secalonic acids, in particular the isomeric hemisecalonic acids B and E (or blennolides A and B, 11 and 19). Their rearrangement products 20–22 (blennolides D–F) are structurally unique new natural products, where a highly substituted γ -lactone moiety is linked to a dihydrobenzopyranone. In blennolide G (23), the usual ergochrome monomer is linked to the deoxy analogue of rearranged monomer 20, extending the secalonic acid family with a novel heterodimer. The structure and the relative stereochemistry of blennolide A (11) were confirmed by single-crystal X-ray analysis, while the absolute configuration was determined as

OH

Δε (M–1cm–1) and φ (mdeg)

25 Experimental solid-state CD (KCl disc, type A crystals)

20 10

HO

Calculated CD on X-ray geometry (type A)

H CH3 α,β-Dehydrocurvularin (10) O

Calculated CD on X-ray geometry (type B)

5

O

O

A

0 B –5 200

250

300 350 λ (nm)

400

450

Figure 6.4. Left: Experimental solid-state ECD spectrum of (15R)-(+)-10,11-dehydrocurvularin (10) recorded using crystals isolated from CDCl3 (type A), as well as ECD spectra calculated with B3LYP/TZVP on the X-ray structures of (15R)-10 for the crystals isolated from CDCl3 (type A) and a mixture of wet CH3 OH and CH2 Cl2 (type B). Right: Overlapped solid-state X-ray geometries of (15R)-10 for type A and B crystals (dark and light structures, respectively).

231

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

O

HO

OMe O

OH

OH

O

O

OH

OH

OH

OH

O H

2

OH

O

O

MeO

O

OH

MeO

Secalonic acid B (18)

OH

O

OH

O

O

O O H

O

MeO

Blennolide B (19)

OH

O

O

O

O

Blennolide D (20) x OMe O OH

H O O

Blennolide E (21)

O

OH

OH

O

9

O

H

OH

2

MeO

OH 9

O

O

MeO

O O

Blennolide F (22)

O

O

OH O

Blennolide G (23)

MeO

O

OH

Chart 6.2. Blennolides extracted from Blennoria sp. [72, 82]. Blennolide A (11) is shown in Chart 6.1.

(5S ,6S ,10aR)-(+)-11 after application of the solid-state CD/TDDFT approach. The situation for blennolides D and E (20 and 21) was more challenging because their Xray structures were not available. The overall stereochemistry of 20 and 21 could be established only by means of a combination of several spectroscopic (NOESY, heteronuclear 3 JC,H couplings, ECD) and computational techniques (MM conformational searches, DFT geometry optimizations, TDDFT calculations). In particular, molecular modeling results were essential to rationalize observed NOEs around the rotatable C2–C9 bond [72]. The family of blennolides offered a striking evidence of the advantage provided by the X-ray crystal structure for establishing both relative and absolute stereochemistry. A further confirmation of this latter point is provided by the two analogue macropodumines B and C (12 and 24, Figure 6.5 and Figure 6.6, respectively), belonging to the family of Daphniphyllum alkaloids, a group of fused-heterocyclic fungal metabolites with significant medicinal properties. Macropodumine B (12) shows an almost unique structural feature for a natural compound: It is in fact a zwitterion containing a rare cyclopentadienyl anion and an iminio counterion (Figure 6.5). Compounds 12 and 24 were isolated from the Chinese medicinal plant D. macropodum, and the solid-state structure of 12 was determined by X-ray analysis [83]. The absolute configuration of macropodumines B and C was established by comparison of their ECD spectra with TDDFT-calculated ones, using the solid-state protocol for macropodumine B (12) and the corresponding solution protocol for macropodumine C (24). In practice, the two compounds offered the option for a direct comparison between the two methods. For macropodumine B (12), the solid-state ECD/TDDFT approach led to a calculated CD spectrum (B3LYP/TZVP) in very good agreement with the experimental solid-state one (KCl disk, Figure 6.5). Other functional/basis set combinations (using BH&HLYP and ADZP) did not improve the observed agreement. After a moderate computational effort (>>

5′ O

O

O

O

Favored conformation I

Unfavored conformation II

Predicted positive CD exciton couplet

Predicted negative CD exciton couplet

Figure 7.7. Macrocyclic 1:1 host–guest complex formed between guest and host. (Reproduced by permission of The Royal Society of Chemistry [16, 41].)

7.3.2.3. Practical Application of the Pfeiffer Effect for Analyzing Chiral Diamines. Recently, Anslyn took advantage of Pfeiffer-related phenomena to develop a rapid assay of enantiomeric excess. Chiral diamines were added to racemic Cu(I) or Pd(II) complexes, resulting in CD spectra corresponding to metal-to-ligand charge transfer (MLCT) bands of the metal complexes [45, 46]. An instrument interfaced to a robotic 96-well plate allowed rapid and convenient measurement of the CD spectra of the compound library. Linear discriminate analysis of the CD spectra then determined the identification, concentration, and enantiomeric excess of the diamines. This study represents a practical application of a chiroptical sensor technique. 7.3.2.4. Anion-Controlled Switching of Amide Complexes. There are several interesting examples of chiroptical metalloswitches triggered by interaction with

261

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

OCH3 O

1.3 nm

0.3 nm

+2 H• (CF3SO3H)

2X

NO3–

2X

N * H OCH3

X X

•

–2 H N

NO3–

N

X: H2O, CH3CN or CF3SO3– “Extended-Λ-form”

H N

N*

H2L5

“Extended-Δ-form”

N* H

N

N*

OCH3

O OCH3

O H3C

“Contracted-Λ-form”

N

H N

CH3

O

H2L6

535 nm Contracted-Λ

Δε (dm3mol–1cm–2)

1

Extended-Λ

0.5

0

–0.5 Extended–Δ 400

600

800

1000

λ (nm)

Figure 7.8. Stretching and inverting dual motions of the CoII complex. Crystal structures CD spectra of [Co(L5)](left, • for CD) and [Co (H2 L5) (CF3 SO3 ) (H2 O)] (CF3 SO3 )–(CHCl3 ) (middle, ), as well as DFT-optimized structure of [Co(H2 L5)(NO3 )]+ (right, ), are illustrated [50]. (Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.)

◦

anions. Interest in such systems is heightened because supramolecular recognition of anions developed later than that of cations, and the abundance of potential anionic analytes in biology and other areas. Yano et al. [47] reported chiral inversion induced around a seven-coordinated cobalt center by interaction between sugars and sulfate anions. New cage-type cobalt(II) complexes that consist of N -glycosides from mannose-type aldoses and tris(2-aminoethyl)amine (tren), [Co((aldose)3 tren)]X2 • nH2 O (X = Cl− , Br− ), and [Co((aldose)3 tren)]SO4 • nH2 O exhibited C3 helical configuration inversion around the Co(II) center. The CD spectral characteristics of [Co((aldose)3 tren)]X2 • nH2 O changed dramatically with the addition of sulfate anions, and even inverted at high sulfate concentrations, suggesting ion pair formation which was confirmed by a crystal structure. When sulfate ion is embedded into the cavity of the sugar hydroxyl groups, the complex adopts a  configuration, while the complex with the halogen anion exhibits a  configuration. When the sulfate anion approaches the sugar complex, the electrostatic attraction between the doubly negative and positive charges of the sulfate anion and complex cation causes the hydrogen bonds between the ligands to be interrupted and brings about a chiral inversion due to the sulfate embedding into the large complex cavity. Reversibility was exhibited when the sulfate ions were removed and replaced with halide ions. Miyake et al. [48] established that the helicity of a chiral tetradentate ligand (ligand L6 in Figure 7.8) chelated to Co(II) was readily inverted by the addition of nitrate anion. Preliminary studies suggest that two molecules of nitrate serve to invert the helicity

262

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the ligand. Chelation of the first equivalent to the Co(II) center displaces the two tertiary nitrogens of the bound ligand, while the second equivalent of nitrate disrupts hydrogen bonding of the amide to solvent [49]. Circular dichroism studies indicated that the initial Co(II) complex exhibited a positive CD signal in the range of the d –d transition (around 530 nm). Upon gradual addition of Bu4 NNO3 the d –d transition the CD response changes from positive to negative. These early findings by Miyake led to the synthesis of a chemical device designed to exhibit dual mode motions [50]. This time, a modified version of chiral tetradentate ligand (ligand L5 in Figure 7.8) including 2,5-dimethoxybenzene moieties attached through amide linkages to the terminals of the ligand was employed. An acid–base reaction of the corresponding cobalt complex triggers an interconversion of coordinating atoms between amide nitrogen atoms and amide oxygen atoms, which causes a stretching (extension/contraction) molecular mode. Inversion of helicity (again from  to ) after addition of five equivalents of Bu4 NNO3 accounts for the second device mode. CD studies of the [Co(L5)] complex in CH3 CN/CHCl3 = 1/9 exhibit positive signals at 433 nm and 918 nm and negative signals at 474 nm, 607 nm, and 1100 nm, which correspond to the contracted -form. Two equivalents of CF3 SO3 H (to form [Co(H2 L5)(CF3 SO3 )(H2 O)](CF3 SO3 ) • (CHCl3 ) then caused a rapid signal shift from 0 to positive in the 530-nm region (∼ 5 s) which indicated a switch to the extended -form. The helicity inversion caused by addition of five equivalents of Bu4 NNO3 gave rise to a negative CD signal around 530 nm. The similarity of this CD signal shift to the original H2 L6 ligand study supports the assertion that helicity is changed in the device. CD signals remained consistent after many deprotonation/protonation cycles, proving that robust reversibility was established. Such a kinetically labile Co(II) complex provides for a dynamic dual mode switch that could potentially be required for sophisticated supramolecular switching devices. Recently, pentapeptide chains were combined in such a chirality-switchable Co(II) complex. The peptide chains experienced helix inversion following the reconfiguration of the octahedral metal center upon addition of the NO3 − anion stimulus (Figure 7.9) [51]. Similar peptide helix inversion by nitrate anion was shown to occur in analogous NiII and ZnII complexes. Selection between ZnII , CoII , or NiII allowed tuning of the rate of the inversion process to occur on a timescale from milliseconds (ZnII ) to hours (NiII ). The estimated half-lifetime (log τ1/2 ) of these metallo-peptide complexes showed a linear correlation with the water exchange lifetime of the aqueous metal cations.

P O O NO3– O O N M M O O Λ Helicity Δ Inversion

4

M Rapid

O N OM O Chirality Δ Transfer

log(t1/2)

P

Ni(II)

2 Co(II) 0 Zn(II)

–2 Λ, P-form

Δ, P-form

Δ, M-form

–8

–7

–6 log(τ)

–5

–4

Figure 7.9. (Left) Helicity inversion around a metal center and sequential chirality transfer to the pentapeptide helical tubes (-Aib-Phe-Aib-Phe-Aib-OCH3 ). (Right) Linear relationship between half-lifetime (t1/2 ) of helicity inversion and cation water exchange lifetime (τ /s) [51]. (Reproduced by permission of the American Chemical Society.)

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

Δε a b1

N S

R R

N R H

c1, d1 He1 g1, h1 R N f1 i1, j1 R i2, j2 R S N f2 g2, h2 N H e2 c2,d2

λ acetonitrile

N

H

Water Δε

b2 a

L7 (RRRR) (RSRS)

λ

Figure 7.10. (Left) Structures of one of the chiral macrocycle isomer L7(RRRR)(RSRS) . (Right) Solvent induced reversible helicity inversion with accompanying CD spectra [54]. The top spectra show the change in ellipticity over time for the compound on the left upon dissolution in acetonitrile; the bottom spectra are for the compound on the right in water. (Copyright American Chemical Society. Reproduced with permission.)

7.3.2.5. Solvent-Controlled Switching of Metal Complexes. Solvent can also induce significant changes in chiroptical response in metal-based systems, either as a result of general differences in polarity or nonspecific solvation or as a result of coordination of solvent in the inner coordination sphere of a metal complex. In a novel study, the Lisowski group has examined chiroptical switches involving lanthanide complexes of chiral hexaazamacrocycles [52, 53]. The hexaazamacrocycle L7 [54] shown in Figure 7.10 was designed to complex large lanthanides such as Yb(III) and Eu(III). Upon addition of Yb(NO3 )3 • 5H2 O in acetonitrile, it was observed that the chiral ligand L7 wrapped around the Yb in a helical -form corresponding to the (R, R, R, R)-(S, R, S, R) L7 isomer. Crystal structure studies of [YbL7(NO3 )2 ]2 -[Yb(NO3 )5 ](NO3 )4 • 5CH3 CN show an improper torsion angle C2–C4–C15–C17 of −13.3◦ , which is unusually high for a lanthanide(III) hexaazamacrocycle complex. Solvation of the same complex in water, though, leads to ligand reorganization presenting a sharp shift in helicity as evidenced by an improper torsion angle C2–C4–C15–C17 of 87.2◦ for the (R, R, R, R)-(S, S, S, S ) isomer. CD studies confirm helicity inversion by solvent effects, demonstrating quantitative conversion in 144 hours. The proposed mechanism of inversion involves an initial exchange of hydrate into a 10-coordinate metal inner sphere, which is followed by slow ligand reorganization into an 8-coordinate sphere. Lisowski argues that the “squeezed” (R, R, R, R)–(S, S, S, S ) isomer is more capable of accommodating smaller water axial ligands whereas the “open” (R, R, R, R)–(S, R, S, R) isomer preferentially binds the bulkier nitrate ligand in the axial position [54]. The study as a whole represents a rare case of reversible solvent induced helicity inversion for a metal-based complex. Recently, Muller, Lisowski, and co-workers [55] reported that a similar but larger chiral nonaazamacrocyclic amine wraps around the lanthanide(III) ions to form enantiopure helical complexes. The NMR and CD spectra show that kinetic complexation product of the (R, R, R, R, R, R) isomer prefers the (M )-helicity. However, the preferred helicity of the thermodynamic product is M for the early lanthanide(III) ions and P for the late lanthanide(III) ions. In solution, the late lanthanide(III) complexes slowly invert their helicity from the kinetically preferred M to the thermodynamically preferred P . The Mamula group reported the diastereoselective self-assembly of the enantiomerically pure pinene-bipyridine-based receptor, (−) or (+) L-, in the presence of Ln(III) ions

263

264

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Ln3+

Ln3+ N CH3OH

N

COO– (+)-L

CH3CN

CH3CN CH3CN + H2O

tris-Ln[(+)-L]2

tetra-Ln4[(+)-L]9

Figure 7.11. Divergent self assembly leading to the synthesis of interconverting trinuclear [Ln3 {(+)-L}6 (μ3-OH)]2+ and tetranuclear [Ln4 {(+)-L}9 (μ3-OH)]2+ complexes [56]. (Reproduced by permission of the American Chemical Society.)

(Figure 7.11) [56]. Upon exposure to La, Pr, Nd, Sm, Eu, Gd, and Tb ions in dry acetonitrile, it forms a C3 -symmetrical, pyramidal tetranuclear species with the general formula [Ln4 (L)9 (μ3 -OH)](ClO4 )2 ) (abbreviated as tetra-Ln4 L9 ). Three metal centers shape the base: an equilateral triangle surrounded by two sets of helically wrapping ligands with opposite configurations. The tetranuclear structure is completed by a capping, helical unit LnL3 whose chirality is also predetermined by the chirality of the ligand. The sign and the intensity of the CD bands in the region of the π –π * transitions of the bipyridine are highly influenced by the helicity of the capping fragment LnL3 . In methanol, it selfassembles to give the trinuclear species [Ln3 (L)6 (μ3 -OH)(H2 O)3 ](ClO4 )2 ) (abbreviated as tris-LnL2 ). The two related superstructures can be interconverted. As is shown by the CD evolution in Figure 7.12, in pure dry acetonitrile, pure tris-LnL2 disassembles and reorganizes gradually to form tetra-Ln4 L9 . If a certain amount of water is added, tris-LnL2 can be reassembled quantitatively. Water stabilizes the trinuclear species to the detriment of the tetranuclear ones. Reducing the amount of water by molecular sieves leads to the tetranuclear species. However, the number of these reversible cycles is limited due to partial decarboxylation of the ligand in the presence of water. Recent reports from Nitschke describe a Cu(I) based solvent-triggered molecular switch [57]. The initial synthesis of the chiral Cu(I) complex in methanol resulted in an equal mixture of both P - and M -diastereomers which was characterized by a weak circular dichroism spectrum bearing similarities to that of the free ligand. A featureless CD spectrum in the region of the MLCT band further established that there was no net diastereomeric excess formed. A stark contrast was then encountered when the Cu(I) complex was dissolved in dichloromethane-d2 and the CD spectrum revealed a positive CE in the MLCT region. Combined studies of CD and NMR suggest that the Cu(I) complex (Figure 7.13) fully converts to the P -diastereomer in nonpolar dichloromethane. Similar studies in DMSO then showed that the M -diastereomer of the complex (Figure 7.13) exists in 20% excess, setting the stage for a reversible metal-based chiroptical molecular switch. The solvent-induced conformational exchange was reasoned to be dominated by hydrogen bonding effects. A weakly polar solvent such as dichloromethane only weakly

265

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

300 pure tetra-Pr4L3

180 min 115 min

Δε (M–1 cm–1)

200

75 min

45 min

100 15 min 10 min

0 5 min

–100

pure tris-Pr[L]2 260

280

300 λ (nm)

320

340

360

Figure 7.12. Time-dependent evolution of the CD profile of tris-Pr[(+)-L]2 in CH3 CN [56]. (Reproduced by permission of the American Chemical Society.)

Figure 7.13. Postulated structures in DMSO (left, M predominating) and CH2 Cl2 (right, P exclusively) [57]. (Reproduced by permission of the Royal Society of Chemistry.)

interacts with the hydroxyl groups of the ligand, allowing for intramolecular hydrogen bonding. Such hydrogen bonding serves to rigidify the structure and lock the complex into the P conformation. Polar solvents such as DMSO, though, interact strongly with the ligand hydroxyl group and push the hydroxyl groups apart, leading to a preference for the M conformation. The authors hypothesize that such a reversible solvent-triggered complex could serve as a means to control stereoselectivity in future metal-catalyzed reactions.

7.3.3. Redox-Triggered Systems The rich coordination chemistry literature offers many avenues for entry into the design of redox-sensitive metal complexes that display rich chiroptical spectra. Redox-active metal ions themselves often show useful electronic spectral changes. However, changes in CD spectra of the organic ligand are also very useful, particularly in complexes that display CD exciton chirality. 7.3.3.1. Iron Translocation in Triple-Stranded Helical Complexes. Shanzer and co-workers [58] reported the first published example of a redox-mediated chiroptical redox switch. The system was based on chemical triggering of iron translocation in

266

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Fe2+

+e–

Fe3+

–e–

(a)

O N

N H

(b)

H N CH3

OH N O

O

O N H

N N 3

Figure 7.14. Tripodal ligand containing two binding sites and the redox-switched chirality of its iron complexes [58]. (Reprinted by permission from Macmillan Publishers Ltd. Copyright 1995.)

triple-stranded helical complexes (Figure 7.14). The design accommodated a single metal ion in one of two sites, either a “hard” binding N3 O3 cavity presenting three hydroxamate moieties or a “soft” N6 -cavity with three bipyridyl ligands. Chemical reduction of Fe(III) to Fe(II) induced the metal to translocate from the hydroxamate binding site to a bipyridyl site, because the “softer” Fe(II) favored the site with more nitrogen ligands. Redox switching of the complex was induced by reduction with ascorbate and oxidation with ammonium persulfate. Pronounced differences in UV–vis and CD spectra were observed corresponding to changes in absorbance associated with Fe(II) versus Fe(III) electronic spectra. A split CD spectrum in the UV region was observed that was three times more intense for the Fe(II) state, suggesting exciton interactions involving the bipyridyl moieties. Reduction was rapid, and oxidation gave the Fe(III) absorbance spectrum after a few minutes (several hours were required to achieve the original Fe(III) CD spectrum). The fact that metal exchange did not occur between control compounds with single metal binding sites suggested intramolecular translocation reaction. Variation of the structure resulted in significantly different translocation rates. 7.3.3.2. Chiroptical Tripodal Ligands. Three-armed, or tripodal, ligand–metal complexes have been found to offer particularly rich stereodynamic behavior, especially when coupled with exciton chirality analysis. The development of redox-triggered chiroptical switches in the Canary laboratory began with the observation that tripodal, N4 ligands form stable coordination complexes with divalent metal cations [59, 60]. In the case of Zn(II) and Cu(II), the ligand, otherwise conformationally mobile with many conformations, wraps around the metal ion to form a propeller-like complex. In ligands with a single stereogenic center on one of the tripod arms, the helicity of the propeller formed by the

267

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

planes of the heterocycles is dictated by this stereocenter. A number of crystallographic structures [61], with few exceptions, established the relationship between the chiral carbon center and the propeller configuration [62]. ECCD established the preponderance of a single propeller conformation in solution, and it tested whether the same configuration was present relative to the carbon center as had been observed in solid-state studies [63]. It was confirmed that Cu(II) and Zn(II) complexes showed ECCD spectra consistent with solid-state propeller-like structures. This method allowed the assignment of the absolute configuration of secondary amines from the sign of the observed excitant couplet. Furthermore, the sign of the couplet discloses the sense of the propeller twist in solution. The dependence of CD exciton chirality upon the strength of the electronic transition moment, the proximity of the coupled transitions, and the angle between them led to the development of several interesting chiroptical molecular switches. An on/off system was studied involving a tripodal ligand containing three quinoline moieties [64]. The tris(quinoline) compound in Figure 7.15 forms a coordination complex with Cu(II) (8) involving the coordination of four nitrogen atoms and affording an exceptionally intense split CD spectrum that results from the additive effect of three ECCD couplets in one molecule. Reduction to the Cu(I) complex in the presence of strongly coordinating thiocyanate ion gave dissociation of one quinoline arm. This resulted in a much weaker ECCD spectrum due to two factors: (1) the dissociation of one quinoline reduces the number of ECCD couplets from three to one, and (2) the less sterically crowded environment around the copper ion allows unwinding of the ligand and reduces the magnitude

S C N N

S C N N N

+e–

CuII H

Cu I H

–e–

N

CH3

N CH3

N (a)

200

66000

100

33000

Δε 0

0 [Θ]

–100

–33000

–200

–66000

–300

–99000

–400

–132000

CuI –165000 CuII –600 –198000 200 210 220 230 240 250 260

0 [Cu(I)]

OFF

–100 –200

Δε

8

–500

N

N

–300 –400

[Cu(II)] –500

ON 0

1

2

3

4

5

Reversibility at 240 nm

λ (nm)

(b)

(c)

Figure 7.15. On/off chiroptical molecular switch. (a) One-electron reduction results in dissociation of an arm of the tripodal ligand. (b) CD spectra of Cu+ and Cu2+ complexes. (c) oxidation and reduction cycles with ascorbate and persulfate [64]. (Reproduced by permission of Wiley-VCH.)

268

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the dihedral angle and therefore diminishes the amplitude of the ECCD couplet. The overall effect is a very large difference in ECCD amplitude between Cu(I) and Cu(II) states. Dependence of the ECCD amplitude on the counterion supported the structural assignment [65]. The complex is highly reversible chemically upon oxidation of the Cu(I) complex with ammonium persulfate and reduction of the Cu(II) complex with sodium ascorbate. Temperature-dependent 1 H NMR studies of this system led to the conclusion that the two arms lacking the chiral carbon center are in rapid equilibrium between associated and dissociated states at room temperature, but slow on the 1 H NMR timescale at low temperature [66]. The arm containing the chiral carbon center, however, remains coordinated. Such tripodal ligands were found to act as chemosensor molecules by virtue of their ability to torque a nematic into a cholesteric liquid crystalline phase increased upon complexation with copper ion [67]. Changes in overall shape of the complexes induced by different metals and counterions were transferred sensitively to the supramolecular level, observed by proportionate changes in the degree of twisting. Redox changes (Cu(I)/Cu(II)) also gave large changes in twisting power. The handedness of the induced cholesteric phase was related to the stereochemistry of the ligand. Interestingly, a direct correlation was observed between helical twisting power and ECCD amplitude, consistent with each technique responding proportionately to the relative twist of the planes of the nitrogen heterocycles. Another related complex containing two chiral carbon centers within a piperidine ring (9) was reported (Figure 7.16) [68]. In this case, the rigidity of the ligand provided control as to which chair form of the piperidine was adopted. In the Cu(I) oxidation state, the ligand adopts a relatively stable cyclohexane chair conformation, with two equatorial and one axial substituent. This conformation places one pyridine moiety remote from the metal ion, but this is accommodated by the lower coordination number of the Cu(I) ion. In the Cu(II) state, strong binding to the higher-coordination number ion brings all three pyridines into association, which forces the piperidine to adopt a higher-energy chair with two axial and one equatorial substituents. The CD spectrum of the Cu(II) complex showed the largest amplitude of any complex in this series, but the Cu(I) spectrum did not give an exciton chirality spectrum. In this case, the Cu(I) structure was characterized by a series of 1 H NMR experiments. In these studies, CD exciton chirality served as a tool to gauge not only the configuration of the propeller conformation but also the degree of twist of the molecule. Relatively few spectroscopic probes are available to report 3D molecular geometry, so it may be expected that this technique should be broadly applicable for the characterization of solution species [16]. Systematic exploration of amino acid derivatives by Canary and co-workers [69, 70] led to the discovery of a molecule that inverts helicity and CD couplet sign upon one-electron redox change [71]. A ligand derived from the amino acid methionine forms a tetradentate complex with Cu(II) involving three nitrogen atoms and a carboxylate (Figure 7.17). In this system, the propeller twist of the molecule is dictated by the asymmetric carbon center by virtue of a gearing mechanism between the methine and methylene carbon atoms, and it can be visualized when viewing down the bond between the tertiary amine nitrogen atom and the Cu(II) ion. Upon reduction to Cu(I), the ligand reorganizes and the sulfide moiety replaces the carboxylate, which is expected due to the preference of Cu(I) for this type of coordination. The reorganization requires a pivot about the bond between the tertiary nitrogen atom and the asymmetric carbon atom. This pivot destroys the gear previously mentioned; to retain the geared conformation, the

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

269

X N

CuI

N

N

X

H

H CuII HN

N H

N

N

N X = sovent 9

800 600 400

Free ligand CuI complex CuII complex CuI complex oxidized CuII complex reduced ZnII complex

Δε

200 0 –200 –400 20

12 8 4

200

210

220

230 240 λ (nm)

250

260

0 270

ε × 10–4

16

Figure 7.16. Redox-triggered inversion of one chair form of a piperidine ring into the other chair and corresponding CD (top) and UV–vis (bottom) spectra [68]. (Reproduced by permission of the American Chemical Society.)

two methylene carbon groups flip, which, in turn, inverts the helical orientation of the two quinoline moieties and, therefore, the exciton chirality spectrum. The CD spectrum appears to give mirror images for the Cu(I) versus Cu(II) complexes. The switching was reversible with cyclical additions of ascorbate and ammonium persulfate. Crystallographic data supported the structural assignments [72]. The Cu(I)/Cu(II) complexes of other tripodal ligands also give inversion of the CD spectrum including derivatives of methioninol and S -methylcysteine [73]. 7.3.3.3. Redox-Controlled Molecular Flipper Based on a Chiral Cu Complex. Copper redox chemistry has also been explored in other ligand platforms for redox-dependent chiroptical effects. A molecular bipaddled flipper based on a tetradentate chiral Cu complex was reported whose paddling motion could be controlled

270

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b) 40

N O CuII H N

H

Ar

H N ArO CO2-H CH3SCH2

H

N

S

+e– –e–

I

Cu N

N -O C 2 H

N

H

–60

[Cu(I)] [Cu(II)]

Cu

5

0

4

–100

3 2

II

Cu –200

(c)

7 6

I

100 Δε239

Ar

H –20 –40 –80

CO2–

S

Ar

CH3SCH2

N

20 0

0

1

2

3

ε × 10–4

X

Δε

+

–

1 210 220 230 240 250 260 λ

n

Figure 7.17. Redox-induced inversion of helicity. (a) As a result of the presence of gearing among the three arms of the tripod near the sterically crowded tertiary amine of the ligand, a pivot about a C–N bond results in the inversion of the propeller. (b) CD and UV–vis spectra of Cu+ and Cu2+ oxidation states. (c) chemical cycling with ascorbate and persulfate [71]. (Reproduced by permission of the American Association for the Advancement of Science.)

N (R) (R) N (R)

CH3

(R) N

CuI N

[Δ-10]+ +e–

[Λ-11]+ +e–

–e–

–e–

Figure 7.18. Crystal structure of [-10]+ (top left) and DFT structures of [-11]+ (top right), [-12]2+ (bottom

CH3

Δ-10PF6 [Δ-12]2+

[Λ-13]2+

left), and [-13]2+ (bottom right) [74]. (Reproduced by permission of the American Chemical Society.)

by reversible oxidation of the metal center (Figure 7.18) [74]. The isomeric pair of Cu(I) complexes -[CuI ((NR , NR , R, R)-L)]PF6 (-10PF6 ) and -[CuI ((NS , NS , R, R)-L)]PF6 (-11PF6 ) interconvert, a slight preference for -11PF6 , (Keq = 1.3), which can be monitored by time-dependent CD starting with a pure -10PF6 in CH2 Cl2 (Figure 7.19a). The amplitude of the Cotton effects decreased with time and eventually rested at smaller amplitudes with opposite signs within a couple of hours. The inversion of the chirality of the aliphatic N atoms resulted in the inversion of helicate chirality. The mechanical motion undergone by -10PF6 and -11PF6 could be switched on/off by reversible metal oxidation and reduction electrochemically or chemically. An oxidation experiment of species -10PF6 with AgBF4 was performed at different times of the isomerization process and the CD of the resulting Cu(II) species (-12PF6 and/or 13PF6 ) recorded (Figure 7.19b). The spectra varied, depending on when the oxidant was added, which corresponded to the degree of [-10]+ /[-11]+ isomerization. However, when AgBF4 was added to a pure -10PF6 sample, all of the -10PF6 was oxidized to -12PF6 and no isomerization to -13PF6 or -11PF6 was observed. The transition

271

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

7 3

40

CD (mdeg)

CD (mdeg)

80

0 –40

t = 30

t = 123

t=0

–7

× 10

–3

–80 –120 235

0

t = 123

t=0 t = 30

335

435 λ (nm) (a)

535

–1 235

436

635

835

λ (nm) (b)

Figure 7.19. (a) CD monitoring of the -10PF6 /-11PF6 isomerization reaction in CH2 Cl2 . (b) CD spectra of the oxidized -10PF6 /-11PF6 couple with AgBF4 at different isomerization times yielding [-12]2+ at t = 0 min and mixtures of [-12]2+ and [-13]2+ at t > 0 min [74]. (Reproduced by permission of the American Chemical Society.)

metal integrated into this device acts as a redox switch that permits one to start/stop the motion at will. 7.3.3.4. Redox-Switchable Pt-Bridged Cofacial Diporphyrins via Carbon– Metal σ Bonds. In one of several porphyrin-containing systems, Shinokubo, Osuka, and co-workers [75] reported the construction of a Pt(IV)-bridged cofacial diporphyrin architecture and its dynamic helical conformational change by reduction of the bridge to Pt(II) (Figure 7.20). Two stable Pt-C σ bonds supported by the pyridyl groups brought two porphyrin macrocycles to be in close proximity in each of these two complexes. The platinum bridge offers conformational flexibility to the complexes due to the susceptibility of platinum toward redox reaction. These complexes also exhibit helical chirality. Reduction of M spiral of 14M mainly yielded the P spiral enantiomer of 15P . The Pt(IV) complex 14M exhibits exciton coupling in both UV–vis and CD, while there is no exciton coupling in the Pt(II) complex 15P . 7.3.3.5. Redox-Triggered Porphyrin Tweezers. Recently, a redox-triggered porphyrin tweezer was reported in an attempt to develop materials with optical properties in the visible region of the electromagnetic spectrum [76]. As shown in Figure 7.21, bis(porphyrin) methioninol derivative (16) gave a strong ECCD couplet upon metallation with Cu(II). The free ligand and Cu(I) complex did not give ECCD. The absence of an ECCD couplet in the Cu(I) complex was rationalized as resulting from relatively weak association of the metal under the conditions studied. The Cu(II) complex, however, showed very strong amplitude, affording an on/off chiroptical molecular switch. Other, nonchiral electrochemically responsive dimeric porphyrin systems have been reported where the redox changes occurred within the porphyrin moieties [77]. 7.3.3.6. Redox-Controlled Dinuclear Ruthenium-Based Switches Monitored by Electronic Near-IR CD. A system showing strong changes in near-infrared (NIR) CD spectra was reported recently [78]. NIR techniques are of interest for several reasons, including the benefit of lower incident light energy on organic materials and greater transparency of NIR light in biological applications. Building on earlier studies of organic-based systems [79–81], the Wang laboratory studied dinuclear ruthenium complexes with 1,2-dicarbonylhydrazizo bridging ligands , -17 and , -17

272

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N N

2e–

N Ni

N

CI

N N

N

N

N Ni

Ni

N

N

N N

CI N

N

Ni

N N

N

14M spiral

15P spiral

200

Δε (M–1cm–1)

100

0

–100 14M –200

15P

–300

–400 300

400

500

600

700

800

Wavelength (nm)

Figure 7.20. (Top) Reduction of the M spiral Pt(V) complexe 14M result in P spiral Pt(II) complex 15P. (Bottom) CD spectra. [75]. (Reproduced by permission of the American Chemical Society.)

(Figure 7.22) that are highly electrochromic with absorption bands near 500, 900, and 1200–1600 nm. Ligand-centered transitions in the UV region and redox-sensitive MLCT bands in the visible region dominate the CD spectra shown in Figure 7.22. A prominent band near 1115 nm observed in the Ru(II)/Ru(III) state, due to metal–metal charge transfer (MMCT), did not give a strong Cotton effect in the CD spectrum. The Ru(III)/Ru(III) state gave a strong MLCT band at 900 nm that gave a relatively strong Cotton effect in the CD spectrum. Reversible redox switch behavior was demonstrated by monitoring the CD signal at 890 nm and cycling up to seven times electrochemically between the Ru(II)/Ru(II) and Ru(III)/Ru(III) states. A variety of systems have thus been examined for redox-active metal ion triggered chiroptical molecular switches. The mechanisms reported involve translocation of a metal

273

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

N N C N N

Cu(CIO4)2

N HN

NH4SCN

NH N

N N

Cu

N

N

SCN

OH H

N HO HCu N N

16 N N 2

SCH3

SCH3

300 3eq of Cu(II) and NH4NCS free ligand

200 Δε

100 0 –100 –200

A

–300 0,16 0,14 0,12 0,10 0,08 0,06 0,04 0,02 0,00

360

390

360

390

420

450

480

420

450

480

λ (nm)

Figure 7.21. Redox-triggered reorientation of porphyrins [76]. (Reproduced by permission of the American Chemical Society.)

ion, changes of lability of ligand rearrangement, or inner sphere ligand rearrangement resulting from change in coordination number or hardness of the metal. The changes in amplitude of observed CD spectra can be dramatic, even leading to complete inversion of the sign of the ECCD couplet.

7.3.4. Photochemically Triggered Chiral Metal Switches Among many interesting studies, the Aida group used ECD to characterize a redoxtriggered system in which chemical or photoreduction of a chiral cerium bisporphyrinate double-decker complex resulted in racemization by acceleration of the porphyrin ligand

274

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

n+

N

N

HN

Ru

N O

N N

N

Pr

N

O N

Ru

NH

Pr

N N

3×10 2×10

5

RuII/RuII RuII/RuIII RuIII/RuIII

1×105 0 –1×105 –2×105 –3×105

300

400 500 600 Wavelength (nm)

Molar Ellipticity (deg x cm2/dmol)

Molar Ellipticity (deg x cm2/dmol)

Λ, Λ-17 5

700

3×104 RuII/RuII RuII/RuIII RuIII/RuIII

2×104 1×104

0 –1×104 –2×104 –3×104 600

800

1000

1200

1400

Wavelength (nm)

(a)

(b)

Figure 7.22. CD spectra of , -isomer of a diruthenium complex 17 at different oxidation states [78]. (Reproduced by permission of the Royal Society of Chemistry.)

rotation. They further showed that oxidation of a chiral zirconium complex resulted in deceleration of acid-induced racemization [82]. 7.3.4.1. Azobenzene-Based Molecular Scissors. The Aida group carried out the synthesis of other complex light-triggered chiroptical molecular switches. Life-sized scissors having a handle, pivot, and blades inspired the preliminary design of a pair of molecular “scissors” [83]. The chemical equivalents to these three units were found to be azobenzene as the handle, ferrocene as a pivot, and phenyl groups as the blades (Figures 7.23a and 7.23c). The operation of the molecular scissor is quite elegant. Under standard conditions, the azobenzene handle is predominantly in the trans state leading to “closed” blades. Under irradiation of UV light, the azobenzene undergoes isomerization to the cis isomer, which then causes a slight rotation of the cyclopentadienyl rings of the ferrocene pivot. This finally moves the attached phenyl rings away from one another, leading to an “open” scissor state. The scissors’ chirality (due to the planar 1,1 ,3,3 -tetrasubstituted ferrocene) allows both open and closed states to be seen using circular dichroism (Figure 7.23b). The authors explain that the trans-to-cis isomerization of [CD(−)280]-trans-18 upon UV-irradiation (λ = 350 nm) after 180 s gave rise to CD spectral changes at 240–300 nm due to the major adsorption of the tetraarylferrocene unit. Upon irradiation with visible light (λ > 400 nm), a reverse spectral change occurred, where the system quickly reached a photostationary state in 15 s. Effective reversibility was also exhibited by the system upon sequential irradiation with UV and visible light.

275

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

+2.0

Handle Pivot

N N

Fe 18 (a)

Δε

+1.0 0.0

–1.0

trans-18 10 sec 20 sec 60 sec 180 sec

–2.0 235 260 285 310 335 360 λ (nm) (b)

Blade

Azobenze Strap

(c)

Figure 7.23. (a) Structural representation of azobenzene controlled ‘‘molecular scissors.’’ (b) CD spectral changes of trans-18 upon irradiation with UV light. (c) Graphic conceptualization of the ‘‘molecular scissors’’ [83, 86]. [Reproduced by permission of the American Chemical Society (a, b) and the Royal Society of Chemistry (c).]

7.3.4.2. Host-Controlled Guest Chirality. Further research into Aida’s molecular scissors proved that they could be applied to the field of host–guest chemistry [84]. When metallated porphyrins were attached to the 4-position of the phenyl blades, it was found that a diisoquinoline guest was able to chelate to the zinc porphyrin units. Upon irradiation of the host–guest complex with UV light (λ = 350 ± 10 nm), the trans-azobenzene unit again isomerizes to the cis-isomer, causing a long-distance conformational twist of the diisoquinoline guest (Figure 7.24). The guest molecule (19) in solution is initially achiral due to its conformational freedom; but when added to the host molecule (trans-20), it binds in a nonplanar CD-active chiral geometry. Overlap of CD bands (275–350 nm) from the host molecule required that differential CD spectra be used to examine the motion of the guest (Figure 7.24b). Irradiation with UV light caused the Cotton effects at 270–350 nm of the guest (19) to diminish and then vanish. It was reasoned that the disappearance of the CD band is caused by the guest molecule being forced into a nearly planar state when bound to the cis-isomer of the host compound. Sequential irradiation of host–guest complex (19 • trans-20) with UV and visible light proved that the complex was controllably reversible (Figure 7.24c). This represents the first instance of a molecular machine causing chirality manipulation in a controllable and reversible manner. 7.3.4.3. Chirality Transfer via Ternary Complex. Recently, Aida and coworkers [85] have created many similar compounds incorporating the molecular scissor as a basis for more elaborate and complex systems. Such systems include a ternary compound, which includes a pyridine-appended dithienylethene derivative as a photochromic module that can again be used to transfer conformational information with UV and visible light as a trigger [85]. Extension of the pivotal ferrocene has also been adopted in reversible self-locking compounds shown in Figure 7.25. In the presence of trans1,2-bispyridine ethylene, the zinc–porphyrin moieties coordinate intramolecularly with the anilines to “lock” the molecule internally [86]. UV light is then used to isomerize to the cis-1,2-bispyridine ethylene that is then capable of coordinating to the zinc porphyrin units, “locking” the molecules externally. The process is again shown to be reversible by alternating UV and visible light irradiation. Such discoveries by Aida and co-workers could help to controllably transmit chiral and mechanical information through long molecular distances.

276

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N Ar N Zn N N Ar

Ar

(a)

N Zn N N N

N

Ar

N

Ar N

19

N

Vis

N N Zn

N

N

N

UV

N

Fe

Ar

Fe

19

N N

N Zn N N Ar

Trans-20

Ar

Cis-20 19·Cis-20

19·Trans-20 100

(b)

[19]/Trans-[20]

Δε

50

0.0 0.3 0.6 1.0 2.0

0

–50

–100

250

300

350

400

450

500

Wavelength (nm) 6.0

(c) Δε

4.5 3.0 1.5

NN N Zn NN

0

Δε

150

N

NN Zn NN

100 50

[trans-20]

[trans-20] + [cis-20]

0 1.00 0.75 0.50 N N

0.25 0.00

0 10 20 30 40 50 60 70 80 90 Times (s)

Figure 7.24. (a) Photoisomerization of a 1:1 complex of molecular pedal 20 with rotary guest 19 (19 • 20). (b) CD spectral changes of trans-20 upon titration with 19. (c) CD visualization of the motions of guest-binding molecular pedals 19•(+)-20 and 19•(−)-20 triggered by light [84]. (Reprinted by permission from Macmillan Publishers Ltd: Copyright 2006.)

277

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

H2 N

NH2

N

N N Zn N

N

N N Zn N

N

UV N N N Zn

N N N Zn N

Vis

trans-22

N

N

N N H2 Internally Double-locked 21

NH2 Externally Locked 21 in 21⊃cis-22

Figure 7.25. Structures of internally double-locked 21 and externally locked 21 ⊃ cis−22, and the self-locking operation in response to photochemical isomerization of 22 [86]. (Reproduced by permission of the Royal Society of Chemistry.)

7.4. DYNAMIC STEREOCHEMISTRY MONITORED BY VCD VCD spectra have been applied to study transition-metal complexes since the pioneering work by Nafie and co-workers [8, 13, 87] VCD spectroscopy was applied to monitoring in situ the photoinduced rewind of supramolecular helices in a liquid crystal (Figure 7.26) [11]. A room-temperature liquid crystal, ZLI-1132, was doped with a chiral Cr(III) complex -[Cr(acac)2 (2C12)] (acac = acetylacetonate; 2C12 = 4, 4 -didodecyloxyated dibenzoylmethanate). The selective reflection wavelength λc = np (n is the average refractive index, p is the pitch length of an induced helix) of the nematic phase was determined to be 5.3 μm. At this wavelength, circularly polarized light reflects from or passes through the sample when it has the same or opposite sense of the induced helix, respectively. Under the illumination of UV light (365 nm), the photoracemization of the Cr(III) complex rewound helices in the chiral nematic phase. In response to this, the VCD spectrum of the system exhibited the transient change. Figure 7.26 shows the time course of the VCD spectrum recorded every 30 min. The peak at 1610 cm−1 increased intensity for the initial 1 h. The change reflected the elongation of the pitch maintaining the relation of λ/λc > 1. The peak underwent a drastic change at 1.5 h: the spectral shape transformed from a Gaussian to a biphasic one, which indicated the relation of λ/λc = 1 was fulfilled at this stage. After 2 h, the peak returned to a Gaussian shape with a negative sign, indicating the relation of λ/λc < 1. Reflecting the further elongation of the helical pitch, the position of a biphasic peak shifted toward the longer wavelength. When the VCD spectrum is regarded as a memory signal for the photoresponsive events in this liquid-crystalline system, it shows high signal-to-noise ratio (S/N) since it changes the sign as well as the intensity. The spectral change conveys the information of time memory because the position of the biphasic shaped peak shifts toward the longer wavelength on continuing irradiation. It

278

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

C12H25O O O

O O

O

Cr

Cr O

O O C12H25O

6h 5.5h 5h 4.5h 4h 3.5h 3h 2.5h 2h 1.5h 1h 0.5h 0h

OC12H25

O O

O O

Δ Cr(acac)2(2C12) Λ

OC12H25

1800

1600

1400 1200 Wavenumber (cm–1)

1000

Figure 7.26. VCD spectra of a chiral nematic sample of ZLI-1132 doped with 0.538 mol% -[Cr(acac)2 (2C12)] at various times after UV light (365 nm) irradiation [11]. (Reproduced by permission of Taylor and Francis.)

shows extremely stable memory since the racemization process is accompanied by an increase of entropy, thus irreversible.

7.5. DYNAMIC STEREOCHEMISTRY MONITORED BY FDCD AND CPE Fluorescence-detected circular dichroism (FDCD) is a method that can measure the CD response by detection in emission if the chromophore is also fluorescent. This method was originally developed by Tinoco and co-workers [88, 89] and studied intensively more recently by Berova, Nakanishi, and co-workers [90–92]. While conventional CD measures the difference in a sample’s absorption of left- and right-circularly polarized light, FDCD measures the difference in fluorescence intensity upon excitation by leftand right-circularly polarized light. Since it is usually true that the excitation spectrum of a fluorophore parallels its absorption, the same circular dichroic information should be able to be extracted from both processes if the artifacts related to undesired fluorescence anisotropy are eliminated by more advanced instrumentation [92]. FDCD has been shown to be more sensitive than absorption CD [90, 93], analogous to the fact that fluorescence spectroscopy is more sensitive than the UV–vis absorption method because fluorescence suffers no background interference from the incident light. Raw FDCD data measured by a JASCO circular dichroism system equipped with FDCD attachment and with the fluorescence detector placed at 90◦ to the excitation beam (i.e., 90◦ to the CD detector) is recorded in two channels [90]. They represent excitation spectra and correspond to the difference in emission F (FL − FR ) and the total emission (FL + FR ) resulting from differential absorption of left- and right-circularly polarized light, respectively. Typically, the data are converted to CD spectra by established methods, which gives a normal CD spectrum if fluorescence polarization is negligible: The FDCD and normal CD of zinc complex of chiral tripodal ligand 23 match perfectly. An adaptation of the FDCD technique can provide a unique and powerful new strategy for sensor applications by using the F , that is, (FL − FR ), component of FDCD data directly, without conversion to CD. To distinguish this new approach from traditional FDCD and to avoid confusion, this method was named differential circularly polarized fluorescence excitation (CPE) [17], although no new instrument is required and all of the advantages and nature of FDCD still pertain. This is different from circularly polarized luminescence (CPL) because CPL is the differential spontaneous emission of left- and right-circularly polarized light and reflects the structural properties of the excited

279

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

state, while CPE is still an indirect reflection of the structural properties of the ground electronic state. The theoretical basis of CPE can be derived from long-established Eq. (7.1) [90]. F =

θ · (FL + FR ) · ln 10 . 33 · 2 · (10A − 1) · k

(7.1)

If A = AL − AR ≤ 0.1 and A/A ≤ 0.1, total emission (FL + FR ) should be proportional to fluorescence induced by nonpolarized light, that is, FL + FR = k2 · F = k2 · F · I 0 · (1 − 10−A ),

(7.2)

Therefore Eq. (7.1) can be simplified as shown in Eq. (7.3), where K is a constant, which incorporates all other constants and F is the fluorescence quantum yield. F =

θ · F · I 0 θ · F · I 0 · k2 · ln 10 = K · . 33 · 2 · 10A · k 10A

(7.3)

Materials with higher ellipticity θ and higher fluorescence quantum yield F will lead to an even larger F . Substances lacking either fluorescence or CD properties will not be observed. Shown in Figure 7.27 are the CD and CPE (F) responses of a chiral piperidine compound (23) to Zn2+ [17]. Apparently, FDCD and CPE may be used to monitor the chirality switching in such metal complexes.

N

N

N

N H

H 23

20 15

0.4 μM incremental 0 μM Zn2+

5

5 ΔF (V)

θ (mdeg)

10

0 –5

–10

5

Zn2+ 5.2 μM Zn2+

Zn2+

0 μM Zn2+ 0.4 μM incremental 5.2 μM Zn2+

–15 200 220 240 260 280 300 320 340 λ (nm) (a)

–5

0 μM Zn2+ 0.4 μM incremental

10 15

Zn2+ 5.2 μM Zn2+ 20 200 220 240 260 280 300 320 340 λ (nm) (b)

Figure 7.27. Spectral responses of 4.8 μM (R,R)-23 to 0–5.2 μM Zn(II) in acetonitrile. (a) CD. (b) CPE (700V, 81 deg, filter: 360 nm) [17]. (Reproduced by permission of the American Chemical Society.)

280

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6 3+

Me

Ph

4

HN R

O

O N Me

N H

N

N

N

Ln

N

N

N

O NH Ph

(SSS)-Δ-[Ln·L25]3+

(IL − IR)

N

2 0 480

Ln = Eu, Gd, or Tb

−2

25a: R = H 25b: R = COOMe

−4

530

580

630

Me

Figure 7.28. (Left) CPL spectra for (SSS)--[Tb.L

Wavelength (nm)

25b 3+

] (solid curve) and in the presence of BSA (broken curve). (Right) Chiral lanthanide metal–ligand complex used to bind human or bovine serum albumin in ‘‘drug site II’’ [18]. (Reproduced by permission of the Royal Society of Chemistry.)

7.6. DYNAMIC STEREOCHEMISTRY MONITORED BY CPL Circularly polarized luminescence (CPL), the anisotropic emission of circularly polarized light originated from nonpolarized excitation, is the emission analogue to CD. The sign and magnitude of CPL are affected by the degree of helical twist of the complex, the nature of the ligand field, and other factors. In this context, excited Ln(III) ions can be considered as “spherical” emitters and avoid the problems associated with anisotropy that can complicate some chiroptical analyses [94]. CPL reflects the time-averaged local helicity around the lanthanide(III) ion. The Parker group has recently utilized the chiral environment of drug site II of serum albumin to induce helicity inversion in complexes of terbium and europium (III) [18]. It was found that chiral complex (S , S , S )--[Tb.L25 ]3+ changed helicity to (S , S , S )--[Tb.L25 ]3+ upon addition of human or bovine serum albumin. Convincing data was supplied by circularly polarized emission (Figure 7.28). When the -isomer is exposed to BSA or HSA, there is an inversion of the sign of emission and 35% reduction of the signal intensity. The authors explain that the emission spectra “are consistent with the inversion of the helicity of the complex in the protein-bound form” [95]. Parallel experiments were run using the -isomer, but no change in emission spectra was found. These results give one of the few immediately biologically relevant examples of a metal-based chiroptical molecular switch. The system could potentially allow protein association to be tracked in vitro in real time.

7.7. SOLID-STATE METAL-BASED CHIROPTICAL SWITCHES 7.7.1. Pressure-based switches Pressure also has been reported to induce chiroptical responses in chiral metal complexes. In solution, high pressure can provide a powerful solvent effect since dispersive interactions depend strongly on density changes. In the solid state, crystal packing plays an additional role. The effect of pressure on circular CD spectra of the octahedral chiral -, -, and (, )tris{O, O -bis[(+)(S )-2-methylbutyl]dithiophosphate}Cr(III)

281

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

300

(Δ,Δ) − Cr[(S)(S)Mebdtp]3

200

S

Me O P O Me

O

Me

P S

S

Me O

O P

Δ − (S,S)(S,S)(S,S)

S M

S S

S

S

M S

1.7

Me

P O Me O Me

S

S P

S

Me O O Me

Me O P O Me

Δ − (S,S)(S,S)(S,S)

CD/mdeg

Me O

100 0

1.2

−100 −200

0.2 GPa

−300 400

500

600 λ (nm)

700

800

CD spectra at different pressures of solid -tris{O, O -bis[(+)(S)-2 -methylbutyl]dithiophosphate}Cr(III) and proposed ,  conversion mechanism [96]. (Reproduced by permission of Taylor and Francis.)

Figure

7.29.

complexes was studied in the pressure range 0–2.5 GPa (Figure 7.29) [96]. Results on polycrystalline samples dispersed in nujol show a pressure-induced -to- inversion of configuration at the metal center above 1.2 GPa, which was suggested to arise from differential crystal packing in the solid-state structure of the diasteromeric complexes. The -form is confirmed to be the most favored crystal packing among different ligand conformations of the chiral complex under high pressure. When the applied pressure exceeded 2.5 GPa, the CD band obtained from polycrystalline Nujol samples of chiral  and -tris-[cyclic O,O , 1(R), 2(R)-dimethylethylene dithiophosphato]chromium (III) complexes inverted from negative to positive, which demonstrated inversion from the -form to the -form by means of pressure [97, 98]. To minimize artifacts, the spectra were obtained from average data from rotating the diamond anvil cell (DAC) around its optical axis, and the reference spectrum was normalized outside the absorption region of the sample. The cycle was reversible as demonstrated by applying repeating pressure cycles. However, the transition pressure varied and was dependent on the amount of -diastereomer present in the sample. Mechanistic explanation of the pressure induced chirality inversion could involve bond breaking or trigonal twisting around the metal center. It would be interesting to see if other solid-state data could be obtained to test the mechanistic hypothesis and exclude artifacts.

7.7.2. Temperature-Induced Dynamic Stereochemistry The compound α-Ni(H2 O)6 · O4 and its selenate derivative exhibit chirality only in the solid state. The Kuroda group observed a remarkable reversible sign inversion of CD in the 3 A2g → 3T1g (P) Ni(II) d –d transition at near liquid nitrogen temperatures, although the crystal structure hardly changes from 300 to 100 K (Figure 7.30) [99]. The change in Ni2+ electronic states at low temperatures might have altered the relative magnitude of the opposite sign first- and second-order rotational strengths.

7.7.3. Photo-induced Switching Switching of molecular chirality under photoirradiation was studied in a cobaloxime complex crystal using CD (Figure 7.31) [100]. The (S -alkyl)(S -base) crystal was irradiated using two different wavelength bands, one with 439–499 nm covering the LMCT transition and the other with 640–900 nm covering the triplet d –d transition of Co(III). After irradiation with either wavelength band, the solid was dissolved in methanol and the changes in its CD spectrum were recorded. Excitation of the d –d transition of the Co(III) ion appeared to be much more effective in inducing the chirality change than

282

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

800 600

3

CD (mdeg)

296 k

A2g→3T1g(P)

3A →3T (F) 2g 1g

400 200 0 –200 300

83 k 400

500

600

700

Wavelength (nm) (a) 200

400 300

CD (mdeg)

CD (mdeg)

100

@384 cm

P41212 (0.062 cm)

200 100 0 –100

Cooling Heating

0 –100 –200

P41212 (0.06 cm)

–200

–300

–300 –400 0

–50

–100

–150

–200

–400

0

–50

–100

–150

Temperature(°C)

Temperature(°C)

(b)

(c)

–200

Figure 7.30. (a) Observed CD spectra of the same α-Ni(H2 O)6 · SO4 single crystal (P43 21 2, 0.062-mm thickness) at different temperatures. (b) CD signal at 384 nm plotted against temperature for the enantiomorphous α-Ni(H2 O)6 · SO4 single crystals. : P43 21 2 (0.62 mm thickness); : P41 21 2 (0.60 mm thickness). (c) Temperature dependence of the CD values of P41 21 2 crystal on cooling () and on heating (•) [99]. (Reproduced by permission of Elsevier.)

excitation of the ligand–metal charge transfer band, although the latter is more effective in breaking the Co–C bond that initiates the chirality switching. The chirality change versus irradiation time showed a step-like behavior suggesting that chirality switching of molecules occurred in correlation with their nearest neighbors. The same group made direct observation of a photoinduced chirality change or switching in the alkyl ligand of the cobaloxime complex in the hydrated and nonhydrated crystals of the cobaloxime complex by direct CD measurements in the solid phase in Nujol-mull and KBr pellets [101]. The CD spectra of the two crystal forms showed clear differences. Additional CD peaks in the spectra of the nonhydrated crystals seemed to arise from exciton splitting of the charge-transfer band. Photoirradiation induced chirality change or switching in the alkyl part of the molecule, but not in the crystal structure. The CD spectra well reflect such behavior.

283

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

(R-Alkyl)(S-Base)

(S-Alkyl)(S-Base)

O

O

H3C

CH3 hν * H O H O O N N Co N N O H O

H * CH3 O H O O N N Co N N O O H

H3C

H3C * HH2 H

H3C * NH2 H

(a) 100

0

–10 200

(S/R-Alkyl)(S-Base) 300

(S-Alkyl)(S-Base)

400 500 Wavelength (nm) (b)

600

75

25

50

50

25

75

(R-AlkyI) (%)

0

(R-Alkyl)(S-Base) (S-AlkyI) (%)

CD (mdeg.)

10

700 0 0

5 10 15 20 25 Irradiation time (hours)

100 30

(c)

Figure 7.31. (a) Molecular structures of a pair of diastereomers of cobaloxime complex that can be converted to each other by photoisomerization. (b) CD spectra of (S-alkyl)(S-base), (R-alkyl)(Sbase), (S/R-alkyl)(S-base) cobaloxime complexes in 1-mM methanol solution. (c) Time variation of photoinduced chirality change in the cobaloxime complex crystals of (S-alkyl)(S-base), (R-alkyl)(Sbase), and (S/R-alkyl)(S-base) under constant irradiation of light with wavelength of 640–900 nm [100]. (Reproduced by permission of the American Institute of Physics.)

7.8. CONCLUSIONS The future of metal-based chiroptical switches is bright, given the high degree of control available, a multitude of triggering mechanisms, and powerful chiroptical spectroscopy tools available for analysis. With the large number of recently discovered systems, these compounds and materials derived from them could potentially be used for applications including optical displays, complex molecular electronics, chiral resolution, and catalysis. The Pfeiffer effect and metal ion templated synthesis provided early chemistry relevant to more recently developed metal-based chiroptical switches. Environmentresponsive switches have been developed using a large variety of metals and ligands triggered by pressure, counterion alteration, light, and solvent changes. Redox triggered switches have been explored primarily using a tripodal ligand motif. Diazobenzeneferrocene systems were designed to reliably switch the conformations of a set of “molecular scissors,” which were then used in an array of interesting and complex supramolecular machines. Polymer systems have been explored illustrated by the use of metal dopants to cause chiroptical changes in oligothiophene polymers. The studies in this area have

284

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

provided much stimulating new chemistry and exemplify the power of modern molecular design and solution characterization techniques. There is no doubt that there are many more opportunities to develop even more imaginative systems. Many applications for these materials have been discussed, especially in the areas of electronics and sensors, and several of the available systems are poised to make a genuine contribution. Although early phenomena were studied by ORD, electronic circular dichroism experiments were used in nearly all experiments to analyze the conformational changes of the chiral compounds. Although other analytical techniques such as NMR are used to study these systems, it is readily apparent that CD experiments provide accurate and dependable read-out for chiral metal-based switches. The exciton chirality method has been particularly useful as a result of the fact that it gives a sizable and interpretable signal. Indeed, few other spectroscopic measurements give a direct report of three-dimensional shape as exciton chirality does for the orientation of chromophoric units. However, care should be taken in the assignment of CD data as arising from exciton coupling. Newer chiroptical spectroscopic methods are beginning to contribute to this field. The first reports of systems employing NIR-CD, which offers low-energy measurement, have appeared. FDCD can provide very sensitive detection if precautions are taken to avoid artifacts. FDCD or CPE may offer more specific information since it may arise from a subset of transitions compared to CD. VCD offers the possibility to probe vibrational phenomena as illustrated in studies of memory in liquid crystalline switches. All of these newer methods are ripe for further development in chiroptical switch detection strategies. It is also high time for computational methods coupled with chiroptical spectroscopy to play a greater role in not only characterization but also design of these systems. Most metal-based chiroptical switches reported to date were studied in solution, but many applications of chiroptical molecular switches involve the solid phase where chiroptical spectra are more difficult to interpret. Fortunately, the development of solid phase characterization tools and accompanying theory is progressing. In this regard, computation has resulted in renewed interest in ORD and other classic methods due to the possibility of making structural conclusions by matching experiment with theory.

ACKNOWLEDGMENTS We are grateful to the National Science Foundation (CHE-0848234) for generous support of our work in this area. ZD thanks Research Corporation for Science Advancement and the donors of the American Chemical Society Petroleum Research Fund for support of this work.

REFERENCES 1. L. Fabbrizzi, A. Poggi, Transition Metals in Supramolecular Chemistry, Kluwer Academic Publishers, Dordrecht, 1994. 2. B. Champin, P. Mobian, J. P. Sauvage, Chem. Soc. Rev . 2007, 36 , 358–366. 3. H. Amouri, M. Gruselle, Chirality in Transition Metal Chemistry: Molecules, Supramolecular Assemblies and Materials, Wiley, Chichester, 2008. 4. H. Miyake, H. Tsukube, Supramol. Chem. 2005, 17 , 53–59. 5. M. Elhabiri, A. M. Albrecht-Gary, Coord. Chem. Rev . 2008, 252 , 1079–1092.

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

6. K. Nakanishi, N. Berova, R. Woody, Circular Dichroism: Principles and Applications, Wiley-VCH, New York, 2000 7. S. Allenmark, J. Gawronski, Chirality 2008, 20 , 606–608. 8. P. L. Polavarapu, L. A. Nafie, S. A. Benner, T. H. Morton, J. Am. Chem. Soc. 1981, 103 , 5349–5359. 9. L. A. Nafie, J. Mol. Spectrosc. 1995, 347 , 83–100. 10. T. B. Freedman, X. Cao, R. K. Dukor, L. A. Nafie, Chirality 2003, 15 , 743–758. 11. T. Taniguchi, K. Monde, S.-I. Nishimura, J. Yoshida, H. Sato, A. Yamagishi, Mol. Cryst. Liq. Cryst. 2006, 460 , 107–116. 12. H. Sato, D. Shirotani, K. Yamanari, S. Kaizaki, Inorg. Chem. 2010, 49 , 356–358. 13. D. A. Young, E. D. Lipp, L. A. Nafie, J. Am. Chem. Soc. 1985, 107 , 6205–6213. 14. S. Abbate, E. Castiglioni, F. Gangemi, R. Gangemi, G. Longhi, Chirality 2009, 21 , E242–E252. 15. C. Guo, R. D. Shah, R. K. Dukor, X. Cao, T. B. Freedman, L. A. Nafie, Charility 2005, 59 , 1114–1124. 16. N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. 17. Z. Dai, G. Proni, D. Mancheno, S. Karimi, N. Berova, J. W. Canary, J. Am. Chem. Soc. 2004, 126 , 11760–11761. 18. C. P. Montgomery, E. J. New, D. Parker, R. D. Peacock, Chem Commun. 2008, 4261–4263. 19. T. Harada, H. Hayakawa, R. Kuroda, Rev. Sci. Instrum. 2008, 79 , 073103. 20. A. Werner, Ann. 1912, 386 , 1. 21. D. B. Shapiro, R. A. Goldbeck, D. Che, R. M. Esquerra, S. J. Paquette, D. S. Kliger, Biophys. J . 2005, 68 , 326–334. 22. M. Krykunov, M. D. Kundrat, J. Autschbach, J. Chem. Phys. 2006, 125 , 194110–194113. 23. H. Yoneda, Y. Nakashima, U. Sakaguchi, Chem. Lett. 1973, 1343–1346. 24. D. W. Urry, H. Eyrin, J. Am. Chem. Soc. 1964, 86 , 4574–4580. 25. W. T. Shearer, R. K. Brown, J. Biol. Chem. 1966, 241 , 2665–2671. 26. E. Cerasoli, B. K. Sharpe, D. N. Woolfson, J. Am. Chem. Soc. 2005, 127 , 15008–15009. 27. K. Pagel, T. Vagt, T. Kohajda, B. Koksch, Org. Biomol.. Chem. 2005, 3 , 2500–2502. 28. I. Otsuka, R. Sakai, T. Satoh, R. Kakuchi, H. Kaga, T. Kakuchi, J. Polymer Sci. A: Polymer Chem. 2005, 43 , 5855–5863. 29. T. Sanji, Y. Sato, N. Kato, M. Tanaka, Macromolecules 2007, 40 , 4747–4749. 30. H. Goto, E. Yashima, J. Am. Chem. Soc. 2002, 124 , 7943–7949. 31. O. Henze, W. J. Feast, F. Gardebien, P. Jonkheijm, R. Lazzaroni, P. Lecl`ere, E. W. Meijer, A. P. H. J. Schenning, J. Am. Chem. Soc. 2006, 128 , 5923–5929. 32. S. Haraguchi, M. Numata, C. Li, Y. Nakano, M. Fujiki, S. Shinkai, Chem. Lett. 2009, 38 , 254–255. 33. L. A. P. Kane-Maguire, G. G. Wallace, Chem. Soc. Rev . 2010, 39 , 2545–2576. 34. M. Funahashi, N. Tamaoki, Chem. Mater. 2007, 19 , 608–617. 35. P. Pfeiffer, K. Quehl, Ber. 1931, 64 , 2667. 36. H. G. Brittain, Polyhedron 1984, 3 , 1087. 37. G. A. Hembury, V. V. Borovkov, Y. Inoue, Chem. Rev . 2008, 108 , 1–73. 38. N. Wei, N. N. Murthy, K. D. Karlin, Inorg. Chem. 1994, 33 , 6093–6100. 39. J. W. Canary, A. E. Holmes, J. Liu, Enantiomer 2001, 6 , 181–188. 40. H. Ishii, Y. Chen, R. A. Miller, S. Karady, K. Nakanishi, N. Berova, Chirality 2005, 17 , 305–315. 41. N. Berova, G. Pescitelli, A. G. Petrovic, G. Proni, Chem. Commun. 2009, 5958–5980.

285

286

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76.

X. Li, B. Borhan, J. Am. Chem. Soc. 2008, 130 , 16126–16127. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, Org. Lett. 2000, 2 , 1565–1568. V. V. Borovkov, G. A. Hembury, Y. Inoue, Acc. Chem. Res. 2004, 37 , 449–459. S. Nieto, V. M. Lynch, E. V. Anslyn, H. Kim, J. Chin, Org. Lett. 2008, 10 , 5167–5170. S. Nieto, V. M. Lynch, E. V. Anslyn, H. Kim, J. Chin, J. Am. Chem. Soc. 2008, 130 , 9232–9233. S. Yano, M. Nakagoshi, A. Teratani, M. Kato, T. Onaka, M. Iida, T. Tanase, Y. Yamamoto, H. Uekusa, Y. Ohashi, Inorg. Chem. 1997, 36 , 4187–4194. H. Miyake, K. Yoshida, H. Sugimoto, H. Tsukube, J. Am. Chem. Soc. 2004, 126 , 6524–6525. H. Miyake, H. Sugimoto, H. Tamiaki, H. Tsukube, Chem. Commun. 2005, 34 , 4291–4293. H. Miyake, M. Hikita, M. Itazaki, H. Nakazawa, H. Sugimoto, H. Tsukube, Chem.-Eur. J . 2008, 14 , 5393–5396. H. Miyake, H. Kamon, I. Miyahara, H. Sugimoto, H. Tsukube, J. Am. Chem. Soc. 2008, 130 , 792–793. J. Gregolinski, J. Lisowski, Angew. Chem. Intl. Ed . 2006, 45 , 6122–6126. J. Gregolinski, T. Lis, M. Cyganik, J. Lisowski, Inorg. Chem. 2008, 47 , 11527–11534. J. Gregolinski, K. Slepokura, J. Lisowski, Inorg. Chem. 2007, 46 , 7923–7934. J. Gregolinski, P. Starynowicz, K. T. Hua, J. L. Lunkley, G. Muller, J. Lisowski, J. Am. Chem. Soc. 2008, 130 , 17761–17773. M. Lama, O. Mamula, G. S. Kottas, L. De Cola, H. Stoeckli-Evans, S. Shova, Inorg. Chem. 2008, 47 , 8000–8015. M. Hutin, J. Nitschke, Chem. Commun. 2006, 1724–1726. L. Zelikovich, J. Libman, A. Shanzer, Nature 1995, 374 , 790–792. J. W. Canary, C. S. Allen, J. M. Castagnetto, Y. H. Wang, J. Am. Chem. Soc. 1995, 117 , 8484–8485. J. W. Canary, Chem. Soc. Rev . 2009, 38 , 747–756. J. W. Canary, C. S. Allen, J. M. Castagnetto, Y. H. Chiu, P. J. Toscano, Y. H. Wang, Inorg. Chem. 1998, 37 , 6255–6262. X. D. Xu, K. J. Maresca, D. Das, S. Zahn, J. Zubieta, J. W. Canary, Chem. Eur. J . 2002, 8 , 5679–5683. J. M. Castagnetto, X. D. Xu, N. D. Berova, J. W. Canary, Chirality 1997, 9 , 616–622. S. Zahn, J. W. Canary, Angew. Chem. Intl. Ed . 1998, 37 , 305–307. S. Zahn, J. W. Canary, J. Am. Chem. Soc. 2002, 124 , 9204–9211. J. Zhang, K. Siu, C. H. Lin, J. W. Canary, New J. Chem. 2005, 29 , 1147–1151. S. Zahn, G. Proni, G. P. Spada, J. W. Canary, Chem. Eur. J . 2001, 7 , 88–93. J. Zhang, J. W. Canary, Org. Lett. 2006, 8 , 3907–3910. A. E. Holmes, S. Zahn, J. W. Canary, Chirality 2002, 14 , 471–477. S. Zahn, J. W. Canary, Org. Lett. 1999, 1 , 861–864. S. Zahn, J. W. Canary, Science 2000, 288 , 1404–1407. S. Zahn, D. Das, J. W. Canary, Inorg. Chem. 2006, 45 , 6056–6063. H. S. Barcena, A. E. Holmes, J. W. Canary, Org. Lett. 2003, 709–711. A. Company, M. Guell, D. Popa, J. Benet-Buchholz, T. Parella, X. Fontrodona, A. Llobet, M. Sola, X. Ribas, J. M. Luis, M. Costas, Inorg. Chem. 2006, 45 , 9643–9645. S. Yamaguchi, T. Katoh, H. Shinokubo, A. Osuka, J. Am. Chem. Soc. 2008, 130 , 14440–14441. A. E. Holmes, D. Das, J. W. Canary, J. Am.Chem. Soc. 2007, 129 , 1506–1507.

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

77. G. Pognon, C. Boudon, K. J. Schenk, M. Bonin, B. Bach, J. Weiss, J. Am. Chem. Soc. 2006, 128 , 3488–3489. 78. D. Li, Z. Y. Wang, D. Ma, Chem. Commun. 2009, 1529–1531. 79. Z. Y. Wang, E. K. Todd, X. S. Meng, J. P. Gao, J. Am. Chem. Soc. 2005, 127 , 11552–11553. 80. J. Zheng, W. Qiao, X. Wan, J. P. Gao, Z. Y. Wang, Chem. Mater. 2008, 20 , 6163–6168. 81. J. Deng, N. H. Song, W. Liu, Q. F. Zhou, Z. Y. Wang, ChemPhysChem 2008, 9 , 1265–1269. 82. K. Tashiro, K. Konishi, T. Aida, J. Am. Chem. Soc. 2000, 122 , 7921–7926. 83. T. Muraoka, K. Kinbara, Y. Kobayashi, T. Aida, J. Am. Chem. Soc. 2003, 125 , 5612–5613. 84. T. Muraoka, K. Kinbara, T. Aida, Nature 2006, 440 , 512–515. 85. H. Kai, S. Nara, K. Kinbara, T. Aida, J. Am. Chem. Soc. 2008, 130 , 6725–6726. 86. K. Kinbara, T. Muraoka, T. Aida, Org. Biomol. Chem. 2008, 6 , 1871–1876. 87. Y. He, X. Cao, L. A. Nafie, T. B. Freedman, J. Am. Chem. Soc. 2001, 123 , 11320–11321. 88. D. H. Turner, I. J. Tinoco, M. Maestre, J. Am. Chem. Soc. 1974, 96 , 4340–4342. 89. I. J. Tinoco, D. H. Turner, J. Am. Chem. Soc. 1976, 98 , 6453–6456. 90. J.-G. Dong, A. Wada, T. Takakuwa, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 1997, 119 , 12024–12025. 91. T. Nehira, C. A. Parish, S. Jockusch, N. J. Turro, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 1999, 121 , 8681–8691. 92. T. Nehira, K. Tanaka, T. Takakuwa, C. Ohshima, H. Masago, G. Pescitelli, A. Wada, N. Berova, Appl. Spectr. 2005, 59 , 121–125. 93. L. Geng, L. B. McGown, Anal. Chem. 1992, 64 , 68–74. 94. J. L. Lunkley, D. Shirotani, Y. Yamanari, S. Kaizaki, G. Muller, J. Am. Chem. Soc 2008, 130 , 13814–13815. 95. R. A. Poole, C. P. Montgomery, E. J. New, A. Congreve, D. Parker, M. Botta, Org. Biomol. Chem. 2007, 5 , 2055–2062. 96. M. Benedetti, P. Biscarini, A. Brillante, High Pressure Res. 2000, 18 , 285–289. 97. M. Benedetti, P. Biscarini, A. Brillante, E. Castiglioni, Enantiomer 1999, 4 , 63–66. 98. M. Benedetti, P. Biscarini, A. Brillante, Physica B 1999, 265 , 203–207. 99. T. Harada, T. Sato, R. Kuroda, Chem. Phys. Lett. 2008, 456 , 268–271. 100. K. Ikeda, W. Liu, Y. R. Shen, H. Uekusa, Y. Ohashi, S. Koshihara, J. Chem. Phys. 2005, 122 , 141103. 101. K. Ikeda, W. Liu, Y. R. Shen, H. Uekusa, Y. Ohashi, S. Koshihara, Chem. Phys. Lett. 2006, 422 , 267–270.

287

8 CIRCULAR DICHROISM OF DYNAMIC SYSTEMS: SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY Angela Mammana, Gregory T. Carroll, and Ben L. Feringa

8.1. INTRODUCTION Chiral systems that undergo controlled dynamic processes including switching the conformation of molecules and the self-assembly of complex structures permeate molecular biology [1]. Studying these phenomena in model synthetic or semisynthetic systems holds great promise in gaining a better understanding of complex and highly organized biological systems. The creation, amplification, and control of chirality [2] is a fundamental issue in chemical biology [3]. Hence, exploring some of the basic pillars regarding dynamic chiral systems at the molecular and supramolecular level, particularly those that undergo conformational switching upon given particular stimuli, provides a strong groundwork by which to develop paradigms to guide forays into the hinterlands of biomolecular and materials sciences. Refining our knowledge regarding conformationally switching molecules and assemblies may provide guiding principles for gaining a deeper understanding of enzyme processes, molecular recognition and self-assembly, the mechanisms behind biological molecular machines [4, 5] and possibly clues regarding the speculative areas of the origins of homochirality and prebiotic structures [3]. Additionally, the study of switchable molecules and assemblies has the potential to provide knowledge innovation applicable to developing new molecular level technologies related to information storage, transport, and optics [6–8]. In this chapter we will explore molecular systems that undergo changes in chiral conformation and configuration upon thermal, chemical, photochemical, or mechanical stimuli. A preview of the kinds of systems we will examine is presented in Figure 8.1. CD spectroscopy is a method “par excellence” to study dynamic chiral systems at different hierarchical levels ranging from molecules to the supramolecular and macromolecular scale. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

289

290

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

(c)

Figure 8.1. Dynamic chirality at the molecular and supramolecular level detected by CD spectroscopy. (a) A chiral molecule can direct achiral molecules to self-assemble into chiral supramolecular structures. (b) A chiral molecular switch or motor undergoes conformational changes that include inversion of molecular helicity. (c) Chiral molecules can self-assemble into chiral supramolecular structures, the chirality of which is determined by the enantiomer in excess. (See insert for color representation of the figure.)

Our survey will emphasize supramolecular systems [9–12]. Chemists have made considerable strides in the ability to create and control covalent bond formation in molecules with a high degree of efficiency, selectivity, and control of chirality, enabling the creation of numerous unique structures. Although the number of reactions that chemists have developed greatly exceeds that utilized by nature, the natural molecular world, constructed over billions of years of evolution, is vastly more complex than synthetic systems. Nature’s qualitative hegemony can be attributed to its mastery of supramolecular chemistry. Supramolecular structures rely on the formation of intermolecular bonds through a variety of interactions including electrostatic, dispersive, hydrophobic, hydrogen bonding, π -stacking, adsorption, or simply entrapment. A variety of molecular spaces that can host a guest molecule and form a supramolecular complex include cyclodextrins, zeolites, buckminsterfullerenes, DNA, micelles, various aggregates, and even transient solvent cages [13]. Our understanding of the intermolecular (noncovalent) bond is much less advanced in comparison with covalent bond formation. Biology has utilized molecular recognition and self-assembly to develop the most complex molecular systems known. As chemistry moves forward in the twenty-first century, a more precise understanding of intermolecular interactions will enable greater control over the properties of self-assembled materials. Supramolecular structures are comprised of two or more molecules forming weak bonds. The noncovalent nature of the interaction allows for greater flexibility of the molecular constituents in comparison with covalent structures; however, polyvalent interactions allow for the stabilization of supramolecular species through the cooperation of many weak interactions [14]. Akin to the advancement of traditional organic and inorganic chemistry, much research has been performed in order to gain an understanding of the dynamics, conformations, and

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

stability of supramolecular structures. Among the various aspects of molecular chemistry that has a supramolecular analogue, chirality is of fundamental importance, particularly in regard to chemical biology. The same principles regarding chirality at the molecular level can be conceptually extended to the supramolecular level. A supramolecular system is chiral if the noncovalent units comprising the system are arranged in an asymmetric manner wherein its mirror image is nonsuperimposable, even if the constituents comprising the structure are achiral. A number of systems have been reported that demonstrate the formation and amplification of supramolecular chirality from achiral molecules. Two current principles regarding chirality at the supramolecular level include the sergeant-andsoldiers [15, 16] and majority-rules [17] effects. In brief, a sergeant-and-soldiers effect involves a small amount of chiral material that enforces a chiral structure on an assembly composed predominantly of achiral molecules which is dictated by the chirality of the sergeant. The majority-rules effect states that in a chiral but nonracemic assembly of two enantiomers, the one in the greatest amount will dictate the chirality of the system. CD spectroscopy provides an invaluable tool in elucidating the underlying themes of molecular and supramolecular chiroptical switching [18]. We will focus on some specific examples in which CD spectroscopy is used to interrogate molecules and assemblies that undergo reversible changes in chirality.

8.2. THERMAL SYSTEMS Modification of chirality through thermal processes is a well-known phenomenon, the most widely studied systems being biomacromolecules that lose their helical secondary structure upon melting [19, 20]. Similarly, synthetic systems show thermoresponsive supramolecular chirality. An organogel-based chiroptical system that reversibly assembles achiral porphyrin molecules into a chiral supramolecular assembly by a thermally controlled aggregation/deaggregation process was realized through the coassembly of glutamic diamide gelators, L1 or D1, and a tetra-alkyl-substituted porphyrin, TPPOC12 H25 (Figure 8.2) [21]. A very well-known property of many porphyrins is the ability to self-assemble under certain conditions, giving rise to two possible types of aggregates: edge-to-edge (J-aggregates) and face-to-face (H-aggregates). TPPOC12 H25 itself does not gelate; however, co-mixing with the gelator in DMSO followed by cooling the sample resulted in a supramolecular gel showing a positive exciton-type CD signal with a crossover at 443 nm and a positive Cotton effect with a maximum at 402 nm, in accordance with the Soret band absorption of J- and H-aggregates, respectively. Using a gelator of opposite chirality resulted in the reverse CD signal. The minimum ratio of gelator to porphyrin that displayed an ICD was 30:1 gelator:porphyrin. As more gelator was added, the signal was increased until a maximum was achieved at a ratio of 150:1. The induced chirality was attributed to assembling the porphyrin in a chiral environment. When porphyrins without long alkyl chains were used (i.e., TPPOH and TPPMe), an ICD could not be obtained, indicating that the cooperative assembly of the porphyrin is necessary for transmitting chirality. Furthermore, when TPPOC12 H25 and L1 were co-gelled in toluene, no ICD was obtained. The UV–vis spectra indicated that the porphyrin does not aggregate in toluene, whereas in DMSO both J- and H-aggregates were formed, indicating that aggregation of the porphyrin is essential for its chiral coassembly with the gelator. Upon melting the gel the CD signal was lost, but was regained upon cooling. The cycle could be repeated several times without a loss of signal.

291

292

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OR

N

NH RO

OR 80

N

OR R=H R = CH3 R = C12H25 H37C18

NH

(TPPOH) (TPPOMe) (TPPOC12H25) HN

C18H37 O

O O

Mol. Ellip. (104)

HN

Gel

40 0

Solution

–40 –80 350

400

450 500 550 Wavelength (nm)

600

NH O L-1 or D-1

Figure 8.2. A supramolecular chiroptical switch comprising achiral porphyrins (top left) that form gels upon co-assembly with a chiral gelator (bottom left). The chiral assembly can be switched on and off by adjusting the temperature as confirmed by CD spectroscopy. On the right, CD spectra of a mixture of achiral porphyrins (90 μM) and gelator (13 mM) in DMSO are shown. At high temperature the compounds do not gel and a CD signal is not obtained. Cooling the solution results in the formation of a chiral gel. The gel displays an exciton-type CD signal with a crossover at 443 nm and a positive Cotton effect at 402 nm, corresponding to the Soret band of J- and H-aggregates, respectively. (Reproduced by permission of The Royal Society of Chemistry [21].)

Temperature does not always show a simple relationship to the formation or disappearance of chiral structures. In order to understand the effect of temperature on the chiral amplification of dynamic supramolecular polymers, a simple and well-studied building block, trialkylbenzene-1,3,5-tricarboxamide, which assembles by triple-hydrogen bonding, was employed (Figure 8.3) [22]. Both sergeant-and-soldiers and majority-rules experiments were performed. A fast dynamic equilibrium exists between monomers and hydrogen-bonded stacks, allowing for modified self-assembled structures to form within one minute of mixing external components into the solution. Sergeant-and-soldiers experiments confirmed that (S )-TABTC could induce chirality in aggregates of achiral TABTC [23]. Similarly, the system was shown to follow the majority-rules principle. As the opposite (S )-TABTC enantiomer is mixed into the solution of the (R)-TABTC, the CD spectrum weakens. At 0% ee the CD disappears. As the ee is further increased to 40%, the intensity of the CD signal reaches a maximum. The authors studied the temperature dependence of the CD intensity at 223 nm as a function of fraction of sergeant (Figure 8.4a) and as a function of enantiomeric excess (Figure 8.4b). Similarly, the temperature dependence of the net helicity as a function of fraction of sergeant and enantiomeric excess was studied (Figures 8.4C and 8.4D). The phenomena were quantified by

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

Figure 8.3. On the top left is presented the structure of a discotic molecule based on benzene1,3,5-tricarboxamide (TABTC). The R substituents can be achiral (TABTC) or chiral [(R)-TABTC or (S)-TABTC]. When R is chiral, self-assembly results in chiral supramolecular polymers. On the right a proposed right-handed supramolecular helix is presented (note that the side chains were replaced by methyl groups for clarity). (Reproduced by permission of the American Chemical Society [22].)

calculating the free energy penalties associated with helix reversal (HRP) in a stack and the introduction of a chiral monomer into a stack of its unpreferred helicity (mismatch penalty, MMP). For both sergeant-and-soldiers and majority-rules experiments, the HRP is associated with the disruption of three hydrogen bonds. In contrast, the MMP has a different physical meaning for the two types of experiments. In the sergeant-and-soldiers experiment, the MMP corresponds to the incorporation of a chiral sergeant into a stack of achiral molecules of its unpreferred helicity. In the majority-rules experiment the MMP is associated with the incorporation of one chiral enantiomer into the helix formed by the opposite enantiomer. The strength of the noncovalent interactions decreases with temperature; however, even at elevated temperatures long stacks composed of 100 monomers were predicted for this cooperative self-assembling system. The HRP was found to change very little with temperature; however, the MMP was found to decrease due to a slight increase of the intermolecular distance which reduces unfavorable steric interactions. The effect of temperature on the MMP value explains why the degree of chiral amplification is reduced for the sergeant-and-soldiers system while it is enhanced for the majority-rules system. A lower MMP reduces the authority of the sergeant in the former case. In the latter case a lower MMP makes it easier for the minor enantiomer to join the helix dictated by the major enantiomer. Consequently, for the sergeant-and-soldiers experiment, upon increasing temperature a higher fraction of sergeant is required in order to obtain a homochiral system. On the contrary, for the majority-rules experiment, a lower ee is required in order to obtain a homochiral system.

293

294

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10°C 20°C 40°C 50°C

20

10

(b) CD (mdeg)

CD (mdeg)

(a) 30

0

10°C 20°C 40°C 50°C

30 20 10 0

0.00

0.04

0.08

0.12

0.16

0.20

0.0

1 0.8 0.6 0.4 0.2 0 0.00

0.04

0.08

0.12

0.16

(d)

1

Net helicity (–)

(c) Net helicity (–)

Fraction of sergeant (–)

0.8

0.20

Fraction of sergeant (–)

0.2 0.4 0.6 0.8 Enantiomeric excess (–)

1.0

0.6 0.4 0.2 0 0.0

10°C 20°C 40°C 50°C

0.2 0.4 0.6 0.8 Enantiomeric excess (–)

1.0

Figure 8.4. Effect of temperature on the CD signal on chiral supramolecular assemblies of benzene-1,3,5-tricarboxamide monomers containing either varying amounts of a chiral sergeant that induces chirality into the assembly of achiral soldiers or a mixture of two enantiomers at various ratios [22]. (a) CD intensity at 223 nm as a function of fraction of sergeant at 10◦ C, 20◦ C, 40◦ C, and 50◦ C. (b) CD intensity at 223 nm as a function of enantiomeric excess at 10◦ C, 20◦ C, 40◦ C, and 50◦ C. (c) Net helicity as a function of fraction of sergeant at 10, 20, 40 and 50◦ C. (d) Net helicity as a function of enantiomeric excess at 10◦ C, 20◦ C, 40◦ C, and 50◦ C. (Reproduced by permission of the American Chemical Society [22].)

8.3. PHOTOACTIVE SYSTEMS Systems that can be addressed by light offer many advantages over systems requiring thermal or chemical stimuli. Light provides a clean, traceless, and noninvasive reagent that leaves behind no byproducts. Second, photochemical reactions are localized to the chromophoric functionalities that absorb at the wavelength employed, allowing for highly specific transformations to occur within a molecule [24]. Finally, the use of light allows for spatially directed transformations of materials through the use of well-established photolithographic techniques [25]. Many photoactive overcrowded alkenes have been shown to contain helical structures that give rise to CD signals [26]. A particularly interesting class of overcrowded alkenes behave as molecular rotary motors [27]; that is, one-half of the molecule can undergo continuous 360◦ unidirectional rotation relative to the other half. Changes in the helicity of the motor during the stages of the rotary process make CD spectroscopy an invaluable tool in characterizing the rotary motion. The overcrowded alkene in Figure 8.5, (P , P )-trans1, contains two aromatic halves connected by a photoisomerizable double bond [28]. Each half also contains a stereogenic center in a pseudoaxial configuration, which is crucial for obtaining unidirectional rotation. The helical nature of the molecule is reflected in the sign

295

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

+400 Meax

≥ 280 nm

Meeq

≥ 380 nm

+200

Meeq

(P, P)-trans-1

(M, M)-cis-2 20°C

60°C

Meeq ≥ 380 nm Meax Meeq

Δε (1mol–1 cm–1)

Meax

0

≥ 280 nm Meax –400

(M, M)-trans-1

(a) (b) (c) (d)

–200

(P, P)-cis-2

220

240 λ (nm)

260

280

Figure 8.5. Rotary cycle of a light-driven molecular motor. A combination of photochemical and thermal isomerizations results in a net 360◦ rotation. Each of the four isomers has a distinct P- or M-helicity and a unique CD signal [28]. The CD spectrum of the initial motor at the start of the cycle, (P, P)-trans-1, is shown on the right (a). Upon absorption of a photon, the motor undergoes trans–cis isomerization to form (M,M)-cis-2, which displays an inversion of the CD signal at 217 nm (b). Thermal isomerization generates (P, P)-cis-2, which shows an inversion of the CD signal at 217 nm (c) in comparison with (M, M)-cis-2. A second photochemical isomerization inverts the CD signal at 217 nm and forms (M, M)-trans-1 (d). The original CD spectrum (a) is restored upon thermal isomerization to generate the starting conformation, (P, P)-trans-1.

of the CD spectrum at 217 nm (Figure 8.5). Upon absorption of a photon, the molecular motor undergoes trans–cis isomerization to form (M , M )-cis-2. CD spectroscopy reveals that the resulting cis motor has inverted helicity relative to the trans. The photochemical isomerization is accompanied by a change of the stereogenic center to a pseudoequatorial orientation, which is more unstable than the original pseudoaxial orientation. The motor undergoes a thermodynamically favorable thermal isomerization and forms (P , P )-cis2, which restores the more stable pseudoaxial conformation and inverts the helicity as reflected in the CD spectrum. The large free energy change provided by the thermal isomerization results in an irreversible conformational change, preventing the motor from rotating in the opposite direction. A second photochemical isomerization followed by thermal isomerization regenerates the initial CD spectrum, indicating the initial stage of the rotary cycle. A key challenge in applying the motor to exert nanomechanical-like forces is to demonstrate the rotational motion of the motor while anchored to a macroscopic surface such as a solid substrate (Figure 8.6). Appending the motor to a solid macroscopic surface puts limitations on the characterization options compared to the traditional solution-phase measurements because both (a) a smaller quantity of molecules is present (typically approximately 1014 molecules per cm2 in monolayers [11]) and (b) the analytical methods are limited to techniques that are applicable to the solid state. Although the use of

296

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

S

Rotor Axle Stator O

Legs

O O

O

n

Au Surface

S

n S

2 CD (mdeg)

hν

hν

0 –2 200

240

280 λ (nm)

320

Figure 8.6. Assembly of thiol-terminated light-driven rotary molecular motors on a semitransparent gold film provides a monolayer of chiroptical material that can be analyzed with CD spectroscopy. The CD signals invert between positive and negative bands, corresponding to changes in the helicity of the molecules comprising the monolayer upon the application of photons and thermal energy. The initial spectrum (solid black) inverts (dotted black) after irradiation with UV light (λmax = 365 nm) at room temperature. After heating the surface (70◦ C, 2 h) the spectrum inverts again to restore the original (solid gray). A second dosage of photons inverts the signal (dotted gray). Heating brings the rotors back to the original orientation relative to the substrate [30]. (See insert for color representation of the figure.)

CD spectroscopy to analyze monomolecular layers of organic molecules is rare [29], it provided an invaluable tool in characterizing the rotary motion of molecular motors attached to a semitransparent gold film [30]. In order to use CD spectroscopy, only a very thin layer of gold could be used in order to minimize the optical absorbance of the system. Therefore, 5 nm of gold was deposited onto both sides of an aminosilane-coated quartz substrate. Despite the low amount of material present in monolayer systems, the motor provides a strong enough CD signal to demonstrate rotary motion, as evidenced by the inversion of the CD spectra upon given photochemical and thermal stimuli (Figure 8.6). In addition to providing evidence that the motor can access the four stages of the rotary cycle, CD spectroscopy was used to uncover the proper length of spacer required to minimize quenching of the photochemical isomerization by the gold film. When spacers of eight atomic units were used, the CD signals did not change sign upon irradiation; however, when the motor contained spacers of 16 atoms, the gold-mounted chromophore was able to undergo photoinduced isomerization followed by thermal helix inversion. In addition to controlling the helicity of the chromophore, the overcrowded alkene can also exert intramolecular control over the helicity of a polymer. The stages of the rotary process have been shown to influence the twist sense of helical poly(hexyl isocyanate) (PHIC) when attached to the end-terminus of the macromolecule (Figure 8.7) [31]. Poly(isocyanates) (PIC) are stiff helical polymers that exist as a racemic mixture of P - and M -helices in the absence of an asymmetric influence [32, 33]. X-ray studies ˚ A similar on poly(n-butyl isocyanate) have revealed an 8/3 helix with a pitch of 5.14 A. helical structure is maintained in solution [32, 34, 35]. Chiral perturbations can favor the

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

Figure 8.7. A single light-driven molecular motor attached to the end-terminus of poly(hexyl isocyanate) (PHIC) is used to control the twist sense of the polymer. The trans isomer of the motor end-group exerts no chiral induction, allowing an equal probability for the P and M helices to form. Photochemical isomerization to the cis form induces a preferred handedness to the polymer backbone. Thermal isomerization inverts the handedness. Restoring the motor to its original trans form brings the polymer solution back to a racemic mixture with no preferred twist sense. (Reprinted by permission of John Wiley & Sons, Inc. [31].)

presence of one helical twist over the other. For example, when a PIC contains chiral repeat units, the two diastereomeric helices, which have the same stereogenic configuration in the repeat units, can have a different energy, making one helix more favorable over the other. Many studies have shown that PICs containing chiral repeat units give rise to CD spectra that indicate that the backbone assumes a preferred helicity [36–38]. A rather striking example is the preference for one helicity from the subtle asymmetric presence of a deuterium in place of a hydrogen atom [33, 35]. The resulting PHIC is optically active and displays a CD spectrum with a band at approximately 250 nm in the region where the recurring amide groups of the backbone absorb. The preference for one helicity is attributed to cooperative effects among the repeat units which amplify a slight energetic preference for one helix. The use of a chiral solvent can also bias the P - or M -helicity [39]. The preparation of PHIC with the presence of a molecular motor at the end-terminus of the macromolecule provides an example in which the helical twist sense can be reversibly controlled [31]. A molecular motor containing a benzamide functionality in the lower half was used to initiate the polymerization of n-hexyl isocyanate from the benzamide’s sodium salt. The use of a molecular motor as the initiator results in a PHIC with a chiral environment at the α-chain end of the polymer that can influence the twist-sense of the macromolecule for certain stages of the rotary cycle in which the helical end-group can interact with the repeat units. In the trans conformation the upper naphthalene shows no detectable interaction with the polymer backbone, and so the helix has no bias (Figure 8.8). No detectable excess of P - or M -helicity is revealed by CD spectroscopy because the signal of the motor-PHIC polymerized from an enantiomerically pure motor matches the signal of the enantiomerically pure motor alone. Photoisomerization of the motor to the cis configuration results in an increase in the intensity of the CD signal and is attributed to an induced preference for one helicity of the polymer. The induced helicity can be attributed to the encroaching naphthalene rotor, which imposes a chiral environment on the nearest repeat unit and favors one helicity over the other. Cooperative interactions along the chain would then amplify the chirality, resulting in a preferred helical twist of the polymer. After thermal

297

298

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

step 1

step 2

polymer: R N PIM = 50:50 Bz

365 nm

stable (2'S)-(M)-trans-2-PHIC (R = PHIC) stable (2'S)-(M)-trans-3 (R = COCH3)

R polymer: N excess M helicity Bz

unstable (2'S)-(P)-cis-2-PHIC (R = PHIC) unstable (2'S)-(P)-cis-3 (R = COCH3)

R polymer: N excess P helicity Bz stable (2'S)-(M)-cis-2-PHIC (R = PHIC) stable (2'S)-(M)-cis-3 (R = COCH3)

(b)

(a)

(c)

4

4

2

2

2

0 –2 –4

θ (mdeg)

4 θ (mdeg)

θ (mdeg)

Δ

hn

0 –2

–4

–4 220

280 λ (nm)

340

0 –2

220

280 λ (nm)

340

220

280 λ (nm)

340

Motor-polymer Motor (control)

Figure 8.8. The helicity of motor-terminated poly (hexyl isocyanate) (PHIC) can be followed with CD spectroscopy (Et2 O, −20◦ C). The CD spectra of the motor without PHIC (R = COCH3 ) and the motor attached to PHIC (R = PHIC) are shown. (a) In the trans form of the motor the CD spectra of the motor and motor polymer are the same, reflecting the lack of preferred helicity in the polymer backbone. (b) Photochemical isomerization (UV lamp, λmax = 365 nm) to the cis form induces a preferred handedness to the polymer helix, which is reflected in an increase in the intensity of the CD signal relative to the motor alone. (c) Thermal isomerization (20◦ C, 30 min) inverts the CD spectrum, indicative of a helix inversion. Reprinted by permission of John Wiley & Sons, Inc. [31].

isomerization the CD spectrum of the motor-PHIC is inverted and more intense than the motor alone. The change in the CD implies that the upper-half rotor remains close enough to the polymer backbone to maintain influence over the helical twist; however, the rotor is now on the opposite side of the polymer and induces the reverse twist. Photochemical and thermal isomerization bring the motor back to the starting state in which the polymer has no preferred handedness. The changes in the helicity of both the motor and polymer can be followed with CD spectroscopy (Figure 8.8). Recent follow-up experiments show that these kinds of systems can be used to control the pitch of cholesteric liquid crystals [40] and form interesting toroidal morphologies when dried on a solid substrate [41]. Similarly, incorporating photoisomerizable molecules into the backbone of a poly(isocyanate) can reversibly switch the helical sense of the polymer. The azobenzene chromophore, which can be reversibly photochemically switched between cis and trans isomers (and thermally isomerized from the cis to trans form), has a large body of literature related to its use as a photoswitch [42, 43]. A PIC containing chiral azobenzene pendant groups in the repeat units, AzoPIC, was shown to induce a preferred helicity when the chromophore was in the trans conformation (Figure 8.9) [44]. Photoisomerization to the cis isomer resulted in a reversal of the CD spectrum as shown in Figure 8.9. The dashed curve represents the polymer after irradiation and displays the mirror image of the spectra of the polymer before irradiation, indicating that the direction of the helicity was reversed. It was found that in order to induce a preferred helical twist the chiral group must be linked to the same phenyl ring of the azobenzene

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

Figure 8.9. A poly(isocyanate) (PIC) copolymer with repeat units that contain an azobenzene pendant group bearing a chiral chain displays a CD signal (THF, 0.5 mg/mL). Photoisomerization of the azobenzene unit reverses the sign of the CD signal. (Reproduced by permission of the American Chemical Society [44].)

as the isocyanate backbone. Variation of the location of the stereogenic center allowed the dominant direction of the helicity to be controlled. A variety of azobenzene PICs were synthesized, and their response to photoisomerization was found to depend on the stereogenic center [45]. The various systems synthesized showed different behavior in response to irradiation. Isomerization from the trans to the cis isomer in some polymers was shown to lead to an increase in the chiral interaction, detected by an increase in the CD spectrum, while in other polymers a decrease in the interaction was deduced based on an attenuation of the CD signal. Additionally, strong changes in the optical rotation were observed for some of the polymers during photoisomerization. Photoisomerization of azobenzene chromophores has been shown to affect the chirality of other helical systems as well. For example, a foldamer, a small oligomer that adopts a secondary structure, was designed by incorporating azobenzene into the core of an oligo(meta-phenylene ethynylene) derivative [46]. In the cis form of the azobenzene the foldamer can obtain a helical conformation showing a bisignate CD signal (Figure 8.10). Photoisomerization to the trans form denatures the helix because the coordinates of the foldamer components compromise the propensity to form a stable helical structure. An attenuation of the CD signal upon thermal isomerization from the trans to the cis form was attributed to depletion of the helical conformation through unfolding. Regeneration of the cis form by photo- or thermal isomerization restores the initial helical structure and is accompanied by the regrowth of the initial CD signal. Photoswitchable molecules have also been shown to modify the conformation of proteins in a reversible manner [47]. Such changes can be monitored by CD spectroscopy. For example, spiropyran modified poly(l-glutamate) (SPPGA) (Figure 8.11) has been shown to be effective in inducing such changes [48]. Spiropyran can exist as a neutral spiro form or as a zwitterionic merocyanine form [49]. The two states contain a considerably different geometry and polarity. SPPGA was shown to undergo conformational changes upon switching between the two forms. When the peptide-appended switch is in the merocyanine form (MEPGA), the polypeptide assumes a random coil conformation. Photoisomerization to SPPGA allows the polypeptide to undergo a transition to an α-helix. In hexafluoro-2-propanol, the spiro form thermally converts to the merocyanine form with a half-life of 2.5 h at 25◦ C, regenerating the disordered conformation. The

299

300

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CO2Tg

Tg

O

O

O

O

O

O

O

O

OMe

H 6

O N N

O

O

O

O

O

O

OMe

H 6

20

CO2Tg helix

θ (mdeg)

10 0 −10

hν, Δ

hν

−20 250

325 λ (nm)

400

Random coil

Figure 8.10. A photoswitchable foldamer composed of an azobenzene core bearing two oligo(meta phenylene ethynylene) pendant groups. The cis form of the switch can be accessed by irradiation with UV light (λmax ∼ 365 nm). The CD spectrum of the cis form is shown (acetonitrile, 5.6 × 10−6 M) and has been attributed to the foldamer obtaining a helical structure. Thermal isomerization to the trans form disrupts the chiral structure and shows a decrease in the CD signal. (Reprinted by permission of John Wiley & Sons, Inc. [46].)

changes in the conformation of the polypeptide manifest as changes in the CD spectrum. The characteristic negative CD signal of an α-helix with minima at 208 and 222 nm is generated and increases during irradiation. Subsequently, the negative CD signal becomes smaller in magnitude as SPPGA converts to MEPGA in the dark. UV–vis and fluorescence measurements indicate that the merocyanine form of the switch dimerizes, which is the proposed driving force for the distortion of the structure. Dithienylethenes (DET) provide another example of an often exploited photoswitch (Figure 8.12) [50]. The dithienylethene unit can undergo reversible photochemical ringopening and closure reactions upon absorption of a photon of the appropriate wavelength. The two forms of the switch absorb at sufficiently different wavelengths to allow for a particular state to be selected through the use of conventional lamps emitting UV and visible light. The open form of the switch absorbs in the UV region of the electromagnetic spectrum, while the closed form of the switch absorbs in both the UV and visible region. DET-1o can be reversibly photochemically closed to form DET-1c, and reopened to form DET-1o without fatigue for at least five cycles, and possesses a thermal stability that prevents the ring-opening and closure reactions from occurring in the dark. In the open form the switch exists as a dynamic structure that rapidly interconverts between a P - and M -helicity. Photochemical ring closure locks the switch in equal amounts of the RR and SS enantiomers; however, the chirality can be controlled by photochemical switching in a chiral environment as will be discussed below.

301

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

NO2

dark N H (CH2)2 O CO (CH2)2 N H

CH

O

NO2

N+ (CH2)2 O CO

hν COOH

(CH2) 2

(CH2)2

N H

C H

N H

CO

SPPGA

CH

N H

–O

COOH (CH2)2 C H

CO

MEPGA

–20

–210

230 λ (nm)

light

–10 dark

[Θ]·10–3

0

250

Figure 8.11. A spiropyran-modified PGA (poly(L-glutamate)), SPPGA, undergoes a ring opening reaction in the dark to the merocyanine form of the dye, MEPGA. Formation of MEPGA is accompanied by a distortion of the helical structure of the peptide (lower right) which is reflected in a change in the CD spectrum (lower left). Top curve (positive signal), solid line: MEPGA in hexafluoro-2-propanol before irradiation. Bottom curve (negative signal), solid line: SPPGA is generated after irradiation with sunlight. Intermediate spectra, dashed lines: thermal isomerization from SPPGA to MEPGA over a time period of 8 h. (Reproduced by permission of the American Chemical Society [47, 48].)

A dithienylethene photochromic unit functionalized with (R)-1-phenylethylaminederived amides (Figure 8.12) self-assembles into supramolecular structures through hydrogen-bond formation [51]. The stereogenic center causes the assembly to form a helical fiber. At room temperature, DET-1o forms a gel in organic solvents such as toluene, benzene, and hexane. Utilizing photochemical processes that control the molecular conformation of the switch and using thermal processes that control the macroscopic aggregation, it is possible to realize four supramolecular chiroptical states of the gel (Figure 8.13). The aggregated molecules form helical fibers and show a CD band at approximately 320 nm (Figure 8.14). Only the negative half of the exciton band is shown because the positive half is obscured by the solvent, which in this case is toluene. The CD signal of the aggregate is attributed to locking of a selected molecular helicity, M or P , of the open form of the switch when it is confined in the self-assembled structure. The enantiomeric (S )-1-phenylethylamine-derived DET-1o forms an aggregate that shows the opposite CD signal. The chirality of the stereocenters at the hydrogen-bonding components of the molecule determines the selection of the helical conformation of the

302

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H N

S O

H H

S

O

DET-1o

Vis H N

S

UV

Vis

UV

H N

S

O

M – DET- 1o

P – DET- 1o

O (S,S) – DET- 1c

(R,R) – DET- 1c

DET-1c

Figure 8.12. A dynamically helical photoswitchable dithienylethene chromophore containing chiral amides can be photochemically switched between an open, DET-1o, and closed, DET-1c, form. The open form of the switch interconverts between two helical conformations. The helicity of the open form during photochemical ring-closure determines the stereochemistry of the two stereogenic centers on the photoswitchable unit of the molecule. The amides allow the molecule to self-assemble into chiroptical fibers [51].

Sol 1

Gel (α) 1

Gel (β) 1 Vis

Vis UV Gel (α) 2 (PSS)

UV

Δ

Sol 2 (PSS)

Gel (β) 2 (PSS)

Figure 8.13. Scheme of the aggregation and switching processes of a DET-1 gel (1–4 mM, toluene), which can access four chiroptical states. Upon cooling an isotropic solution of the open form of DET-1 (Sol 1), a stable Gel (α) 1, with a P-helicity, is obtained. It is possible to reversibly close and open the central ring of DET-1, cycling between the stable Gel (α) 1 and the metastable Gel (α) 2 (PSS) with high diastereoselectivity (96% DE) and P-helicity, by irradiating with UV light (313 nm) and visible light (> 420 nm), respectively. Note that during the photochemical step the helicity of the gel is preserved; however, the photochemical ring-opening or ring-closure changes the rigidity and chirality (fixed or dynamic) of the central unit and as a consequence the stability of the chiral aggregate. Heating of the metastable Gel (α) 2 (PSS) leads to an isotropic solution of 2 [sol 2 (PSS)], which, upon cooling, results in stable Gel (β) 2 (PSS) with M-helicity. The thermal processes (50◦ C required to fully melt gel) irreversibly convert the gel from a metastable to a stable aggregate with an inversion of helicity. Again, the photochemical step is reversible and occurs with retention of the supramolecular chirality on going from the stable Gel (β) 2 (PSS) to the metastable Gel (β) 1, and vice versa. Finally, heating of the metastable Gel (β) 1 gives the isotropic solution of DET-1 (sol 1) closing the cycle shown [51].

core, M or P , during aggregation. The CD signal disappears when the aggregate is melted by heating to 50◦ C (Figure 8.15) and reaches a maximum at temperatures below 0◦ C. The attenuation of the CD bands correlates with changes in the molar fraction of DET-1o existing as a free monomer compared to the amount aggregated as measured by NMR. When completely dissolved, DET-1o photochemically cyclizes with no diastereomeric

303

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

–100

0

(mdeg)

(mdeg)

200

–200

0

–400

–100

–600

–200

300

350 400 Wavelength (nm) (a)

450

300

400 500 600 Wavelength (nm)

700

(b)

Figure 8.14. (a) CD spectra of stable gel DET-1o obtained by cooling a hot solution of DET-1o (1.8 mM) in toluene (solid line) and an unstable gel of DET-1o obtained by photochemically ringopening (λ > 420 nm) a stable gel of DET-1c (3.6 mM) (dotted line). The dashed line corresponds to DET-1o (0.35 mM) in solution and shows that when DET-1o is not aggregated, it does not produce a CD spectrum. (b) CD spectra of an unstable gel of DET-1c with 96% DE obtained by irradiating a stable gel of DET-1o (3.6 mM) (solid line), a stable gel of DET-1c with 96% DE obtained by heating (100◦ C) and cooling (0◦ C) the aforementioned unstable gel of DET-1c (dashed line) and a gel of DET-1c with no DE obtained by photochemical ring-closure (λ = 313 nm) of DET-1o and cooling (dash–dotted line) [51].

Figure 8.15. Temperature dependency (−15◦ C to 70◦ C) on the intensity of the CD band of DET1o at 320 nm. The extent of aggregation as probed by NMR and the %DE during photochemical cyclization of DET-1o to DET-1c decrease with increasing temperature in correlation with the attenuation of the CD signal [51].

excess (DE); however, in the gel form, DET-1o undergoes photochemical cyclization to DET-1c with 96% DE. The DE correlates with the intensity of the CD and the molar fraction of aggregated DET-1o. In the closed form the chirality of the methyl substituents at the photochromic core are locked. The aggregate of DET-1c can be photochemically ring-opened to regenerate the aggregate of DET-1o.

304

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Aggregation of a solution of DET-1c containing 0% DE forms a gel that gives a similar CD spectrum obtained upon photochemical ring closure of the aggregate of DET-1o. NMR reveals that the aggregate contains only one diastereomer of the closed switch. Interestingly, the diastereomer of DET-1c that gels is opposite to that formed upon stereoselective photocyclization of DET-1o in the aggregated state. Melting an aggregate of DET-1c formed by photochemical ring closure of an aggregate of DET1o, followed by cooling to regenerate the gel, results in an aggregate with the opposite helicity of that present after photochemical ring closure. The results suggest that in the open form, the chirality of the peripheral amide groups dictates the chirality of the selfassembled aggregate. However, in the closed form the chirality of the central photoactive part of the switch dictates the thermodynamically most stable helical conformation of the aggregate. Photochemical ring closure in the gel results in a metastable aggregate of DET-1c. Melting of the gel and re-cooling is necessary in order to access the more thermodynamically stable form of the supramolecular fiber. Similarly, when a stable aggregate of DET-1c is formed, photochemical ring opening forms a metastable aggregate of DET-1o. The CD signal of the metastable aggregate inverts upon successive heating and cooling. In addition to studies of switch DET-1, switch DET-2 (Figure 8.16), which differs from switch DET-1 by the presence of two methylene units between the outer phenyl rings and the stereogenic center, can also form chiral gels [52]. Interestingly, the gels formed by DET-2 have the opposite chirality of those formed by DET-1. The combination of fixed stereogenic centers in the hydrogen bonding units and a dynamic helicity that can be switched “on” and “off” allows for a four-state switching system in which molecular chirality and supramolecular chirality communicate. The self-assembling system described shows how the molecular and supramolecular chirality in a chemical system can influence each other. The self-assembling DET system described above was further shown to amplify chirality in aggregates of the isostructural, achiral switch DET-3o, shown in Figure 8.16 [52]. While both DET-1o and DET-2o can form chiral gels, DET-3o, which lacks a stereocenter, forms an achiral aggregate. Photochemical ring closure of the gel results in equal amounts of the RR and SS isomers. However, the chirality of DET-1o or DET-2o can induce DET-3o to aggregate in a chiral

H N O

H N n

H N

S

S

n

n

H N

S

O

O

H N

S

H N O

O

S

O

DET-1o n = 0 DET-2o n = 2

DET-3o (M)

DET-3o (P)

UV Vis

UV Vis

UV Vis

S

S

O O DET-1c n = 0 DET-2c n = 2

H N n

H N

S O

H N

S O

DET-3c (SS)

H N

S O

H N

S

H N

S O

DET-3c (RR)

Figure 8.16. Structure of photoresponsive organogelators for dynamic chiral selection and amplification. The compounds aggregate via hydrogen bond formation. Irradiation with UV (λmax = 313 nm) and visible (λ > 420 nm) light [52] produces photochemical ring-closure and ring-opening reactions. DET-3o lacks a stereocenter and forms achiral aggregates; however, in the presence of DET-1o or DET-2o, DET-3o aggregates in a chiral manner as shown in Figure 8.17 [52].

305

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

manner. When DET-3o is coassembled with DET-1o or DET-2o, a supramolecular chiral aggregate results, with the chirality transferring and propagating to achiral DET-3o. When DET-3o is coassembled with DET-1o or DET-2o, the resulting CD spectrum shows a more intense signal compared to DET-1o or DET-2o assembled alone (Figure 8.17), indicating that DET-3o coassembles with its chiral analogue and that the selected helicity of the chiral switch is imparted to the achiral switch. The chirality induced in DET-3o was locked by photochemical ring closure and shown to proceed with 94% ee of DET-3c. As the ratio of DET-3o increases, the intensity of the CD signal decreases and correlates with a decrease in ee. The transfer to and amplification of molecular chirality in DET3o by DET-1o or DET-2o is in line with previous sergeant-and-soldiers systems. The contrasting effects of DET-1o and DET-2o, which differ only in the presence or absence of two pairs of methylene units, show how subtle differences in molecular structure can influence both the resulting molecular and supramolecular properties of the system. The dithienyl systems described above utilize switches containing an inherent chirality. The resulting chiral supramolecular structures in turn influence the chirality of the molecules constituting the assembly. In order to initiate the chirality transfer, a stereogenic center needs to be present within the constituents. It is also possible to induce chirality in an achiral DET by using an external helical template. dsDNA provides a well-studied, chiral nanoscale building block that affords a high level of control over the structure [53, 54] and possesses charge-transport capabilities [55]. Its use as a scaffold for building new classes of organic materials merits attention [56]. Among the number of ways to functionalize dsDNA, electrostatic binding provides perhaps the simplest approach [57]. The construction of a DET switch bearing two primary amines provided a suitable candidate for binding studies (Figure 8.18) [58]. Molecular models show that the amines can access coordinates that closely match the locations of negative charge density on the DNA base pairs due to the phosphate groups. Similar to the DET switches described above, both DET-4o and DET-4c do not display CD spectra when dissolved in solution. However, addition of a poly(dGdC)2 2+3 150

CD (mdeg)

100

2

50 0 –50 –100

1 1+3

–150 300

320

340

360

380

400

Wavelength (nm)

Figure 8.17. CD spectra of chiral gels of DET-1o (1.3 mM, toluene) (solid line 1, negative band), DET-2o (1.2 mM, toluene) (solid line 2, positive band) and coassemblies of the achiral DET-3o (1 equiv.) with the respective chiral switch DET-1o (dotted line 1 + 3) or DET-2o (dotted line 2 + 3). In both cases the addition of DET-3o increases the CD intensity of gels containing DET-1o or DET-2o. (Reproduced by permission of the American Chemical Society [52].)

306

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

UV O

S

Vis

HN

NH H2N

O

O

S

DET-4o

NH2

S

S

H2N

+

O HN

NH

NH2

DET-4c

DET-4o

–

+ –

dG

dC

Figure 8.18. A photoswitchable chiroptical DNA complex. At the top is shown the photoequilibrium between the open (DET-4o) and closed (DET- 4c) forms of a dithienylethene molecular switch that contains pendant ammonium groups to confer water solubility and allow the switches to bind electrostatically to the polyanionic backbone of DNA when the amine is protonated. Photochemical ring closure was accomplished through the use of a UV lamp (λmax = 365 nm) using a 340-nm cutoff filter. Photochemical ring opening was performed with visible light using a 520-nm cutoff filter. Molecular models (created using Hyperchem®) show that the distance between the terminal ammonium functionalities closely resembles the distance between the anionic phosphate groups of a guanosine (G)–cytosine (c) base pair [58]. (See insert for color representation of the figure.)

oligonucleotide to a solution containing DET-4o or DET-4c results in an ICD corresponding to the switch (Figure 8.19). The chirality of the DNA double helix is transmitted to the orientation of the switches comprising the supramolecular complex. Additionally, the CD signal corresponding to the DNA attenuates, indicating that both components modify the structure of the other. The intensity of the CD grows until the amount of switch is approximately 79% the amount of base-pairs. The results are similar to studies regarding the binding of simple mono- and divalent cations to DNA, of which the charge compensation was found to be no larger than 85% [59]. Similar results were found for both the open and closed form. The titration experiments were used to calculate the binding constants of DET-4o and DET-4c, both of which are 2 × 105 . The CD measurements were complemented by UV–vis absorption measurements. The spectra of DET-4o and DET-4c show a hypochromic effect and a red shift. The strength of the binding could be controlled by adjusting the pH of the solution. As the pH was increased, the CD signal decayed, consistent with the hypothesis that the switch binds via electrostatic interactions between the ammonium groups of the switch and the negatively charged phosphate groups exposed at the outer surface of the DNA double helix. When the pH increases above 9.12, the CD signal corresponding to the switch is completely removed. The switches were also shown to bind to poly(dAdT)2 . Interestingly, the interaction showed enhanced chiroptical activity. As found for the poly(dGdC)2 , when poly(dAdT)2 is added to a solution of DET-4o or DET-4c the UV–vis absorption spectra undergo hypochromic effects and the CD spectra show ICDs corresponding to the isomer of the

307

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

– – – (a)

(c)

– – – – (b)

(d)

Figure 8.19. Changes in both the CD (a) and absorption (c) spectra of the dithienylethene (DET-4)–poly (dGdC)2 complex due to cyclically performing photochemical ring-opening and ring-closure reactions. All spectra were taken at room temperature in aqueous buffer at a pH of 6.5. Photochemical ring closure was accomplished through the use of a UV lamp (λmax = 365 nm) and a 340-nm cutoff filter. Photochemical ring opening was performed with visible light, using a visible light-emitting lamp equipped with a fiber optic and a 520-nm cutoff filter. The CD spectrum of the open DET- 4o- poly (dGdC)2 complex shows a clear ICD corresponding to the open form of the switch. Irradiation with visible light results in the attenuation of the band corresponding to the open DET- 4o and the growth of a signal in the visible region corresponding to the closed DET- 4c. Several cycles of photochemical switching can be performed as indicated by the reversible changes in the CD signal (b) at 350 nm and the UV–vis signals (d) at 331 nm (DET4o) and 560 nm (DET- 4c) [58].

switch. The optimum ICD was obtained at a ratio of 1:1.7 switch:base pairs compared to a ratio of 1:1.3 for the poly(dGdC)2 complex. The ICD spectra for the poly(dAdT)2 appear more intense, especially when the bisignate signals for the closed form of the switch are compared. Also, because the signal for the DNA is slightly blue-shifted, a clear bisignate signal for the open form is visible. A more complicated relationship between the ratio of switch to DNA was observed. The intensity of the ICDs did not simply increase and plateau as DNA was added. Furthermore, for some ratios, unique ICD signals could be obtained, indicating that intermolecular interactions between the bound switches seem to play a role. The higher charge density associated with poly(dAdT)2 may allow for multiple orientations of the switch, the most stable of which depends on interactions between the bound switches. Regardless of the sequence used, both DET-4o and DET4c retain their photochemical switching abilities when complexed to the DNA. The chiroptical response can be modulated with photons as shown by the reversible changes in the CD spectra, allowing for a multitude of unique chiral states to be generated (Figure 8.19).

308

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8.4. CHEMICAL AND MECHANICAL SYSTEMS A porphyrin system displaying a dynamic memory in which supramolecular chirality can be reversibly stored and released provides an elegant example of a switchable molecular assembly that responds to chemical stimuli (Figure 8.20) [60]. Achiral porphyrins have been shown to form chiral aggregates in the presence of chiral noncovalent amino acid polymers. The use of cationic and anionic porphyrins permits hetero-aggregation to occur via electrostatic interactions of the oppositely charged porphyrins. The strength of the interactions allows the assembly of achiral molecules to maintain supramolecular chirality after removal of the asymmetric template. By utilizing porphyrins containing

–

–

SO3

O3S

–

N

N

NHHN

N–H H–N

N

N

O3S

–

H2TPPS anionic porphyrin

SO3

N

N

NH

+4H+

N

–4H+

N HH N

+H N

N H+ TpyP4+

H2TpyP neutral porphyrin

route a

+

HN

N

+

–H+

+

N

N

H6 cationic porphyrin

+H+

monomers nonchiral aggregate

route b

chiral aggregate

+H+

cationic porphyrins anionic porphyrins neutral porphyrins

monomers

chiral seeds chiral aggregate

Figure 8.20. On the top are shown the structures of, respectively: the anionic meso-tetrakis (4-sulfonatophenyl) porphyrin (H2 TPPS), the neutral meso-tetrakis (4-pyridyl) porphyrin (H2 TpyP), and its protonated form, the cationic (H6 TpyP)4+ . On the bottom the system is described schematically. Cationic and anionic porphyrins form a chiral aggregate in the presence of one enantiomer of the chiral template phenyl alanine (either L or D). After removal of the template, the aggregate maintains chirality. The aggregate can be disassembled by deprotonation. Two routes of disassembly can be hypothesized. In route a, the aggregate fully disassembles, leaving only monomers in solution. Reprotonation would result in an achiral aggregate. In route b, the aggregate disassembles; however, the disassembly is not complete due to the slow rate of deprotonation and leaves small chiral aggregates in solution. Upon reprotonation, the small chiral aggregates can be thought of as chiral ‘‘seeds’’ that direct the regrowth of the template in a chiral manner. (Reproduced by permission of the American Chemical Society [60].)

309

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

ionizable groups in the meso positions, a system for which it is possible to imprint, store, release, and restore chirality was developed. The porphyrins employed contain four protonated pyridine groups in the meso positions. Deprotonation results in the loss of the cationic charge and consequent disassembly of the chiral aggregate. The CD signal of the aggregate disappears upon increasing the pH (Figure 8.21). The deprotonation step is kinetically slow, allowing for chiral seeds to remain in solution, provided that 24 hours have not elapsed. The chiral seeds are very efficient templates for their own selfpropagation. Reprotonation therefore leads to the regrowth of the supramolecular chiral assembly, which displays a CD signal matching the signal of the initial chiral aggregate. The two states can be cyclically addressed even in the absence of the chiral template. When a template containing the opposite chirality is employed, the CD spectrum shows the mirror image. The above concept was extended to anionic porphyrin J-aggregates templated from  or -[Ru(phen)3 ]2+ [61]. It is known that H2 TPPS4 is zwitterionic at a pH lower than 3 and in the presence of a millimolar concentration of salt. Under these conditions H2 TPPS4 self-assembles to give both H- and J-aggregates, displaying absorption maxima at 422 and 490 nm, respectively. The authors showed that J-aggregates formed in the presence of one enantiomer of the title ruthenium complex can maintain memory of the chirality imprinted at pH 6 and can be switched on and off by simply modulating the pH. The chiroptical

15 a1,2,3

10 L

5

b1,2

0

0

–40

–5 D –10 –15 350

450 λ (nm)

a 0.0 370 390 410 430 450 λ (nm) (b)

–80

a1,2,3

400

b

0.8

40 Δε

CD (mdeg)

80

Abs (a.u.)

1.6

500

550

(a)

Figure 8.21. (a) CD spectroscopy is used to detect the disappearance and reappearance of the chiral porphyrin aggregate. The sign of the CD signal is dependent on the chirality of the template that is used to initially imprint the chirality. CD spectra of the erasure and restoration of the chiral aggregate upon changing the pH are shown. Aggregates were made using either L- or D-phenyl alanine templates at a pH of 2.3 and show CD spectra that are mirror images. Increasing the pH to 9 results in the disappearance of the CD signal (curve a1 becomes b1 for either the L- or Reducing the pH back to 2.3 restores the signal (a2 ). The cycle can be continued to

D-aggregate).

give b2 (no signal) and a3 (restored signal). At least 10 cycles can be performed. (b) The absorption spectra of the aggregate at pH 2.3 (a) and at pH 9 in which the aggregate disassembles to predominantly monomer (b). (Reproduced by permission of the American Chemical Society [60].)

310

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

response could be controlled by cycling the pH between 2.5 and 6 of a solution containing chiral J-aggregates of H4 TPPS4 templated from -[Ru(phen)3 ]2+ in the presence of an excess of the -enantiomer of the complex. Upon raising the pH to approximately 6, the ICD of the J-aggregate disappears. Decreasing the pH back to 2.5 restores the CD signal with the same sign of the exciton couplet as that of the starting J-aggregate. As in the previous example, the propensity to reproduce the chiral conformation of the original aggregate during regrowth despite the introduction of a template that possesses the opposite chirality of the original is due to the remarkable inertness of these chiral aggregates and consequently to the presence of chiral seeds that have a stronger driving force to reconstruct the memorized supramolecular architecture compared to the ability of the template of opposite chirality to direct the converse structure. A thin film based on the layer-by-layer (LBL) assembly of DNA and poly(allylamine hydrochloride) (PAH) was shown to induce chirality in tetrakis(N -methylpyridinium-4yl)porphine upon its addition to the preexisting film (Figure 8.22) [62]. The induced chirality was evident by the appearance of a bisignate CD signal in the Soret band region with positive and negative Cotton effects at 423 and 445 nm, respectively, and a crossover at 432 nm. The authors attribute the Cotton effects to intercalation into the DNA and electrostatic binding [63]. A negative Cotton effect at 268 nm was assigned to DNA. Differences between the DNA CD band in the film compared to solution were attributed to polymer-salt-induced aggregation during assembly with PAH. The induced

Θ (mdeg)

12

(a)

0

–12

NH3 and H2O

HCI

–24 HCI Absorbance

(b) 0.6 As-prepared film (ICD)

A (no ICD)

B (ICD)

0.3 TMPyP 0.0 200

300

400 500 Wavelength (nm)

600

Protonated TMPyP

700

Figure 8.22. A chiral switch based on a layer-by-layer assembled DNA/poly (allylamine hydrochloride) film containing tetrakis (N-methylpyridinium-4-yl) porphine (TMPyP) additives is shown. The interaction between the dye and DNA is characterized by the appearance of an induced CD (ICD) signal corresponding to the main absorption band (Soret band) of the porphine. On the right is schematically presented the initial film (As-prepared film) that shows an ICD signal; the film after exposure to HCl gas (a), which breaks the interaction between the DNA and dye (TMPyP) and consequently cancels the ICD signal; the film after exposure to ammonium followed by H2 O (B), which restores the interaction and the ICD signal. On the left are shown the CD (a) and UV–vis (b) spectra of the films exposed to the same conditions described in the scheme on the right: signal of the film as prepared (dashed line); after exposure to HCl gas (dotted line); after subsequent exposure to NH3 gas (dashed–dotted line); and finally after exposure to water vapor (solid line). (Reproduced by permission of the American Chemical Society [62].)

311

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

chirality could be switched off by exposure to HCl, which can protonate both the DNA and dye [62]. The loss of the CD signal was attributed to deintercalation of the dye due to possible electrostatic repulsions and changes in the structure of the DNA. Loss of the ICD was accompanied by a change in the color of the film from yellow to green. Exposure to ammonium gas restores the yellow color of the film; however, the ICD was not recovered until the film was subsequently exposed to water. A pH-switchable DNA complex was designed that utilizes an achiral naphthalene derivative (P) bearing a diaminopurine hydrogen bonding unit that can bind to an oligothymine template (Figure 8.23) [64]. Upon mixing a ssDNA template with the diaminopurine, the UV–vis spectrum underwent a blue shift and hypochromic effect. The CD spectrum of the complex showed a positive Cotton effect in the region where naphthalene absorbs, with a zero-point crossing at 338 nm. The authors deduced that the naphthalene guests are arranged in a right-handed helix [65]. When the pH is decreased from 9 to 2, the Cotton effect reverses, suggesting that a left-handed helix is formed. The Cotton effect begins to reverse at pH 5. A pH titration revealed that the pKa of P is 4.8, which is in the range of the value at which the helix inversion occurs, suggesting that protonation of the purine induces a rearrangement in the complex that results in an inversion of the helicity. P

Tn

H H N

O N

H N H

HO

O

O

O

N

OH

O OPO O

N N O

O

O

O

O n-1H

OH

O N N H O O

N O

O N

H

75 pH = 9

CD (mdeg)

50 25 0 –25 pH = 2

–50 –75 200

250

300

350

400

450

500

λ (nm)

Figure 8.23. A naphthalene–diaminopurine derivative, P (top left), forms a dynamic, helical hydrogen bonding assembly with an oligothymine template, Tn (top right), where n is the number of residues. At higher pH values a right-handed helix forms. At lower pH values a left-handed helix forms. CD spectra (bottom) of the complex ([P] = 2[T]T40 = 0.5 mM) at 268 K at various pH values show that the handedness of the complex can be controlled by pH. As the pH is changed from 9 to 2, the sign of the CD bands invert. (Reprinted by permission of John Wiley & Sons, Inc. [64].)

312

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Amphiphilic thin films composed of achiral molecules have shown supramolecular chirality, possibly resulting from mechanical or other effects during film preparation (Figure 8.24) [66]. Regardless of the mechanism, the induced chirality can be controlled by chemical stimuli. 5-Octadecyloxy-2-(2-thiazolylazo)phenol (TARC18) amphiphile forms a chiral film at the air–water interface that gives rise to a CD signal (Figure 8.25). The origin of the chirality is thought to arise from a spontaneous overcrowded packing of the functional groups into a helical sense during compression of the film. Two different types of chiral films could be formed; however, no control over the selection of the type of chiral film could be achieved. FT-IR analysis indicated that the films differed in their orientation to the plane of the film as well as their trans/gauche conformation. Interestingly, the chirality of the film could be erased and restored upon exposure to HCl gas and air. Several cycles could be repeated before the film began to peel from the substrate. An explanation for the observed behavior is that exposure to HCl gas protonates the nitrogen of the thiazolyl group, changing the conjugation of the molecule that is accompanied by a change in the packing of the film to a state where the chirality is lost. Exposure to air restores the conjugation and allows the film to reform the original chiral arrangement. Although the initial direction of the chirality is selected at random, switching only occurs between the initial direction of chirality and an achiral structure. The opposite chirality is not obtained upon exposing the film to air after loss of chirality through exposure to HCl. Films composed of tetrakis(4-sulfonatophenyl)-porphine or an amphiphilic benzthiazolyl derivative also showed supramolecular chirality; however, these films could not restore the chiral structure after its removal with HCl [67, 68]. Interest in the relationship between macroscopic spinning motion and chirality can be traced back to Louis Pasteur’s attempts at controlling optical activity by performing chemical reactions in a centrifuge or growing plants while rotating in a given direction [2]. Although no induction or inversion of chirality could be detected, more recently the relationship between vortices or macroscopic rotation and the chirality of supramolecular assemblies has shown interesting results that may lead to new insights regarding the origins of homochirality in nature [69–72]. Several reports have dealt with this

Interface

Interface

Compression

M-Chiral

Compression

Achiral

P-Chiral

Figure 8.24. A supramolecular chiroptical switch composed of achiral amphiphiles. Space-filling structures of achiral amphiphile (TARC18), which forms a Langmuir–Schaefer film at the air–water interface, and chiral supramolecular structures formed upon interface compression. (Reprinted by permission of John Wiley & Sons, Inc. [66].) (See insert for color representation of the figure.)

313

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

0.6 b

OC18H37

12 a

a N

c

N

6 CD (mdeg)

Absorbance

OH

0.4

N S

0.2

c

0 –6

b

–12 0.0

300

400 500 Wavelength (nm)

600

(a)

300

400

500

600

Wavelength (nm) (b)

Figure 8.25. (a) Absorption and (b) CD spectra of 70 layer films on quartz or CaF2 substrates. The same spectra are obtained regardless of substrate used; however, the resulting films show one of two possible structures, both of which are formed by chance: (a) film I and (b) film II. The chirality can be removed by exposure to HCl gas: (c) film II after exposure to HCl gas. (Reprinted by permission of John Wiley & Sons, Inc. [66].)

challenging topic and include both static and dynamic systems. In the static systems the spinning sense of a solution results in the formation of one enantiomeric form of an assembly. For example, rotoevaporation of a solution of aggregating porphyrins showed a supramolecular chirality that was dependent on the direction of rotation of the spinning flask [73]. Similarly, spin-coated films of porphyrin–dendrimer wedges displayed a chirality dependent on the spinning direction of the substrate, while films prepared without spinning showed no chiral selection [69]. In dynamic systems the chirality of the system changes upon removal or reversal of the spinning perturbation. Stirring of noncovalent Jaggregates of protonated meso-tetrakis(4-sulfonatophenyl) porphyrin (H2 TPPS4) shows both static and dynamic induced asymmetry (Figure 8.26) [71]. The authors hypothesized that J-aggregates are inherently chiral and exist as racemic mixtures and their distribution can be influenced by chiral vortices. CD spectra taken of solutions stirred in a clockwise (CW) direction showed that  aggregates were preferentially formed. Reversing the direction of spinning to counterclockwise (CCW) showed a CD signal that indicates that the  aggregates are transformed to  aggregates. When the solutions are stirred for 24 h, a static induced chirality is realized. It was shown that in a mixture of J-aggregate enantiomers, the predominant supramolecular enantiomer deposits on the wall of the cuvette after stirring for 24 h. The monomers and minor enantiomer remain in solution. Applying a macroscopic vortex to a solution of J-aggregates results in the adhesion of one enantiomeric J-aggregate. Spinning in the opposite direction results in the deposition of the opposite enantiomer. The results indicate that stirring affects the distribution of enantiomers present in an initially racemic solution. Clockwise (CW) stirring prefers  aggregates, whereas counterclockwise (CCW) stirring prefers  aggregates. The effect of stirring in the presence of a chiral stimulus that favors the enantiomer that is opposite to that favored by the vortex was tested by the addition of a chiral ruthenium complex, [Ru(phen)3 ]2+ (vide supra), with a small enantiomeric excess of the -enantiomer. Stirring in a CW direction resulted in the deposition of -aggregates, despite the opposing driving force to form -aggregates. However, upon increasing the concentration of

314

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1000

600

JΛ

E

Λ

Δ

(A) No Stirring

CD

400 CD (mdeg)

JΔ

CCW CW

200

λ (nm)

λ (nm) JΛ JΔ

0

JΔ

E

–200

CD

800

(B)

E

JΛ (C)

–400 CW

–600

CCW

–800 –1000 460

470

480 490 λ (nm)

500

510

Figure 8.26. The meso-tetrakis(4-sulfonatophenyl)porphyrin (H2 TPPS4) (10 μM) in aqueous solution at pH = 3 and [NaCl] = 0.3 M forms J-aggregates. On the left are shown CD spectra of the J-aggregates recorded at different stirring conditions: without stirring (continuous line), clockwise (CW) stirring (dashed–dotted line), and counterclockwise (CCW) stirring (dotted line). On the right, energy diagrams for the three mentioned conditions are shown: (a) In the absence of stirring, neither enantiomer is favored; (b) CW stirring favors the -enantiomer as shown by the negative CD couplet shown in the small inset; (c) CCW stirring favors the -enantiomer as shown by the positive CD couplet shown in the small inset. (Reprinted by permission of John Wiley & Sons, Inc. [71].)

-[Ru(phen)3 ]2+ , the -aggregate was deposited when CW spinning was performed. The results show that the conformational fate of a supramolecular aggregate is dependent on a competition between macroscopic mechanical forces and molecular level chiral interactions.

REFERENCES 1. B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts, P. Walter, Molecular Biology of the Cell , 4th ed ., Garland Science, New York, 2003. 2. B. L. Feringa, R. A. van Delden, Angew. Chem. Int. Ed . 1999, 38, 3419–3438. 3. W. A. Bonner, Origins of Life and Evolution of Biospheres 1991, 21, 59–111. 4. H. C. Berg, R. A. Anderson, Nature 1973, 245, 380–382. 5. T. Kubori, N. Shimamoto, S. Yamaguchi, K. Namba, S. Aizawa, J. Mol. Biol . 1992, 226, 433–446. 6. B. L. Feringa, J. Org. Chem. 2007, 72, 6635–6652. 7. V. Balzani, A. Credi, F. M. Raymo, J. F. Stoddart, Angew. Chem. Int. Ed . 2000, 39, 3348–3391. 8. T. Verbiest, S. Van Elshocht, M. Kauranen, L. Hellemans, J. Snauwaert, C. Nuckolls, T. J. Katz, A. Persoons, Science 1998, 282, 913–915. 9. J. Elemans, A. E. Rowan, R. J. M. Nolte, J. Mater. Chem. 2003, 13, 2661–2670. 10. A. Ulman, Chem. Rev . 1996, 96, 1533–1554.

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

11. A. Ulman, An Introduction to Ultra-thin Organic Films: From Langmuir–Blodgett to SelfAssembly. Academic Press, Boston, 1991. 12. J. N. Israelachvili, Intermolecular and Surface Forces: With Applications to Colloidal and Biological Systems. Academic Press, New York, 1992. 13. N. J. Turro, Proc. Natl. Acad. Sci. USA 2005, 102, 10766–10770. 14. M. Mammen, S.-K. Chio, G. M. Whitesides, Angew. Chem. Int. Ed . 1998, 37, 2755–2794. 15. M. M. Green, M. P. Reidy, R. J. Johnson, G. Darling, D. J. Oleary, G. Willson, J. Am. Chem. Soc. 1989, 111, 6452–6454. 16. H. Gu, Y. Nakamura, T. Sato, A. Teramoto, M. M. Green, S. K. Jha, C. Andreola, M. P. Reidy, Macromolecules 1998, 31, 6362–6368. 17. M. M. Green, B. A. Garetz, B. Munoz, H. P. Chang, S. Hoke, R. G. Cooks, J. Am. Chem. Soc. 1995, 117, 4181–4182. 18. N. Berova, R. Koji Nakanishi, W. Woody, Circular Dichroism: Principles and Applications. 2nd ed., Wiley-VCH, Weinheim, 2000. 19. W. Saenger, Principles of Nucleic Acid Structure. Springer-Verlag: New York, 1984. 20. A. Mammana, G. Pescitelli, T. Asakawa, S. Jockusch, A.G. Petrovic, R.R Monaco, R. Purrello, N.J. Turro, K. Nakanishi, G.A. Ellestad, M. Balaz, N. Berova, Chem. Eur. J . 2009, 15, 11853–11866. 21. Y. G. Li, T. Y. Wang, M. H. Liu, Soft Matter 2007, 3, 1312–1317. 22. M. M. J. Smulders, I. A. W. Filot, J. M. A. Leenders, P. van der Schoot, A. R. A. Palmans, A. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2010, 132, 611–619. 23. L. Brunsveld, A. Schenning, M. A. C. Broeren, H. M. Janssen, J. Vekemans, E. W. Meijer, Chem. Lett. 2000, 292–293. 24. N. J. Turro, Modern Molecular Photochemistry. University Science Books: Sausalito, CA, 1991. 25. E. J. Park, G. T. Carroll, N. J. Turro, J. T. Koberstein, Soft Matter 2009, 5, 36–50. 26. B. L. Feringa, R. A. van Delden, N. Koumura, E. M. Geertsema, Chem. Rev . 2000, 100, 1789–1816. 27. B. L. Feringa, R. A. van Delden, M. K. J. ter Wiel, Pure Appl. Chem. 2003, 75, 563–575. 28. N. Koumura, R. W. Zijlstra, R. A. van Delden, N. Harada, B. L. Feringa, Nature 1999, 401, 152–155. 29. C. Nuckolls, T. J. Katz, T. Verbiest, S. Van Elshocht, H. G. Kuball, S. Kiesewalter, A. J. Lovinger, A. Persoons, J. Am. Chem. Soc. 1998, 120, 8656–8660. 30. G. T. Carroll, M. M. Pollard, R. v. Delden, B. L. Feringa, Chem. Sci . 2010, 1, 97–101. 31. D. Pijper, B. L. Feringa, Angew. Chem. Int. Ed . 2007, 46, 3693–3696. 32. A. J. Bur, L. J. Fetters, Chem. Rev . 1976, 76, 727–746. 33. M. M. Green, N. C. Peterson, T. Sato, A. Teramoto, R. Cook, S. Lifson, Science 1995, 268, 1860–1866. 34. U. Shmueli, W. Traub, K. Rosenhec, J. Polym. Sci. A 1969, 7, 515–524. 35. M. M. Green, C. Andreola, B. Munoz, M. P. Reidy, K. Zero, J. Am. Chem. Soc. 1988, 110, 4063–4065. 36. M. Goodman, S. C. Chen, Macromolecules 1970, 3, 398–402. 37. M. Goodman, S. C. Chen, Macromolecules 1971, 4, 625–629. 38. M. M. Green, J.-W. Park, T. Sato, A. Teramoto, S. Lifson, R. L. B. Selinger, J. V. Selinger, Angew. Chem., Int. Ed . 1999, 38, 3139–3154. 39. M. M. Green, C. Khatri, N. C. Peterson, J. Am. Chem. Soc. 1993, 115, 4941–4942. 40. D. Pijper, M. G. M. Jongejan, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2008, 130, 4541–4552.

315

316

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.

66. 67. 68. 69. 70. 71. 72. 73.

G. T. Carroll, M. Jongejan, D. Pijper, B. L. Feringa, Chem. Sci . 2010, 1, 469–472. A. Natansohn, P. Rochon, Chem. Rev . 2002, 102, 4139–4175. G. S. Kumar, D. C. Neckers, Chem. Rev . 1989, 89, 1915–1925. G. Maxein, R. Zentel, Macromolecules 1995, 28, 8438–8440. M. Muller, R. Zentel, Macromolecules 1996, 29, 1609–1617. A. Khan, C. Kaiser, S. Hecht, Angew. Chem. Int. Ed . 2006, 45, 1878–1881. O. Pieroni, A. Fissi, N. Angelini, F. Lenci, Acc. Chem. Res. 2001, 34, 9–17. F. Ciardelli, D. Fabbri, O. Pieroni, A. Fissi, J. Am. Chem. Soc. 1989, 111, 3470–3472. H. Duerr, H. Bouas-Laurent, Photochromism: Molecules and Systems, revised edition. Elsevier, Amsterdam, 2003. M. Irie, Chem. Rev . 2000, 100, 1685–1716. J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J. H. van Esch, B. L. Feringa, Science 2004, 304, 278–281. J. J. D. de Jong, T. D. Tiemersma-Wegman, J. H. van Esch, B. L. Feringa, J. Am. Chem. Soc. 2005, 127, 13804–13805. N. C. Seeman, Mol. Biotechnol . 2007, 37, 246–257. N. C. Seeman, Biochemistry 2003, 42, 7259–7269. C. J. Murphy, M. R. Arkin, Y. Jenkins, N. D. Ghatlia, S. H. Bossmann, N. J. Turro, J. K. Barton, Science 1993, 262, 1025–1029. N. C. Seeman, Chem. Biol . 2003, 10, 1151–1159. A. D’Urso, A. Mammana, M. Balaz, A. E. Holmes, N. Berova, R. Lauceri, R. Purrello, J. Am. Chem. Soc. 2009, 131, 2046–2047. A. Mammana, G. T. Carroll, J. Areephong, B. L. Feringa, J. Phys. Chem. B , 2011, 115 , 11581–11587. G. S. Manning, Q. Rev. Biophys. 1978, 11, 179–246. A. Mammana, A. D’Urso, R. Lauceri, R. Purrello, J. Am. Chem. Soc. 2007, 129, 8062–8063. R. Randazzo, A. Mammana, A. D’Urso, R. Lauceri, R. Purrello, Angew. Chem. Int. Ed . 2008, 47, 9879–9882. S. Jiang, M. Liu, Chem. Mater. 2004, 16, 3985–3987. J. Lang, M. Liu, J. Phys. Chem. B 1999, 103, 11393–11397. P. G. A. Janssen, A. Ruiz-Carretero, D. Gonzalez-Rodriguez, E. W. Meijer, A. P. H. J. Schenning, Angew. Chem. Int. Ed . 2009, 48, 8103–8106. P. G. A. Janssen, S. Jabbari-Farouji, M. Surin, X. Vila, J. C. Gielen, T. F. A. De Greef, M. R. J. Vos, P. H. H. Bomans, N. A. J. M. Sommerdijk, P. C. M. Christianen, P. Lecl`ere, R. Lazzaroni, P. D. Van Schoot, E. W. Meijer, A. P. H. J. Schenning, J. Am. Chem. Soc. 2009, 131, 1222–1231. P. Guo, L. Zhang, M. Liu, Adv. Mater. 2006, 18, 177–180. L. Zhang, Q. Lu, M. Liu, J. Phys. Chem. B 2003, 107, 2565–2569. P. Guo, M. Liu, Langmuir 2005, 21, 3410–3412. T. Yamaguchi, T. Kimura, H. Matsuda, T. Aida, Angew. Chem. Int. Ed . 2004, 43, 6350–6355. A. Tsuda, M. A. Alam, T. Harada, T. Yamaguchi, N. Ishii, T. Aida, Angew. Chem. Int. Ed . 2007, 46, 8198–8202. A. D’Urso, R. Randazzo, L. Lo Faro, R. Purrello, Angew. Chem. Int. Ed . 2010, 49, 108–112. M. Wolffs, S. J. George, Z. Tomovic, S. C. J. Meskers, A. P. H. J. Schenning, E. W. Meijer, Angew. Chem. Int. Ed . 2007, 46, 8203–8205. J. M. Rib`o, J. Crusats, F. Sagu`es, J. Claret, R. Rubires, Science 2001, 292, 2063–2066.

9 ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS Cheng Yang and Yoshihisa Inoue

9.1. INTRODUCTION The last decade has witnessed rapid progress in supramolecular chirality research [1]. Chiral phenomena in supramolecular systems, such as chirality induction, chiral association, recognition, memory and amplification, which primarily originate or develop from molecular chirality, are often much more complicated and therefore more challenging than those at the molecular level. Besides the significantly larger number of incorporated building blocks (atoms and molecules), supramolecular chiral structures integrated by noncovalent interactions, such as electrostatic, hydrogen bonding, and van der Waals, π –π stacking interactions, are free from the strict geometrical restrictions of covalent bonds dominating the molecule chirality. Therefore, chiral supramolecular structures are usually more flexible, diverse, and adjustable in geometry, leading to novel chiral phenomena and architectures beyond the limitations of individual molecules. Furthermore, supramolecular chirality is not necessarily based on molecular chirality, but can also be generated upon aggregation of achiral molecules. Elucidating such unconventional higher-order chiral phenomena is crucial for understanding a wide variety of chiral architectures in the nature. Biomolecules by themselves are chiral supramolecular systems, and most of the physical and chemical events occurring in biosystems, including enzyme catalysis, molecular recognition, pharmacological action, and helix formation in DNA and proteins, are intrinsically chiral and supramolecular. Studies on supramolecular chirality provide insights into the biological superstructures and functions, and they may shed light on the origin of biomolecular homochirality. Supramolecular chirality is also important in chemical and materials science and technology, and it finds practical

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

317

318

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

applications in asymmetric catalysis, chiral separation and sensing, data storage, optical devices, and liquid crystal displays. Electronic circular dichroism (ECD) is the most powerful and versatile tool among various spectroscopic techniques employed in the study of molecular and supramolecular chiral phenomena. ECD spectroscopy, relying on the differential absorption of rightand left-handed circularly polarized light, is indispensable in particular for the study of enantiomeric supramolecular systems and events that produce significant changes in chiroptical properties. Other spectroscopic methods, such as NMR, UV–vis, fluorescence, and IR spectroscopy, are achiral in nature and hence applied to the study of diastereomeric supramolecular systems. X-ray crystallography can directly provide the absolute structures of chiral molecules and molecular assemblies, but is applicable only to crystalline samples and not suitable for the observation of dynamic chiral behavior. In contrast, ECD is more widely applied not only to liquid samples but also to amorphous and crystalline samples and even to gaseous samples, and it can monitor the kinetic and dynamic chiral processes under a variety of conditions. Other chiroptical techniques that use circularly polarized light include optical rotatory dispersion (ORD) [2, 3], vibrational circular dichroism (VCD) [4], fluorescence-detected circular dichroism (FDCD) [5, 6], and circularly polarized luminescence (CPL) [7]. ECD is undoubtedly more widely applicable than these methods: ECD has obvious advantageous over ORD, especially for relatively weak transitions; VCD in the infrared region provides chiral information about the relevant covalent bonds but is less sensitive than ECD; FDCD and CPL are more sensitive than ECD but applicable only to fluorescent species. ECD plays an irreplaceable role in supramolecular chirality research as a highly sensitive tool for determining the chiral sense of the relative orientation of chromophores in a supramolecular system. Chiral chromophores are inherently CD-active and often exhibit significant changes in CD intensity upon aggregation or complexation with other molecules, enabling us to quantitatively investigate the kinetics, thermodynamics, and structural changes associated with such processes. Achiral chromophores incorporated in a chiral supramolecular environment may also show appreciable induced CD through extrachromophoric chiral perturbation upon noncovalent interactions with chiral entities in supramolecular system. ECD spectral studies can provide insights into the relative orientation of the relevant chromophore and chiral center and is therefore a highly efficient tool for analyzing complicated chiral supramolecular phenomena. The significance of ECD in supramolecular chirality research is underscored by the vast number of publications devoted to the application of ECD to supramolecular systems. In this chapter, we will review the ECD studies on chiral supramolecular systems mainly in solution phase, which are sorted by the chiral event or method in supramolecular system. Several important rules and interesting phenomena relevant to the ECD spectra of chiral supramolecular systems will also be introduced with examples.

9.2. CHIRALITY SENSING WITH ACHIRAL CHROMOPHORE Detecting the chiral sense of asymmetric molecule or molecular assembly with achiral chromophore through noncovalent interactions represents an important application of ECD in supramolecular chemistry [8, 9]. Achiral chromophore is CD-silent by itself but can function as a chirality sensor or reporter when situated in chiral environment. This strategy can be used for the determination of the absolute configuration of a chiral molecule or the enantiomeric excess (ee) of a chiral compound. The CD intensity induced

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

to achiral chromophore is not very strong in general, but can be significantly amplified by introducing two chromophores to the chiral system to cause an exciton coupling interaction. Such amplification enables sensing of chiral molecules that are barely CDactive or limited in quantity. Possible overlap with the inherent CD signals at shorter wavelengths may also be avoided by choosing an appropriate chromophore that absorbs at longer wavelengths. An ideal chirality-sensing host should bear a strong binding site(s) highly selective to the target chiral molecule and intensely absorbing chromophore(s) near the binding site(s). Hydrogen-bond and coordination are the most frequently employed interactions in the design of a sensing host, primarily due to their strong binding and directing properties. Other weak interactions, such as π –π , hydrophobic, and electrostatic interactions, may play major or supplementary roles under certain conditions. The exciton chirality method, originally proposed by Harada and Nakanishi, is the most frequently used method for analyzing supramolecular chirality by ECD [10, 11]. Exciton-coupling interaction of two transition moments leads to a splitting of absorption band (Davydov splitting) in UV–vis and ECD spectra. The coupling of two transitions arranged in P (plus, right-handed)-helicity gives a “positive” couplet, displaying a positive Cotton effect peak at a longer wavelength and a negative one at a shorter wavelength, and vice versa for the M -helicity. The couplet amplitude produced, being a function of the distance and angle of the two coupling transitions, is usually very strong and enables chirality sensing of a very small amount of chiral sample.

9.2.1. Chirality Sensing Through Hydrogen-Bonding Interaction Molecular recognition of free-rotating achiral biphenyl-2,2 -diols 1a–c with chiral diamines 2a–c (Figure 9.1) through hydrogen-bonding interaction was investigated by using ECD [12, 13]. Complexation of stereolabile 1 with enantiopure 2 hinders the free rotation about the interaromatic bond in 1 to achieve the point-to-axial chirality transfer from 2 to 1. A negative CD couplet centered at 324 nm was induced to the major transition band of 1a, which is assignable to the 1 La band of phenol units, upon complexation with (1R, 2R)-2a in toluene at a 1a/2a ratio of 2, while antipodal (1S , 2S )-2a gave a mirror-imaged spectrum (Figure 9.2). The CD amplitude was a critical function of the binding strength and the steric interactions between 1 and 2, and therefore it was maximized for the most bulky diamine 2a in toluene. However, the use of polar solvents, such as acetone, acetonitrile, and ethanol, significantly reduced the CD signal due to the weakening of the hydrogen bonds. Interestingly, the nature of the interaction between 1 and 2 is altered from hydrogenbonding to electrostatic by lowering the temperature. The proton transfer from 1a to 2a was confirmed in toluene at −80◦ C by the appearance of a phenolate absorption band at

Figure 9.1. Chirality sensing of (1R, 2R)-diaminocyclohexane 2a–c with biphenyl derivatives 1a–c.

319

320

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1.5

(1S,2S)-2a

Δε (M–1cm–1)

1.0 0.5 0.0 –0.5 –1.0 –1.5

(1R,2R)-2a 280

320

400 360 Wavelength (nm)

440

Figure 9.2. Circular dichroism spectra of the complexes of host 1a with guest (1R,2R)- and (1S,2S)2a in toluene at 25◦ C. [1a] = 6.83 × 10−5 M, [2a] = 1.37 × 10−4 M. (Reprinted with permission from reference 14. Copyright Royal Society of Chemistry.)

412 nm [14]. The phenolate band and the corresponding Cotton effect were observed only upon addition of an excess amount of 2a (1.2 equivalents or more). The stoichiometric study suggested that 1a and 2a form a 1:2 complex at low temperatures. The anisotropy factor for the 1:2 proton-transfer complex is one order of magnitude larger than that for the 1:1 hydrogen bonding complex, suggesting that the axial chirality is more effectively fixed in the 1:2 complex. Although a solution containing 1a and 2a in 1:1 ratio did not show any CD signal, addition of an equimolar amount of achiral diisopropylamine to this solution induced CD signals, intensity of which was half of that for a 1:2 complex of 1a with 2a. 1,8-Naphthyridine derivatives 3a,b (Figure 9.3), in which the naphthyridine moiety acts as a proton acceptor while the pyrrole or indole moiety as a proton donor, were prepared as sensors for detecting monosaccharides [15, 16]. Complexation of 3a,b with 4c–h in dichloromethane led to significant changes in UV–vis, CD and fluorescence

OH

HO O

O

HO HO

OC8H17 C8H17O

OH

OH OH

HO

4b

4a N

N 3a

NH

OH HO

OH

HN

O

HO HO

OH

4c

HO HO

O OC8H17

O-C8H17

4d

OH OH

NH

N 3b

HO

Me

O

O

N

OH

OC8H17

C8H17O

OH HO

4e

HN

HO HO

4f O

O OC8H17

OH

4g

OH

CH2OH

HO HO

OCH3 OH

4h

Figure 9.3. Chirality sensing of monosaccharides 4a–h with achiral naphthyridines 3a,b.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

spectra. The two pyrrole rings in 3 takes a dual inward conformation when bound to chair-form pyranoside. Complexation of 3b with octyl-β-d-glucopyranoside 4a induced a positive CD signal at 442 nm, a strong negative one at 344 nm, and a much weaker one at 317 nm, while antipodal octyl-β-l-glucopyranoside 4b gave the mirror-imaged CD. However, the conjugated chromophore in 3b, involving a naphthyridine and two indoles connected with ethynylene linkers, appears to hinder the unambiguous analysis of the observed CD signals due to the overlapped transitions. Nevertheless, these sensing hosts may be used as a tool for distinguishing the monosaccharide enantiomers.

9.2.2. Chirality Sensing through Coordination A large number of achiral porphyrin-based sensors have been developed, exploiting the strong coordination ability and the large extinction coefficients of the Soret band. Nakanishi and Berova employed achiral bis(zinc porphyrin)s linked with a flexible tether as chirality probes for enantiomeric diamines and aminoalcohols [17–24]. Figure 9.4 shows typical tweezer porphyrin 5, which binds a variety of chiral diamines and aminoalcohols, including 6a–n, to produce bisignate CD signals at the Soret band as a result of the exciton coupling interaction of the twisted porphyrin transitions. The sign of CD couplet is nicely correlated with the absolute configuration of guest 6a–h, except for l-lysinol 6f, which, however, gives a consistent result by acylating the hydroxyl to give 6g [17]. As shown in Figure 9.5, an extremely large amplitude of > 1000 cm−1 M−1 was observed upon complexation of 6l with 5, for which the 1:1 stoichiometry was confirmed by a Job plot [18]. The couplet intensity was related to the difference in size between the smallest and largest substituents at the stereogenic center. Recently, this methodology was theoretically investigated by computer simulation with the potential optimized for liquid simulation [22, 23]. These studies not only provide

Figure 9.4. Chiral diamines and aminoalcohols 6a–n examined by chirality sensor ZnZn-5.

321

322

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

16

435 nm + 578

400 200 Δε

ACD = + 1078

0 –200 –400 423 nm - 500

400

MCH

420 440 Wavelength (nm)

CD amplitude (mdeg)

(a)

(b)

12 8 4 0

0

0.2

0.4

0.6

0.8

1.0

Molar fraction of host 5

Figure 9.5. CD spectrum of tweezer complex of 6l (20 equiv.) with ZnZn-5 (1 μM) in methylcyclohexane; ACD denotes the amplitude of the CD exciton couplet. (b) Job plot of ACD as a function of the molar fraction of host 5; [5] + [6l] = 1 μM. (Reprinted with permission from reference 18. Copyright American Chemical Society.)

an important support for the tweezer approach but also offer crucial insights into the structural factors governing the complexation and the couplet amplitude. The computational method was found useful for predicting the absolute configuration of more complicated chiral compounds with multiple stereogenic centers, such as cis- and trans-3-hydroxyl4-aryl/alkyl-β-lactams 6m and 6n, indicating that the remote stereogenic center has only a secondary effect on the inter-porphyrin twist. Inoue, Borovkov, and co-workers [25–38] proposed the chirality sensing by achiral bis(metalloporphyrin) 7 (Figure 9.6), relying on the exciton chirality method. As illustrated in Figure 9.6, coordination of an enantiopure monodentate ligand (Figure 9.7) leads

Figure 9.6. Chirality sensing mechanism for bis(metalloporphyrin)s. (Reprinted with permission from reference 36. Copyright American Chemical Society.)

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.7. Mono- and diamines 8 and 9 examined with chirality sensor ZnZn-7.

to the conformational switching of bis(zinc octaethylporphyrin) ZnZn-7 from stacked syn to twisted anti [39]. Due to the steric repulsion between the ligand’s substituents and the ethyl substituents on the adjacent porphyrin, the anti -form thus produced is not linear but twisted to afford strong bisignate CD signals (Figure 9.8) at the Soret band [25–27]. The sign of the couplet is uniquely correlated with the absolute configuration of chiral ligand—that is, positive couplets for monodentate (S )-ligands and negative ones for (R)-ligands. Solvent does not significantly affect the CD spectra of ZnZn-7 complexes with simple amines 8a–d. However, the CD spectral behavior of the complexes with amino acid esters 8e–g strongly depended on the solvent polarity, affording a negative couplet in nonpolar solvents, in good agreement with the results for 8a–d, but a positive couplet in polar media [34], for which the increased effective size of the ester moiety by solvation would be responsible (Figure 9.9). The ECD spectrum is significantly temperature-dependent in this chirality sensing system, showing a gradual increase of the couplet amplitude at lower temperatures, due to the increased affinity of chiral alcohol and amine ligands. Other factors, such as the central metal and the binding stoichiometry, are also crucial in determining the sign and intensity of the induced ECD [35, 40]. MgMg-7, rather than ZnZn-7, exhibits a higher affinity to chiral alcohols to display a strong CD couplet even at room temperature [41]. Bis-porphyrin ZnZn-7 forms a 1:1 tweezer complex with suitable bidentate ligands, such as trans-1,2-(1R, 2R)-diaminocyclohexane 9a, exhibiting a bisignate CD signal. The tweezer complex is transformed to a 1:2 host–ligand complex in the presence of an excess amount of the ligand. The CD amplitude (ε) of the tweezer complex amounts to 500 M−1 cm−1 as a result of the rigid and optimized geometry, while the corresponding

323

324

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20

0

Δε (cm–1 M–1)

10

ε (105cm–1 M–1)

–10

–20

2

0 300

400

500 Wavelength (nm)

600

Figure 9.8. UV–vis and CD (inset) spectra of 7 in the absence (dotted lines) and presence of 8a (solid lines) and 8e (dashed lines) in CH2 Cl2 (black lines) and in cyclohexane. (Reprinted with permission from reference 34. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.).

1:2 complex still gives a huge, but significantly smaller, ε of 200 M−1 cm−1 , due to the more flexible and dynamic structure. Similar stoichiometry-dependent chiroptical behavior is seen upon chirality sensing by tweezer bis(Zn porphyrin)s 10a–c with longer linkers (Figure 9.10) [42, 43]. All of these tweezers show high affinities (>105 M−1 ) to trans-1,2-diaminocyclohexane (9a) upon 1:1 complexation to give strong bisignate CD signals. However, further addition of 9a led to a red shift of the Soret band and a dramatic decrease of CD intensity with sign inversion, indicating switching of the bis-porphyrin conformation from tweezer to open form upon 1:2 host–ligand complexation (Figure 9.11).

9.2.3. Chirality Sensing through Other Noncovalent and Covalent Interactions Supramolecular complexation through electrostatic, hydrophobic, and van der Waals interactions also induces ECD signals to achiral chromophores. Water-soluble achiral calix[n]arenes 11n (Figure 9.12) form host–guest complexes with chiral ammonium ions through electrostatic and cation–π interactions. Addition of (R)-12 to 114 produced a strong negative CD couplet (Figure 9.13) [44], which was interpreted by a twisted array of the benzene rings in 114 induced primarily by the largest naphthylethyl substituent in the chiral guest.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.9. Effects of strongly and weakly interacting solvents on the mechanism of supramolecular chirality induction in 7. (Reprinted with permission from reference 34. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.).

Figure 9.10. Chirality sensing with tweezer bis(Zn porphyrin)s 10a–c.

Boronic acids bind to organic 1,2- and 1,3-diols with high affinities through the reversible boronate formation. As shown in Figure 9.14, the diboronic acid-bearing 13 gives 1:1 adducts with several monosaccharides, including glucose, mannose, galactose, and talose, through the synergic formation of two boronate esters [45, 46]. Most of the

325

326

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

Figure 9.11. UV–vis (top) and CD (bottom) spectral changes of (a) 10a and (b) 10b in the Soret region upon titration with (1R, 2R)-9a (0, 1, 2, 10, 50, 100 equivalent). (Reprinted with permission from reference 42. Copyright Royal Society of Chemistry.).

Figure 9.12. Achiral calixarene 11 for chirality sensing of 12.

d-saccharides, except for d-galactose, give a negative CD couplet upon esterification with 13, while the l-form afforded a positive couplet. Rosini and co-workers [47] used 4-biphenylboronic acid 14 for sensing chiral 1-arylethane-1,2-diols via boronate ester formation. The conformationally rigid fivemembered boronate ring can fix the two aromatic groups in a well-defined orientation to produce a strong CD couplet. Upon esterification with 13, all of the examined (R)and (S )-1-arylethane-1,2-diols consistently exhibited negative and positive couplet, respectively.

9.3. CHIRAL CONFORMATION OF MACROMOLECULE Macromolecules are usually conformationally diverse in solution, and ECD is a crucial tool for elucidating the static properties and dynamic processes of macromolecules, in particular those with helical conformation. Intense ECD signals are often observed for macromolecules composed of chiral subunits. Macromolecules made from achiral monomers may also take chiral conformations upon interaction with external chiral effectors. Even if the target macromolecule has no UV–vis absorption, CD spectral techniques may be applied by covalently or noncovalently attaching a chromophore. ECD spectroscopy is an important tool for analyzing the structures of biopolymers, such as DNA and proteins. Readers should consult (a) the sections in Chapter 4 of the present volume

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

327

10000

[θ] (deg cm2dmol–1)

5000 114 – (S) –12 0

–5000 114 – (R) –12

Figure 9.13. Induced CD spectra of 114 in the presence of (R)- and (S)-12 in pH 7 phosphate

–10000

200

300 λ (nm)

400

buffer at 25◦ C. (Reprinted with permission from reference 44. Copyright Royal Society of Chemistry.)

Figure 9.14. Chirality sensing through boronate ester formation with chiral 1,2- and 1,3-diols.

2 and (b) relevant literatures for comprehensive information regarding the application of ECD to various biomolecules [48]. In this section, we will focus on the application of ECD to synthetic macromolecules, such as polymers and dendrimers.

9.3.1. Inherently Chiral Macromolecule Chiral polymer is obtained by polymerizing monomers with chiral substituent(s) or by inducing chirality in the main chain upon polymerization of achiral monomers under the influence of chiral initiator, additive, and so on. Copolymers of chiral and achiral monomers also show optical activities; and two intriguing phenomena, leading to the

328

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

15

[Θ] x 10–4 deg cm2 dmole–1

10

5

0

–5

–10

–15 180

200

220

240

260

280

300

λ (nm)

O

O N Δ

H H 3C

N

C H 3C H

(R)

(S)

O

O C

N

N

C Y

X H H3C

C

H 3C (R)

H

(S)

15 X

Y

O

49

51

×

56

44

Figure 9.15. CD spectra of polyisocyanate with chiral side chain of different enantiomeric ratio: R:S = 100 : 0 (triangle), 0:100 (square), 49:51 (circle), and 56:44 (cross). (Reprinted with permission from reference 49. Copyright American Chemical Society.)

“majority rule” and the “soldiers and sergeants principle,” have been discovered upon ECD measurement of mixed chiral or chiral–achiral copolymers. The majority rule was first experimentally demonstrated by Green et al. [49] in their study on the polymerization of isocyanates with chiral substituent. The polyisocyanates thus obtained formed helices in solution, handedness of which was controlled by the ee of the chiral side chain. They found that the CD intensity of chiral copolymer 15 (Figure 9.15) obtained from a mixture of (R)- and (S )-monomers in varying ratios is not proportional to the ee of the used monomer, but behaves nonlinearly against the ee.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.16. Chiral copolymers 16 that obey the 16

majority rule.

Thus, the CD spectrum of 2% (S )-rich copolymer (i.e., R/S ratio = 49 : 51) was identical in sign and shape to that of homopolymer of (S )-monomer, but the CD intensity was unexpectedly strong, amounting to one-third of the homopolymer (Figure 9.15). It was interpreted that the energy barrier for helix reversal is much higher than the chiral bias caused by the pendant group [50], and therefore the helical sense will be controlled by the configuration of the majority pendant configuration. Thus, the CD spectrum of 12% (R)-rich copolymer (i.e., R/S ratio = 56 : 44) became exactly the same as that of (R)-homopolymer, indicating that the helix reversal was completely suppressed at that ee. The “soldiers and sergeants” principle is another interesting phenomenon related to the chirality induction in copolymer of chiral and achiral monomers. According to the principle, a minute amount of chiral monomer incorporated in copolymer with achiral monomer can dominate the overall helical sense of the copolymer. This phenomenon was first observed by Green and Reidy in their study of the specific rotation and CD intensity of copolymer 16 (Figure 9.16) obtained from the copolymerization of chiral and achiral monomers in various ratios [51, 52] Copolymer 16 containing 4% chiral component (x = 0, y = 4, and z = 96) showed a specific rotation comparable to that of 100% chiral homopolymer (x = 0, y = 100, z = 0), while by incorporating 1% chiral unit (x = 1, y = 0, z = 99) 16 achieved half the specific rotation of homopolymer (x = 100, y = 0, z = 0). Strong CD signal was observed even at a chiral unit content as low as 0.12% (x = 0.12, y = 0, z = 99.88), demonstrating that a minute amount of the chiral unit embedded in a long sequence of achiral N -hexylisocyanate units can appreciably bias the equilibrium between the right- and left-handed helices. The chiroptical properties of the copolymer did not change even if the chiral unit content was reduced to 15%. These observations indicate that a very small number of chiral inductor units, acting as the sergeants, can command a much large number of achiral soldier units in the copolymer to align in one direction. Dendrimers are repeatedly branched, monodisperse, and usually highly symmetric spherical macromolecules. Its highly branched 3D structure features high surface functionality and versatility. Dendrimers often show good conformational cooperativity and can convey local structural or chiral information to the next hierarchical level of structural organization. Dendrons 17–19 (Figure 9.17) are moderately soluble in water and highly soluble in organic solvents [53]. An intense negative couplet was observed for dendron 17 in both THF and water (Figure 9.17), indicating an M -helical conformation of anthranilate components in the dendron. The second-generation dendron 18 showed a negative couplet in THF but a positive one in water, indicating an M -to-P helix transition

329

330

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Me O O

O

O NH CO2R

O

OMe CI

O

(a)

N

O O

NH O

OR

Me

CI

O

OR

17 O

P

N HN

O O

O

HN

NH

O

O

RO

O

Me O

O

OR

Me

Me 18

OR

RO RO

OR

Me

O

OR

O HN O

O

O

1 0 –1 –2

240 260 280 300 320 340 360 380 400 λ (nm)

O

O O

O

2.4

RO

O

O

N H

H N

O

O NH N

O HN O RO

N

N HN

O

O O

Me

NH O Me RO

O

O

N HN

CI

θ/(105deg cm2 mol–1)

HN

O

17 THF 17 H2O 18 THF 18 H2O

O

NH N

19

18 UV (THF) π π*

2

HN

N

H N

HN N

O

Me

O

NH

O

Me

O O

Me

Me

O

M

(b)

O

O

RO

N

RO

OR O

NH N Me

H2O

O O O HN N O OH

HN

NH

O

O

O

N

RO THF

N O O N H HO O N

θ/(105deg cm2 mol–1)

R:/

OR

O

NH

UV (THF) 1.6

π

π*

19 THF 19H2O

0.8 0 –0.8 –1.6

O O

Me

230 250 270 290 310 330 350 370 390

OR

λ (nm)

Figure 9.17. (a) Direction of the electronic transition moment of the π –π ∗ transition at 316 nm and the sign of the corresponding CD couplet. (b) CD spectra of dendrons 17-19 in H2 O and THF, normalized for the concentration and the number of chiral terminal groups. (Reprinted with permission from reference 53. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.).

in water. Similar M -to-P transition was observed with the third-generation dendron 19 upon going from THF to water. In good agreement, the experimental IR spectral studies and the computational simulations indicate an increase of the gauche C–C and anti C–O bonds of the poly(oxyethylene) chains on going from organic to aqueous solution. The solvent-induced ECD change reflects the conformational fluctuations of the terminal chains that are coupled with the dendron’s helical secondary structure through correlated dendron-chain motions.

9.3.2. Chirality Induction in Achiral Polymer Polymerization of achiral monomer may give polymers with helical segments, handedness of which is, however, randomly populated in the polymer chain to give no ECD signal on

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.18. Condensation polymerization of 21 with 22 in chiral liquid crystal 20.

Figure 9.19. Chirality induction in achiral polymer 24 by chiral alcohols 25 and 26.

the whole. Polymerization with chiral catalysts or in chiral environment can give chiral polymer efficiently [54–56]. Akagi and co-workers [57] reported copolymerization in chiral liquid crystal 20 of achiral monomers 21 and 22 (Figure 9.18). The nematic phase of 20 is stable over a wide temperature range and remains stable upon addition of 21 and 22. Copolymers 23 obtained by condensation in (+)-20 and (−)-20 showed CD spectra with positive and negative exciton couplet, respectively. However, the couplet of copolymer 23 seems to originate from the chiral aggregation rather than the helical structure of the main chain (axial chirality) [57]. Achiral polymer can adopt a chiral conformation upon noncovalent interaction with chiral molecules. Fujiki and co-workers [58] reported the induced CD for achiral polysilylene 24 upon complexation with chiral alcohols 25 and 26 (Figure 9.19). Notable negative or positive CD couplet was induced when 24 was mixed with (S )- or (R)-25, for which the hydrogen-bonding interaction between the hydroxyl group in 25 or 26 and the ether oxygen in 24 is responsible. Interestingly, the majority rule does not work in this system, because a good linear correlation was observed between the ee of 24 and the induced CD intensity, which, however, offers a possibility to quantitatively determine the optical purity of chiral alcohols. Sada and co-workers [59] investigated the aggregation of chiral bis(dioxazolylpyridinyl)porphyrin 27 with achiral poly(trimethylene iminium) 28 (Figure 9.20). The formation of double-stranded structure, in which bidentate ligand 27 bridges two polymer chains of 28, was revealed by AFM and UV titration, while the ECD spectra provided the information about the 3D structure and the handedness of the supramolecular chirality. In the absence of polymer 28, 27 exhibited an apparent positive CD couplet in the Soret region, which was attributed to the dipole coupling between the porphyrin and terminal dioxazolylpyridine transitions. Upon addition of 28 to 27, the positive couplet was inverted in sign and decreased in intensity. Such an inversion was not caused by adding monomeric 29. This confirms the formation of helical double-stranded structure, in which the porphyrin rings are twisted in order to avoid the steric hindrance between the isopropyl groups in 27.

331

332

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Helical aggregate of porphyrins

O N N

Hydrogen-bonding

H N NH

O

Achiral synthetic polymer

Hydrogen-bond pair

Double-stranded helix

Figure 9.20. Double-helical chiral aggregation of achiral polyamine 28 with bidentate chiral inductor 27. (Reprinted with permission from reference 59. Copyright 1998 American Chemical Society.)

30

31

Figure 9.21. Helical tubular structure formation of polythiophene 31 with schizophyllan 30.

ECD was used for detecting the assembling of water-soluble achiral polythiophene 31 with chiral polysaccharide, schizophyllan 30 (Figure 9.21) [60]. Upon addition of 30 to an aqueous solution of 31, both UV–vis and fluorescence peaks of polythiophene showed significant bathochromic shifts, reflecting the increased effective conjugation length of the polythiophene backbone. Formation of a tubular structure, in which polythiophene is included in the helical schizophyllan tube, was proposed based on the UV–vis titration and AFM measurement. An intense positive CD couplet observed in the polythiothene’s π –π * transition region indicates a right-handed twist of the backbone of 31 in the tubular structure.

333

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

9.3.3. Chiral Memory in Polymer Supramolecular chirality of conjugated polymers often exhibits unique dynamic features. Yashima and co-workers [61–78] investigated the chiral properties of a series of conjugated polymers (Figure 9.22). Achiral conjugated polymer 32, originally a mixture of at least four conformers of cis-transoid, cis-cisoid, trans-transoid , and trans-cisoid , forms a single dynamic helical structure upon complexation with chiral amines, showing a characteristic CD couplet (Figure 9.23) [62]. The amplitude of the couplet is augmented with increasing bulkiness of the chiral amine and also with decreasing distance between the chiral center and the amino group, indicating the importance of steric effect in the chirality transfer. One of the most intriguing phenomena associated with this conjugated polymer–chiral amine system is the memory of macromolecular chirality, which is preserved even after replacing the chiral amine attached to 32 with an achiral amine. As shown in Figure 9.23, addition of (R)-naphthaleneethylamine 35 to a solution of polymer 32 induces a negative CD couplet [64, 70]. Further addition of chiral amino alcohol (S )-36 to this solution leads to an inversion of the couplet sign from negative to positive. Surprisingly, the original negative couplet of [32·(R)-35] complex is not affected by the addition of achiral amine 37 and the subsequent removal of (R)-35 from the solution by gel permeation chromatography, and even upon further addition of (S )-36 [64]. The CD

Figure 9.22. Achiral conjugated polymers 32–34, which give helical conformation upon complexation with chiral amines.

334

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(b) (32-(R)-35 complex) + (S)-36

3

[θ] (104 degree cm2 dmol–1)

2 1 0 –1

Figure 9.23. CD spectra in DMSO of (a) [32·(R)-35] complex, (b) a mixture of [32·(R)-35] complex with (S)-36, (c) a mixture of [32·(R)-35] complex with 37,

(d) Fractionated 32

–2

and (d) 32 isolated by GPC with a DMSO eluent containing 37 (0.8 M); [32] = 3.0 mg/mL (20.4 mmol in monomer unit/mL) for traces a–c and 0.13 mg/mL

(c) (32-(R)-35 complex) + 37 –3

(a) (32-(R)-35 complex) 310

350

400

450

Wavelength (nm)

500

550

for trace d. (Reprinted with permission from Macmillan Publishers Ltd. [64], copyright 1999.)

signal lasts for a long period of time to show only a 5% decrease in intensity after 3 months. The memory efficiency of the macromolecular helicity of 32 preserved by amino alcohols is a critical function of the number of methylene groups in amino alcohol but almost independent of its affinity to the carboxyl group. Similarly, intense CD was induced to achiral polymer 33 by keeping its solution containing chiral aminoalcohol at 50◦ C for 29 days. Once the helical polymer structure was induced, the helicity was memorized even after removal of the chiral amine and only about 10% decrease in CD intensity was observed after 29 days without any assistance of achiral amine [70, 78, 79]. Polymer 34 with bulky aza-18-crown-6 pendants forms a single helical structure upon complexation with chiral amines 39–41 (Figure 9.22), displaying a characteristic split-type CD at the absorption band of the backbone of 34.[76] Achiral cyanine dye 38 is entrapped in polymer 34 to form J aggregates, which afford CD signals when L-39 is coincluded. Interestingly, the supramolecular chirality of the J-aggregates was maintained even after the chirality of the polymer backbone was inverted by adding D-39.

9.3.4. Chiral Molecular Recognition with Polymer Inouye and co-workers [80] studied the CD spectral detection of saccharides with achiral water-soluble polymer 42 (Figure 9.24) in aqueous protic media. Meta-ethynylpyridine polymers 42 form helical structures in protic media through intramolecular solvophobic interactions. Hydroxyl groups of saccharide form a hydrogen bond to the pyridine nitrogens of 42 even in MeOH–H2 O mixture. In contrast to the weak CD signal induced upon complexation with d-glucose, 42 exhibits much stronger induced CD for octylβ-d-glucopyranoside. As shown in Figure 9.25a, the CD signal is inverted in sign by changing the solvent composition from MeOH/H2 O = 5 : 1 to 10:1, due to the varied ratio of d- and l-glucose in solutions of different composition. Furthermore, the CD spectra of 42 obtained immediately after the addition of d- and l-glucose are opposite in sign but gradually reduced in intensity to eventually converge to a common spectrum

335

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

OR

complexation N

I

N

N

N

N

N

N

N

NOH

N OH

N mutarotation of glucose

N

N

42a: R = (C2H4O)8CH3 42b: R = n-C4H9

complexation

N

N

n

42

N

N

N

N

N OH

N N N

OH

left-handed helical complex

right-handed helical complex

Figure 9.24. Chiral self-aggregation of achiral polymer induced by a saccharide. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.) (See insert for color representation of the figure.)

3 MeOH / water

CD/(mdeg)

CD/(mdeg)

2

1

0

–1 300

8

5:1 6:1 7:1

8:1 9:1 10:1 310

320

330

340

350

360

6 β-glucose +7 +2.4 mdeg 4 (337 nm) 2

time 0h 1h 3h 8h 15 h 24, 48 h 15, 24, 48 h 8h

0

5h 3h α-glucose –2 –3 1h +2.4 mdeg 0h (337 nm) –4 300 310 320 330 340 350 360

λ (nm)

λ (nm)

(a)

(b)

Figure 9.25. (a) Induced CD spectra of a mixture of 42a (1 mM in monomer unit) and D-glucose (0.3 M) in 5:1–10:1 MeOH/H2 O at 25◦ C. (b) Time-dependent CD spectra of a mixture of 42a (1 mM in monomer unit) and α- or β-D-glucose (0.3 M) in 5:1 MeOH/H2 O at 25◦ C. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.) (See insert for color representation of the figure.)

after standing the solution at 25◦ C for 24 h (Figure 9.25b). These phenomena were rationalized in terms of the anomerization of d-glucose between the α- and β-forms in the protic solvent.

9.4. SUPRAMOLECULAR COMPLEXATION WITH CHIRAL MOLECULAR HOST ECD is a powerful tool for studying the supramolecular complexation behavior of inherently chiral host molecules, because achiral chromophoric guests often become CD-active when bound to a chiral host. The CD signals thus induced provide crucial information about the spatial arrangement of the chromophoric guest(s) included. On the other hand, the guest inclusion may also cause a change in host conformation, which can be reflected in the CD spectrum. Hence, ECD spectral study enables us not only to qualitatively detect the supramolecular complexation and guest orientation but also to quantitatively determine the binding stoichiometry and affinity.

336

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

9.4.1. Binding with Cyclodextrin Cyclodextrins, a family of cyclic oligosaccharides typically composed of 6–8 glucose units linked by α-(1 → 4) glycosidic bonds, are water-soluble truncated cone-shaped macrocyclic hosts with a hydrophobic cavity that can include a wide range of organic guests through hydrophobic interactions. Cyclodextrin is inherently chiral, and the complexation of chromophoric guest in its cavity induces appreciable CD signals in the guest’s absorption region. By using the ECD response as a tool for detecting supramolecular interaction, Tokura and co-workers [81] investigated the complexation behavior of benzoylbenzoates with cyclodextrins in mid-1970. Mason [82] predicted that the anisotropy (g) factor induced to an achiral chromophore that is oriented randomly to a chiral molecule will be much smaller (10−5 –10−6 ) than the one in fixed orientation (10−2 –10−3 ). Harata and Uedaira [83, 84] attempted to calculate the sign of ECD induced to a chromophoric guest complexed with cyclodextrin by using Kirkwood’s [85] oscillator theory. Later, Kajtar et al. [86] proposed empirical “sector rule” to correlate the sign of induced CD signal with the orientation of transition moment of a chromophoric guest complexed with cyclodextrin. According to the sector rule (Figure 9.26), a positive CD signal is induced when the transition moment of a chromophore is located in the conical sector along the cavity axis, while a negative one to a more slanted transition located outside the cone. Kodaka and others further discussed the correlation of induced CD with the location and orientation of a chromophore complexed by various cyclodextrins [87–91]. Thus, the “Kodaka rule” says that when a chromophoric guest is accommodated inside the cavity, a transition parallel to the cavity axis induces positive CD and a perpendicular transition causes negative CD, but exactly the opposite is true for a chromophore located outside the cavity. These rules have been well-confirmed by the experimental results and therefore used as a standard tool for analyzing or interpreting the orientation of chromophoric guest in and around the cyclodextrin cavity [92–98]. For instance, methyl orange 43 with an azobenzene chromophore (Figure 9.27) exhibits a positive CD upon inclusion by β-cyclodextrin with its π –π * transition being aligned along the cavity axis, whereas the azobenzene moiety in compound 44, being laterally positioned above the cyclodextrin portal, affords a positive CD for the π –π * transition [99]. Brinker and co-worker [94]

Figure 9.26. Sector rule for predicting the sign of induced CD upon complexation in cyclodextrin cavity.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.27. Methyl orange 43 azobenzene-bridged 43

cyclodextrin 44.

Δε (M–1cm–1)

44

λ (nm)

Figure 9.28. Induced CD and orientation of azi-adamantane 45 upon complexation with β- (top) and γ -cyclodextrin (bottom). (Reprinted with permission from reference 94. Copyright 1998 American Chemical Society.)

investigated the orientation of azi-adamantane 45 (Figure 9.28) included in the cavity of α-, β-, and γ -cyclodextrin by ECD. Complexation of 45 with β-cyclodextrin induced positive CD, while γ -cyclodextrin complex only afforded a much weaker negative CD (Figure 9.28). By taking into account the size and shape of 45, the azo chromophore is deduced to be located near the portal of β-cyclodextrin or inside the γ -cyclodextrin cavity, as illustrated in Figure 9.28, which induces the positive or negative CD, respectively, according to the sector rule applied to the π –π * transition which is perpendicular to the N=N bond. The weaker CD intensity observed with γ -cyclodextrin is ascribed to the higher guest mobility in the γ -cyclodextrin cavity. CD spectral titrations with 45 gave the association constants of 6150 and 2740 M−1 for β- and γ -cyclodextrin, respectively.

337

338

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4

HO

θ/(mdeg)

2

O

OH O OH HO

O HO O OH

0

O OH O HO

HO

–4

O

–6

HO

O

46

OH O OH

–2

OH OOH

O O

OH O OH MeO OH OH O O

OH

O

200

220 240 260 280 Wavelength (nm)

HO

Figure 9.29. CD spectral changes of m-methoxybenzyl-β-cyclodextrin (0.1 mM) upon addition of cyclooctene.

θ/(mdeg)

By appending a chromophore, cyclodextrin becomes CD-active. In aqueous solution, the chromophore appended to cyclodextrin is often included in its own cavity, but can be driven out of the cavity by adding an appropriate guest as a competitor. The CD spectral changes thus induced can be used as a measure of guest inclusion. As illustrated in Figure 9.29, 6-O-m-methylbenzoyl-β-cyclodextrin shows a negative CD at the 1 Lb band and a weak positive CD at the 1 La band, indicating shallow penetration of the benzoate moiety into the cavity. Addition of cyclooctene leads to a gradual increase of CD intensity at both the 1 Lb and 1 La bands, suggesting the change of orientation of the methylbenzoyl moiety caused by the guest inclusion. A quantitative CD spectral titration gives the association constant of 11,120 M−1 [100, 101]. γ -Cyclodextrin, possessing a cavity larger than its lower homologues, can accommodate two planar aromatic guests, which are usually stacked in a chiral fashion to produce a split ECD. Inoue and co-workers [97, 102–110] investigated the 1:2 host–guest complexation of anthracenecarboxylic acid (AC) with γ -cyclodextrin derivatives, before examining the asymmetric photocyclodimerization of AC. The intimate stacking of ACs in the cavity caused significant changes in UV–vis, NMR, fluorescence, and CD spectra. As shown in Figure 9.30, the complexation of AC with monoaltro-γ -cyclodextrin 47 led

Wavelength (nm)

Figure 9.30. CD spectral change upon addition of AC (0–0.3 mM) to the solution of monoaltroγ -cyclodextrin 47 (2 mM) in pH 9.0 aqueous buffer.

339

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

to a strong positive couplet at the 1 Bb band, indicating the P -helical arrangement of two ACs in the cavity.

9.4.2. Binding with Biomolecule DNA forms three types of duplexes with different helical structures (i.e., A-, B-, and Z-DNA) through multiple hydrogen-bonding and stacking interactions of A–T and G–C ˚ width × 8-A ˚ depth nucleobase pairs. B-DNA has the major and minor grooves of ∼13-A ˚ ˚ and 4.5-A width × 6-A depth, respectively. The major groove comprises more nucleobase substituents and phosphodiesters, while the minor groove is walled by the hydrophobic part of sugar. A guest molecule may be bound to the major or minor groove or intercalate in between two base pairs. The CD intensity induced to a groove-bound chromophore is thought to be one or two orders of magnitude greater than that of an intercalated one [111]. Because of the existence of multiple binding sites, the complexation behavior with DNA is sometimes complicated and a critical function of the guest concentration and the ionic strength [112, 113]. For example, acridine orange intercalates to DNA to give negative CD at low concentrations, but much stronger positive CD at higher concentrations due to groove binding. Furthermore, the intercalating and groove-binding ligands couple to each other to give a split CD. The CD signal induced to a chromophore upon interaction with DNA crucially depends on the DNA structure. As shown in Figure 9.31, addition of sulfonated Ni porphyrin 48 to right-handed B-DNA of poly(dG-dC)2 induces no appreciable CD signal at the Soret band (∼400 nm) of 48 [114]. However, once the B-DNA is converted to left-handed Z-form by adding spermine, an intense negative couplet emerges at the Soret band, enabling selective sensing of Z-DNA. The electrostatic interaction of negatively charged 48 with the protonated spermine bound to Z-DNA is the main driving force for the ternary complexation of 48 with a Z-DNA–spermine complex, in which the exciton

20 15

– O3S

48 + Z-DNA 48 + B-DNA

SO3

–

N N Ni

10

N

CD/(mdeg)

N

5

– O3S

– SO3

48

0 –5 –10 –15

250

300

350

400

450

500

550

Wavelength (nm)

Figure 9.31. CD spectrum of NiTPPS 48 (4 μM) in the presence of poly(dGdC)2 (50 μM) in righthanded B-form (gray) and in left-handed Z-form induced by adding spermine (black). (Reprinted with permission from reference 114. Copyright 2009 American Chemical Society.)

340

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(M)-49

(P)-49

Figure 9.32. (M)- and (P)-helicene 49.

coupling arises from the through-space interaction of bound porphyrins. The induced CD of a ternary DNA–spermine-48 complex disappears by raising the pH to 8.2, but is recovered by lowering the pH back to 6.9, suggesting that the induced CD can be modulated reversibly by pH. Enantioselective complexation with DNA is expected to occur for chiral species, and ECD is a powerful probe for investigating the chiral recognition upon complexation with DNA [115–117]. Sugiyama and co-workers studied the complexation behavior of (P )and (M )-helicene 49 with the B- and Z-form DNA of d(CGCm8 GCG)2 (Figure 9.32). The CD intensity of (P )-49 was reduced by 70% upon complexation with Z-DNA, but no appreciable change was induced by B-DNA. In contrast, antipodal (M )-49 did not show any chiral discrimination upon interaction with Z-DNA, for which the five-fold smaller affinity for (M )-49 would be responsible. Complexation with protein is a crucial issue in various biological phenomena, such as enzyme catalysis, antibody–antigen interaction, and drug delivery. Folding of polypeptide chain often produce crevices or cavities that function as binding sites for organic guest molecules. These binding sites are usually hydrophobic in nature and surrounded by a set of amino acid side chains that are arranged to optimize the noncovalent interactions with specific ligands. Commonly, more than one binding site will be created near the surface of a protein, and therefore the complexation of a guest with protein is elaborate in general. The binding affinity and stoichiometry of a guest in different binding sites rely on the size, shape, and functional group of the guest and the noncovalent interaction operating. In view of the intrinsically chiral nature of protein, ECD is one of the most crucial and widely employed tools for studying the interaction of organic guests with proteins. Inoue and co-workers [118–120] have studied the binding of AC to bovine (BSA) and human serum albumin (HSA) by means of ECD. As illustrated in Figure 9.33, complexation of AC with BSA induced well-structured CD at 330–400 nm. The intensity of positive CD induced was almost proportional to the AC concentration to reach a maximum at AC/BSA = 1 and then decreased gradually to eventually give negative CD upon further addition of AC. Detailed Job plot and titration experiments using CD, UV–vis, and fluorescence spectroscopy revealed the presence of four independent binding sites for AC in BSA, which respectively accommodate 1, 3, 2, and 3 AC molecules in the following order of affinity: K = 5.3 × 107 , 1.3 × 105 , 1.4 × 104 , and 3.0 × 103 M−1 . Similarly, a CD spectral titration with HSA showed four inflection points at AC/HSA = 1, 2, 5 and 10, indicating the presence of five binding sites that accommodate 1, 1, 3, 5, and >10 AC molecules in the order of decreasing affinity [120].

9.4.3. Binding with Synthetic Chiral Host Shinkai and co-workers [121–123] investigated the binding of achiral guests 50 and 51 by per-(S )-2-methylbutylated calixarenes of different ring sizes, 52[n] (Figure 9.34). In

341

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

AC/BSA 1

(a)

10

5

q (mdeg)

10

(c)

0

0

5 (b) 1

5

q391 (mdeg)

10

0

0 –5 –5 10 –10

–10 300

350

400

0

2

Wavelength (nm)

4

6

8

10

AC/BSA

Figure 9.33. CD spectral change upon addition of AC to a phosphate buffer solution (pH 7) of BSA (0.08 mM) at 25◦ C; (a) [AC] = 0–0.08 mM (from bottom to top); (b) [AC] = 0.08–0.8 mM (from top to bottom); (c) CD intensity at 391 nm as a function of AC/BSA ratio. (Reprinted with permission from [118]. Copyright 2003 American Chemical Society.)

50

51

52[n]

Figure 9.34. Chiral calixarenes 52[n] and achiral azobenzene guests 50 and 51.

the absence of guest, positive CD is observed for 52[4], but split CD for 52[6] and 52[8]. Addition of aliphatic alcohols did not appreciably change the CD spectrum of 52[4], but considerably reduced the CD intensities of 52[6] and 52[8]. 4-Cyano-4 -(diethylamino) azobenzene 51 showed a negative CD couplet upon complexation with 52[6]. However, a positive couplet was observed with 52[8], suggesting that the CD-active species are not monomeric but are instead aggregates of 51, which are arranged counterclockwise with 52[6] and clockwise with 52[8]. Synthetic chiral host 53 shows a dramatic CD change upon complexation with sulfate anion [124]. In the absence of sulfate, chiral guanidium host 53 exhibits a simple positive Cotton effect. Addition of sulfate leads to the formation of 2:1 complex (Figure 9.35),

342

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N N

N O N H O

N O

N H

OTBDPS

SO42– O

OTBDPS

N H O

N H O S

53 TBDPSO

O H N

O H N

O O

N

N

Figure 9.35. 1:2 Complexation of sulfonate anion with chiral guanidine host 53.

inducing a bisignate Cotton effect. Further addition of sulfate switches the complex stoichiometry from 2:1 to 1:1 with accompanying CD spectral change from bisignate to simple negative one.

9.5. CHIRAL MOLECULAR ASSEMBLY Chiral chromophoric compounds often form aggregates with accompanying CD spectral changes. Thus, CD spectral study provides not only the evidence for aggregation but also the structural information of the supramolecular assembly [125]. Achiral chromophore included in well-defined chiral aggregates, such as liquid crystal, may also display induced CD signals, from which the chiral supramolecular structure can be deduced [126–129].

9.5.1. Chiral Homo-aggregate Cyclodextrins modified with a long rigid chromophore of appropriate size are prone to thread together to give linear self-aggregates or supramolecular polymers, in which the chromophore group penetrates into the cavity of other cyclodextrin. Harada and coworkers [130] revealed that β-cyclodextrin 4-aminocinnamate 54 (Figure 9.36) forms a tail-to-tail dimeric aggregate in aqueous solution, while 4-(trinitroanilino)cinnamate 55, possessing a bulkier terminal group, forms head-to-tail aggregates to give a gel. The induced CD signals at 220–350 nm (Figure 9.37) are attributed to the inclusion of the chromophore in the β-cyclodextrin cavity. Interestingly, addition of urea (2 M), which is known to break the hydrogen bond, does not alter the CD spectrum, but the supramolecular polymer is disassembled by adding a better guest, adamantanecarboxylic acid, to give much weaker CD.

9.5.2. Chiral Hetero-aggregate The sergeants and soldiers principle and the majority rule originally found for conventional copolymers are also applied to supramolecular polymers [131–149]. Meijer and co-workers [145] studied the chiral supramolecular aggregation driven by hydrogenbonding and π –π stacking interactions. C3 -symmetric 56 (Figure 9.38), possessing nine identical chiral lipophilic chains at the periphery, forms a columnar structure through intermolecular hydrogen-bonding and stacking interactions. The anisotropy (g) factor of a solution of 56 is not proportional to the ee of 56 but obeys the majority rule (Figure 9.39). The free energy penalty for helix inversion is eight-fold larger than that

343

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

OR

R:

54

55

Figure 9.36. Chromophore-appended β-cyclodextrins 54 and 55.

(a)

30

AdCA 25 20 (b) θ (mdeg)

15 Urea 10 5 0 –5 –10 250

300 λ (nm)

350

400

Figure 9.37. Circular dichroism spectra of 1 mM 55 (black line), in the presence of 2 M urea (gray line) and in the presence of an excess of adamantanecarboxylic acid (AdCA) (dashed line). The inset shows the gel-to-sol transition upon addition of (a) 40 mM AdCA and (b) 2 M urea to a 20 mM solution of 55. (Reprinted with permission from reference 130. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

for incorporating an antipodal monomer with keeping the original helicity, which is the origin of the majority-ruled behavior of this system. The self-assembling mechanism of structurally resembling chiral (R)-57 and achiral 58 (Figure 9.40) was also investigated by ECD spectroscopy [136, 147]. A positive CD couplet was observed for (R)-57 aggregates, indicating the formation of right-handed

344

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N

Figure 9.38. C3 -symmetrical disk-shaped dendrimer (R)- and (S)-56.

helical columnar structure. Temperature-dependence behavior of the CD intensity of (R)57 revealed two distinct regimes of nucleation and elongation [150]. Above the critical elongation temperature, (R)-57 exists in the nonhelical nucleation state and gives almost no CD, and the chiral aggregates start to elongate only below that temperature. A mixture of (R)-57 and 58 obeys the sergeants and soldiers principle, because all supramolecular columns are in the same handedness upon addition of only 4% (R)-57 to 58,

9.5.3. Chirality Memory in Supramolecular Assembly Chiral supramolecular assembly originally constructed from chiral subunits may preserve its chiral character even after the chiral component is removed or replaced by an achiral substitute, provided that the kinetic and dynamic requirements are met. Reinhoudt and co-workers [151, 152] discovered an interesting chirality memory phenomenon in the hydrogen-bonded aggregation of calix[4]arene dimelamines 59 (Figure 9.41) with chiral cyanurates. Calix[4]arene dimelamine 59 and (R)-barbiturate ((R)-BAR) forms M -helical complex (M )-[593 ·(R)-BAR6 ] in benzene through multiple hydrogen-bonding interactions, while the assembly of 59 with (S )-BAR affords antipodal (P )-[593 ·(S )BAR6 ]. The chiral barbiturate components of (M )-[593 ·(R)-BAR6 ] can be substituted for achiral cyanurates. Thus, addition of achiral butylcyanurate (BuCYA) to a solution of (M )-[593 ·(R)-BAR6 ] leads to the displacement of (R)-BAR by BuCYA with only accompanying a slight CD change, indicating preservation of the helical structure and sense. The rate-determining step of the racemization involves the dissociation of 59 from an intact assembly, followed by a quick disk-rotation and reassembling to the antipodal assembly.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

345

100 80

Δε (L/mol.cm)

60 40 20 0 –20 –40 –60 –80 –100 200

250

300

350

400

450

500

λ (nm) (a)

0.002

g

0.001

0.000

–0.001

Figure 9.39. (a) CD spectra of octane solutions of dendrimer (S)-56 (open circles) and (R)-56 (closed circles); c = 2.49 × 10−5 M. (b) Anisotropy (g) factor as a function of the enantiomeric excess of

–0.002 –100

–50

0

50

100

Enantiomeric excess [%] (b)

56 at 20◦ C (closed circles) and 50◦ C (open circles). (Reprinted with permission from reference 145. Copyright 2005 American Chemical Society.).

9.6. SPONTANEOUS SYMMETRY BREAKING IN SUPRAMOLECULAR SYSTEM Controlling microscopic chiral event through macroscopic operation provides an important and intriguing tool for readily manipulating molecular and supramolecular chirality. Several recent studies based on ECD spectral analysis have drawn much attention to this possibility. Sodium chlorate crystallizes in l- and d-chiral forms in statistically equal numbers when crystallized from an aqueous solution without stirring. Interestingly, the crystals formed from a stirred solution are exclusively l- or d-chiral, although the handedness is not controlled [153]. It is believed that the autocatalytic secondary nucleation—that is, the formation of new crystal nuclei in the vicinity of an existing parent crystal—is responsible for this chiral symmetry-breaking processes. Chiral

346

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R HN

O

O

H N R

NH O

R

57: R =

Figure 9.40. Self-assembling chiral 58: R =

benzenetricarbamides (R)-57 and achiral 58.

N HN R

2

NH2

H2N

N

N

N NH

1

R

N NH

N

HN

R

R

2

1

(R)-BAR O

BuCYA

O O O

59: R1 = NO2; R2 = (CH2)3CH3 O HN O

X: NH

X

N

C CH3

O BuCYA (R)-BAR (S)-BAR

Figure 9.41. Chirality memory in a hydrogen-bonded assembly.

autocatalysis was also observed with the random generation of large ee in the crystallization of 1,1 -binaphthyl melt [154]. Ribo and co-workers [155, 156] investigated the spontaneous chiral symmetry breaking in the vortex motion-induced chiral J-aggregation of 5,10,15-tris(4-sulfonatophenyl)20-phenylporphyrin 60 during rotary evaporation, with accompanying strong exciton coupling ECD signals at the Soret band. Thus, the clockwise/counterclockwise rotation affords negative/positive CD couplet, respectively, at high 85% probabilities, while unstirred aggregation leads to no chirality dominance. It was concluded that the motion of oligomeric blocks formed during the aggregation of 60 is diastereotopic, when a particular vortex direction is externally prescribed, and such a preferential asymmetric accretion is imprinted into the aggregated material as the newly arriving blocks weld at definite arrangements.

347

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Me O

Me O

Me

Me O

Me O

Me O

Me O

O Me

O

O

O

Me

O

O

Me O

O

H O

O

O O

N

O

Zn N

Me O

O

O Me

O N

O Me

O

O

O

O O

O H

O O Me

O

O Me

O Me O

O

O Me

O

O

O

O

O Me

O Me

O

O

N

O

Me O

O Me

O

Me O

O

Me O O

O Me O O

Me O

O Me

O Me

O Me

O O

Me

O Me O Me

O Me

61

200 CCW

CD (mdeg)

100 0 OFF –100 CW –200 350

400

450

500 λ (nm)

550

600

650

Figure 9.42. CD spectra emerged upon rotary stirring in clockwise (CW) and counterclockwise (CCW) directions and without stirring (OFF). (Reprinted with permission from reference 157. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

Aida and co-workers [157] reported the induction of negative/positive CD couplet at the Soret band upon clockwise/counterclockwise rotary stirring of a solution containing dendric zinc porphyrin 61 (Figure 9.42). This stirring-induced exciton coupling was attributed to the macroscopic chiral alignment of the nanofibers formed by J-aggregation of 60 upon rotary stirring.

9.7. SUMMARY Supramolecular chirality is a highly intriguing, rapidly growing area of chemistry and biology. Investigations of supramolecular chirality not only provide valuable insights into the chiral phenomena occurring in natural and artificial supramolecular systems, but also

348

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

offer guiding principles for designing advanced supramolecular materials and for better understanding and mimicking the biological and pharmacological processes. This is also a challenging interdisciplinary subject that requires theoretical and experimental knowledge and techniques in quantum chemistry, synthetic chemistry, stereochemistry, supramolecular chemistry, and analytical chemistry. ECD spectroscopy is a powerful indispensable analytical tool for investigating chiral supramolecular phenomena, owing to the development of theoretical and instrumental CD spectral tools applicable to supramolecular systems. Nakanishi and Harada’s exciton chirality theory is the most widely applied approach that allows the determination of chiral spatial arrangement of chromophores as well as the design and construction of chiral supramolecular architectures. Other principles, such as the sector rule for cyclodextrin complexation, are also crucial for elucidating the detailed supramolecular orientation and conformation in certain systems. ECD is applicable to most of the chiral supramolecular phenomena, including the complexation of chiral/achiral hosts with achiral/chiral guests and the aggregation of molecules with chiral elements. Chromophores with high extinction coefficients are normally favored in ECD measurement, and chiral/achiral chromophores are commonly introduced to such supramolecular systems that lack absorption at appropriate wavelengths. Quantitative interpretation and prediction of supramolecular ECD are still a significant challenge at the moment. There are reasons to believe that broader implementation in the near future of the fast advancing quantum mechanical methodologies for predicting the chiroptical properties will bring considerable success to the field.

ACKNOWLEDGMENTS The authors are grateful to the supports of this work by PRESTO, JST (CY) and JSPS (YI).

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

G. Seeber, B. Tiedemann, K. Raymond, Top. Curr. Chem. 2006, 265 , 147–183. H. Eyring, H. Liu, D. Caldwell, Chem. Rev . 1968, 68 , 525–540. K. Wong, J. Chem. Educ. 1974, 51 , 573. T. Keiderling, Appl. Spectros. Rev . 1981, 17 , 189–226. D. Turner, I. Tinoco, Jr., M. Maestre, J. Am. Chem. Soc. 1974, 96 , 4340–4342. J. Dong, A. Wada, T. Takakuwa, K. Nakanishi, N. Berova, J. Am. Chem. Soc 1997, 119 , 12024–12025. J. Riehl, F. Richardson, Chem. Rev . 1986, 86 , 1–16. T. Bell, N. Hext, Chem. Soc. Rev . 2004, 33 , 589–598. A. Hembury Guy, V. Borovkov Victor, Y. Inoue, Chem. Rev . 2008, 108 , 1–73. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy: Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, 1983. K. Nakanishi, N. Berova, Exciton chirality method: Principles and applications, in Circular dichroism—Principles and Applications. Wiley-VCH, New York, 2000, pp. 337–382. T. Mizutani, H. Takagi, O. Hara, T. Horiguchi, H. Ogoshi, Tetrahedron Lett. 1997, 38 , 1991–1994. J. Etxebarria, H. Degenbeck, A. S. Felten, S. Serres, N. Nieto, A. Vidal-Ferran, J. Org. Chem. 2009, 74 , 8794–8797.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

14. H. Takagi, T. Mizutani, T. Horiguchi, S. Kitagawa, H. Ogoshi, Org. Biomol. Chem. 2005, 3 , 2091–2094. 15. J.-M. Fang, S. Selvi, J.-H. Liao, Z. Slanina, C.-T. Chen, P.-T. Chou, J. Am. Chem. Soc. 2004, 126 , 3559–3566. 16. W. Lu, L.-H. Zhang, X.-S. Ye, J. Su, Z. Yu, Tetrahedron 2006, 62 , 1806–1816. 17. X. Huang, B. Rickman, B. Borhan, N. Berova, K. Nakanishi, J. Am. Chem. Soc 1998, 120 , 6185–6186. 18. X. Huang, N. Fujioka, G. Pescitelli, F. Koehn, R. Williamson, K. Nakanishi, N. Berova, J. Am. Chem. Soc 2002, 124 , 10320–10335. 19. X. Huang, B. Borhan, B. Rickman, K. Nakanishi, N. Berova, Chemistry—Eur. J . 2000, 6 , 216–224. 20. G. Proni, G. Pescitelli, X. Huang, K. Nakanishi, N. Berova, J. Am. Chem. Soc 2003, 125 , 12914–12927. 21. G. Pescitelli, S. Gabriel, Y. Wang, J. Fleischhauer, R. Woody, N. Berova, J. Am. Chem. Soc 2003, 125 , 7613–7628. 22. Y. Chen, A. Petrovic, M. Roje, G. Pescitelli, M. Kayser, Y. Yang, N. Berova, G. Proni, Chirality 2010, 22 , 140–152. 23. A. Petrovic, Y. Chen, G. Pescitelli, N. Berova, G. Proni, Chirality 2010, 22 , 129–139. 24. X. Huang, K. Nakanishi, N. Berova, Chirality 2000, 12 , 237–255. 25. V.V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Phys. Chem. A 2000, 104 , 9213–9219. 26. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, Org. Lett. 2000, 2 , 1565–1568. 27. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Am. Chem. Soc. 2001, 123 , 2979–2989. 28. D. Hesek, G. A. Hembury, M. G. B. Drew, V. V. Borovkov, Y. Inoue, J. Am. Chem. Soc. 2001, 123 , 12232–12237. 29. V. Borovkov Victor, T. Harada, Y. Inoue, R. Kuroda, Angew. Chem. Int. Ed . 2002, 41 , 1378–1381. 30. V. Borovkov Victor, M. Lintuluoto Juha, H. Sugeta, M. Fujiki, R. Arakawa, Y. Inoue, J. Am. Chem. Soc. 2002, 124 , 2993–3006. 31. V. Borovkov Victor, M. Lintuluoto Juha, M. Sugiura, Y. Inoue, R. Kuroda, J. Am. Chem. Soc. 2002, 124 , 11282–11283. 32. V. V. Borovkov, T. Harada, Y. Inoue, R. Kuroda, Angew. Chem., Int. Ed . 2002, 41 , 1378–1381. 33. V. Borovkov Victor, T. Harada, A. Hembury Guy, Y. Inoue, R. Kuroda, Angew. Chem. Int. Ed . 2003, 42 , 1746–1749. 34. V. Borovkov Victor, A. Hembury Guy, Y. Inoue, Angew. Chem. Int. Ed . 2003, 42 , 5310–5314. 35. V. Borovkov Victor, I. Fujii, A. Muranaka, A. Hembury Guy, T. Tanaka, A. Ceulemans, N. Kobayashi, Y. Inoue, Angew. Chem. Int. Ed . 2004, 43 , 5481–5485. 36. V. V. Borovkov, G. A. Hembury, Y. Inoue, Acc. Chem. Res. 2004, 37 , 449–459. 37. V. V. Borovkov, Y. Inoue, Top. Curr. Chem. 2006, 265 , 89–146. 38. P. Bhyrappa, V. Borovkov Victor, Y. Inoue, Org. Lett. 2007, 9 , 433–435. 39. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Phys. Chem. B 1999, 103 , 5151–5156. 40. V. V. Borovkov, J. M. Lintuluoto, G. A. Hembury, M. Sugiura, R. Arakawa, Y. Inoue, J. Org. Chem. 2003, 68 , 7176–7192. 41. J. Lintuluoto, V. Borovkov, Y. Inoue, J. Am. Chem. Soc 2002, 124 , 13676–13677. 42. J. Etxebarria, A. Vidal-Ferran, P. Ballester, Chem. Commun. (Cambridge, U.K.) 2008, 5939–5941. 43. J. Etxebarria, H. Degenbeck, A. S. Felten, S. Serres, N. Nieto, A. Vidal-Ferran, J. Org. Chem. 2009, 74 , 8794–8797.

349

350

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

44. T. Morozumi, S. Shinkai, J. Chem. Soc., Chem. Commun. 1994, 1994 , 1219–1220. 45. K. Kondo, Y. Shiomi, M. Saisho, T. Harada, S. Shinkai, Tetrahedron 1992, 48 , 8239–8252. 46. Y. Shiomi, M. Saisho, K. Tsukagoshi, S. Shinkai, J. Chem. Soc., Perkin Trans. 1 1993, 1993 , 2111–2117. 47. S. Superchi, M. Donnoli, C. Rosini, Org. Lett 1999, 1 , 2093–2096. 48. G. Fasman, Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum Press, New York, 1996. 49. M. Green, B. Garetz, B. Munoz, H. Chang, S. Hoke, R. Cooks, J. Am. Chem. Soc. 1995, 117 , 4181–4182. 50. K. Ute, Y. Fukunishi, S. Jha, K. Cheon, B. Munoz, K. Hatada, M. Green, Macromolecules 1999, 32 , 1304–1307. 51. M. Green, N. Peterson, T. Sato, A. Teramoto, R. Cook, S. Lifson, Science 1995, 268 , 1860. 52. M. Green, M. Reidy, R. Johnson, G. Darling, D. O’Leary, G. Willson, J. Am. Chem. Soc. 1989, 111 , 6452–6454. 53. A. Hofacker, J. Parquette, Angew. Chem. 2005, 117 , 1077–1081. 54. K. Akagi, G. Piao, S. Kaneko, K. Sakamaki, H. Shirakawa, M. Kyotani, Science 1998, 282 , 1683–1686. 55. H. Goto, K. Akagi, Macromolecules 2005, 38 , 1091–1098. 56. R. Ohta, F. Togashi, H. Goto, Macromolecules 2007, 40 , 5228–5230. 57. H. Goto, K. Akagi, Angew. Chem. 2005, 117 , 4396–4402. 58. H. Nakashima, J. Koe, K. Torimitsu, M. Fujiki, J. Am. Chem. Soc 2001, 123 , 4847–4848. 59. T. Sugimoto, T. Suzuki, S. Shinkai, K. Sada, J. Am. Chem. Soc. 2007, 129 , 270–271. 60. C. Li, M. Numata, A.-H. Bae, K. Sakurai, S. Shinkai, J. Am. Chem. Soc. 2005, 127 , 4548–4549. 61. E. Yashima, K. Maeda, T. Nishimura, Chemistry—Eur. J . 2004, 10 , 42–51. 62. E. Yashima, T. Matsushima, Y. Okamoto, J. Am. Chem. Soc 1997, 119 , 6345–6359. 63. E. Yashima, Y. Maeda, Y. Okamoto, J. Am. Chem. Soc. 1998, 120 , 8895–8896. 64. E. Yashima, K. Maeda, Y. Okamoto, Nature 1999, 399 , 449–451. 65. H. Goto, Y. Okamoto, E. Yashima, Chemistry—Eur. J . 2002, 8 , 4027–4036. 66. H. Goto, Y. Okamoto, E. Yashima, Macromolecules 2002, 35 , 4590–4601. 67. H. Goto, E. Yashima, J. Am. Chem. Soc. 2002, 124 , 7943–7949. 68. M. Ishikawa, K. Maeda, E. Yashima, J. Am. Chem. Soc. 2002, 124 , 7448–7458. 69. H. Goto, Q. Zhang Hao, E. Yashima, J. Am. Chem. Soc. 2003, 125 , 2516–2523. 70. K. Maeda, K. Morino, Y. Okamoto, T. Sato, E. Yashima, J. Am. Chem. Soc. 2004, 126 , 4329–4342. 71. T. Nishimura, K. Tsuchiya, S. Ohsawa, K. Maeda, E. Yashima, Y. Nakamura, J. Nishimura, J. Am. Chem. Soc. 2004, 126 , 11711–11717. 72. E. Yashima, K. Maeda, Y. Takeyama, H. Mochizuki, K. Sakajiri, PMSE Prepr. 2004, 91 , 125. 73. M. Ikeda, Y. Tanaka, T. Hasegawa, Y. Furusho, E. Yashima, J. Am. Chem. Soc. 2006, 128 , 6806–6807. 74. T. Kajitani, K. Okoshi, S.-i. Sakurai, E. Yashima, Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 2006, 47 , 153. 75. S.-I. Sakurai, K. Okoshi, J. Kumaki, E. Yashima, J. Am. Chem. Soc. 2006, 128 , 5650–5651. 76. T. Miyagawa, M. Yamamoto, R. Muraki, H. Onouchi, E. Yashima, J. Am. Chem. Soc. 2007, 129 , 3676–3682. 77. T. Maeda, Y. Furusho, S.-I. Sakurai, J. Kumaki, K. Okoshi, E. Yashima, J. Am. Chem. Soc. 2008, 130 , 7938–7945.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

78. Y. Hase, K. Nagai, H. Iida, K. Maeda, N. Ochi, K. Sawabe, K. Sakajiri, K. Okoshi, E. Yashima, J. Am. Chem. Soc. 2009, 131 , 10719–10732. 79. M. Ishikawa, K. Maeda, Y. Mitsutsuji, E. Yashima, J. Am. Chem. Soc 2004, 126 , 732–733. 80. M. Waki, H. Abe, M. Inouye, Angew. Chem. Int. Ed . 2007, 46 , 3059–3061. 81. S. Takenaka, N. Matsuura, N. Tokura, Tetrahedron Lett. 1974, 15 , 2325–2328. 82. S. F. Mason, Chem. Phys. Lett. 1975, 32 , 201–203. 83. K. Harata, H. Uedaira, Bull. Chem. Soc. Jpn. 1975, 48 , 375–378. 84. K. Harata, Bioorganic Chemistry 1981, 10 , 255–265. 85. J. Kirkwood, J Chem Phys 1937, 5 , 479. 86. M. Kajtar, C. Horvath-Toro, E. Kuthi, J. Szejtli, Acta Chim. Acad. Sci. Hung. 1982, 110 . 87. M. Kodaka, T. Shimizu, M. Hatano, Bull. Chem. Soc. Jpn. 1983, 56 , 523–527. 88. M. Kodaka, T. Fukaya, Bull. Chem. Soc. Jpn. 1986, 59 , 2032–2034. 89. M. Kodaka, T. Fukaya, Bull. Chem. Soc. Jpn. 1989, 62 , 1154–1157. 90. M. Kodaka, J. Phys. Chem. 1991, 95 , 2110–2112. 91. M. Kodaka, J. Am. Chem. Soc. 1993, 115 , 3702–3705. 92. M. Barra, C. Bohne, J. C. Scaiano, J. Am. Chem. Soc. 1990, 112 , 8075–8079. 93. S. Hamai, T. Ikeda, A. Nakamura, H. Ikeda, A. Ueno, F. Toda, J. Am. Chem. Soc. 1992, 114 , 6012–6016. 94. D. Krois, U. H. Brinker, J. Am. Chem. Soc. 1998, 120 , 11627–11632. 95. B. Mayer, X. Zhang, W. M. Nau, G. Marconi, J. Am. Chem. Soc. 2001, 123 , 5240–5248. 96. M. Miyauchi, Y. Takashima, H. Yamaguchi, A. Harada, J. Am. Chem. Soc. 2005, 127 , 2984–2989. 97. C. Yang, T. Mori, Y. Origane, Y. H. Ko, N. Selvapalam, K. Kim, Y. Inoue, J. Am. Chem. Soc. 2008, 130 , 8574–8575. 98. Y.-L. Zhao, W. R. Dichtel, A. Trabolsi, S. Saha, I. Aprahamian, J. F. Stoddart, J. Am. Chem. Soc. 2008, 130 , 11294–11296. 99. A. Ueno, H. Yoshimura, R. Saka, T. Osa, J. Am. Chem. Soc. 1979, 101 , 2779–2780. 100. R. Lu, C. Yang, Y. Cao, Z. Wang, T. Wada, W. Jiao, T. Mori, Y. Inoue, Chem. Commun. (Cambridge, U.K.) 2008, 374–376. 101. R. Lu, C. Yang, Y. Cao, L. Tong, W. Jiao, T. Wada, Z. Wang, T. Mori, Y. Inoue, J. Org. Chem. 2008, 73 , 7695–7701. 102. C. Yang, M. Nishijima, A. Nakamura, T. Mori, T. Wada, Y. Inoue, Tetrahedron Lett. 2007, 48 , 4357. 103. C. Yang, A. Nakamura, T. Wada, Y. Inoue, Org. Lett. 2006, 8 , 3005. 104. C. Yang, A. Nakamura, G. Fukuhara, Y. Origane, T. Mori, T. Wada, Y. Inoue, J. Org. Chem. 2006, 71 , 3126. 105. C. Yang, T. Mori, Y. Inoue, J. Org. Chem. 2008, 73 , 5786. 106. C. Yang, C. Ke, K. Fujita, D.-Q. Yuan, T. Mori, Y. Inoue, Aust. J. Chem. 2008, 1–4. 107. C. Yang, G. Fukuhara, A. Nakamura, Y. Origane, K. Fujita, D.-Q. Yuan, T. Mori, T. Wada, Y. Inoue, J. Photochem. Photobiol. A: chem. 2005, 173 , 375. 108. A. Nakamura, Y. Inoue, J. Am. Chem. Soc. 2005, 127 , 5338. 109. A. Nakamura, Y. Inoue, J. Am. Chem. Soc. 2003, 125 , 966. 110. C. Ke, C. Yang, T. Mori, T. Wada, Y. Liu, Y. Inoue, Angew. Chem. Int. Ed . 2009, 48 . 111. R. Pasternack, Chirality 2003, 15 , 329–332. 112. B. Nord n, T. Kurucsev, J. Mol. Recognit. 1994, 7 , 141–155. 113. D. Fornasiero, T. Kurucsev, J. Phys. Chem. 1981, 85 , 613–618.

351

352

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

114. A. D’Urso, A. Mammana, M. Balaz, A. Holmes, N. Berova, R. Lauceri, R. Purrello, J. Am. Chem. Soc 2009, 131 , 2046–2047. 115. J. Barton, J. Biomol. Struct. Dyn. 1983, 1 , 621. 116. B. Nord´en, F. Tjerneld, FEBS Lett. 1976, 67 , 368. 117. Y. Xu, Y. X. Zhang, H. Sugiyama, T. Umano, H. Osuga, K. Tanaka, J. Am. Chem. Soc. 2004, 126 , 6566–6567. 118. T. Wada, M. Nishijima, T. Fujisawa, N. Sugahara, T. Mori, A. Nakamura, Y. Inoue, J. Am. Chem. Soc 2003, 125 , 7492–7493. 119. M. Nishijima, T. Pace, A. Nakamura, T. Mori, T. Wada, C. Bohne, Y. Inoue, J. Org. Chem 2007, 72 , 2707–2715. 120. M. Nishijima, T. Wada, T. Mori, T. Pace, C. Bohne, Y. Inoue, J. Am. Chem. Soc 2007, 129 , 3478–3479. 121. T. Arimura, H. Kawabata, T. Matsuda, T. Muramatsu, H. Satoh, K. Fujio, O. Manabe, S. Shinkai, J. Org. Chem. 1991, 56 , 301–306. 122. S. Shinkai, T. Arimura, H. Kawabata, H. Murakami, K. Iwamoto, J. Chem. Soc., Perkin Trans. 1 1991, 1991 , 2429–2434. 123. T. Arimura, S. Shinkai, Bull. Chem. Soc. Jpn. 1991, 64 , 1896–1900. 124. K. Kobiro, Y. Inoue, J. Am. Chem. Soc 2003, 125 , 421–427. 125. G. Gottarelli, S. Lena, S. Masiero, S. Pieraccini, G. Spada, Chirality 2008, 20 , 471–485. 126. F. Saeva, J. Wysocki, J. Am. Chem. Soc. 1971, 93 , 5928–5929. 127. F. Saeva, G. Olin, J. Am. Chem. Soc. 1973, 95 , 7882–7884. 128. K. Tang, M. Green, K. Cheon, J. Selinger, B. Garetz, J. Am. Chem. Soc 2003, 125 , 7313–7323. 129. T. Kato, T. Matsuoka, M. Nishii, Y. Kamikawa, K. Kanie, T. Nishimura, E. Yashima, S. Ujiie, Angew. Chem. Int. Ed . 2004, 43 , 1969–1972. 130. W. Deng, H. Yamaguchi, Y. Takashima, A. Harada, Angew. Chem. Int. Ed . 2007, 46 , 5144–5147. 131. L. J. Prins, P. Timmerman, D. N. Reinhoudt, J. Am. Chem. Soc. 2001, 123 , 10153–10163. 132. M. Mateos-Timoneda, M. Crego-Calama, D. Reinhoudt, Chem. Soc. Rev . 2004, 33 , 363–372. 133. M. Crego-Calama, D. N. Reinhoudt, M. G. J. ten Cate, Top. Curr. Chem. 2005, 249 , 285–316. 134. M. A. Mateos-Timoneda, M. Crego-Calama, D. N. Reinhoudt, Supramol. Chem. 2005, 17 , 67–79. 135. A. Mateos-Timoneda Miguel, M. Crego-Calama, N. Reinhoudt David, Chemistry 2006, 12 , 2630–2638. 136. L. Brunsveld, A. Schenning, M. Broeren, H. Janssen, J. Vekemans, E. Meijer, Chem. Lett. 2000, 29 , 292–293. 137. A. R. A. Palmans, J. A. J. M. Vekemans, E. E. Havinga, E. W. Meijer, Angew. Chem. Int. Ed . 1997, 36 , 2648–2651. 138. L. Brunsveld, J. H. K. K. Hirschberg, A. R. A. Palmans, J. A. J. M. Vekemans, E. W. Meijer, Book of Abstracts, 217th ACS National Meeting, Anaheim, California, March 21–25 1999, POLY-300. 139. B. Langeveld-Voss, R. Waterval, R. Janssen, E. Meijer, Macromolecules 1999, 32 , 227–230. 140. L. Brunsfeld, B. G. G. Lohmeijer, J. A. J. M. Vekemans, E. W. Meijer, J. Inclusion Phenom. Macrocyclic Chem. 2001, 41 , 61–64. 141. L. Brunsveld, E. W. Meijer, R. B. Prince, J. S. Moore, J. Am. Chem. Soc. 2001, 123 , 7978–7984. 142. L. Brunsveld, J. A. J. M. Vekemans, J. H. K. K. Hirschberg, R. P. Sijbesma, E. W. Meijer, Proc. Natl. Acad. Sci. U.S.A. 2002, 99 , 4977–4982.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

143. J. J. van Gorp, J. A. J. M. Vekemans, E. W. Meijer, J. Am. Chem. Soc. 2002, 124 , 14759–14769. 144. M. Masuda, P. Jonkheijm, P. Sijbesma Rint, E. W. Meijer, J. Am. Chem. Soc. 2003, 125 , 15935–15940. 145. J. van Gestel, A. R. A. Palmans, B. Titulaer, J. A. J. M. Vekemans, E. W. Meijer, J. Am. Chem. Soc. 2005, 127 , 5490–5494. 146. A. J. Wilson, M. Masuda, R. P. Sijbesma, E. W. Meijer, Angew. Chem. Int. Ed . 2005, 44 , 2275–2279. 147. M. M. J. Smulders, A. P. H. J. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2008, 130 , 606–611. 148. M. M. J. Smulders, I. A. W. Filot, J. M. A. Leenders, P. van der Schoot, A. R. A. Palmans, A. P. H. J. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2010, 132 , 611–619. 149. M. M. J. Smulders, P. J. M. Stals, T. Mes, T. F. E. Paffen, A. P. H. J. Schenning, A. R. A. Palmans, E. W. Meijer, J. Am. Chem. Soc. 2010, 132 , 620–626. 150. P. Jonkheijm, P. van der Schoot, A. Schenning, E. Meijer, Science 2006, 313 , 80. 151. L. J. Prins, J. Huskens, F. De Jong, P. Timmerman, D. N. Reinhoudt, Nature (London) 1999, 398 , 498–502. 152. L. J. Prins, F. De Jong, P. Timmerman, D. N. Reinhoudt, Nature (London) 2000, 408 , 181–184. 153. D. Kondepudi, R. Kaufman, N. Singh, Science 1990, 250 , 975. 154. D. Kondepudi, J. Laudadio, K. Asakura, J. Am. Chem. Soc 1999, 121 , 1448–1451. 155. J. Ribo, J. Crusats, F. Sagues, J. Claret, R. Rubires, Science 2001, 292 , 2063. 156. C. Escudero, J. Crusats, I. Dez-Prez, Z. El-Hachemi, J. Rib Angew. Chem. 2006, 118 , 8200–8203. 157. A. Tsuda, A. Alam Md, T. Harada, T. Yamaguchi, N. Ishii, T. Aida, Angew. Chem. Int. Ed . 2007, 46 , 8198–8202.

353

10 THE ONLINE STEREOCHEMICAL ANALYSIS OF CHIRAL COMPOUNDS BY HPLC-ECD COUPLING IN COMBINATION WITH QUANTUM-CHEMICAL CALCULATIONS ¨ Gerhard Bringmann, Daniel Gotz, and Torsten Bruhn

10.1. INTRODUCTION The combination of HPLC with analytical methods, like NMR, MSn , and CD is one of the most powerful analytical tools for the structural elucidation of chiral compounds, especially in natural product chemistry, where one often has to deal with complex mixtures, small product quantities, and/or chemically or stereochemically unstable substances [1]. In 1980 Mason and co-workers [2] reported the first hyphenation of HPLC with a CD dichrograph. By this combination it became possible to measure absorptions and optical activities simultaneously, in the on-flow mode. Salvadori et al. [3] were the first to describe the determination of absolute configurations of simple and known chiral compounds by recording their CD signals at a suited single wavelength and interpreting the obtained CD effect by empirical or nonempirical rules like, for example, the octant rule or the exciton chirality method. A huge step forward was provided by Mannschreck and co-workers [4, 5] in 1992: They succeeded for the first time in the online measurement of full CD spectra by HPLC in the stopped-flow mode. The main advantage of the technique is that stereoisomers no longer need to be separated in a time-consuming semipreparative way for offline CD measurements. In an analogous way, the HPLC-NMR technique was refined in the 1980s and 1990s and in 1998 Bringmann et al. were the first to report on HPLC-ROESY-NMR measurements in the online structural analysis of natural products [6]. Together with the HPLC-MSn technique the concept of the “analytical triad” LC-MSn -NMR-CD was born [7], permitting the elucidation of full absolute stereostructures directly from the peak in crude mixtures, thus saving the often laborious, time- and money-consuming isolation procedures. In the following years, several convincing examples that evidenced Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

355

356

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the efficiency of this method were reported, such as the online structural elucidation of the configurationally semi-stable biaryl alkaloid dioncophylline E [8] and the dimeric metabolite ancistrogriffithine A, which possesses two stereocenters and two chiral axes [9]. Another interesting application of the analytical triad is the direct monitoring of biotransformations of metabolites in plant cell cultures and their structural analysis as described by Iwasa et al. [10–12]. HPLC-CD has also become a valuable analytical method in the pharmaceutical industry for drug screening and quality management [13, 14]. Thus, the scope of this chapter is, to demonstrate the huge potential of HPLC-CD coupling in modern analytical chemistry.

10.2. HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY Between 1960 and 1970, high-performance liquid chromatography (HPLC) was developed as a new and efficient analytical tool. Today, chiral resolution by HPLC is the most widely practiced analytical method for determination of optical purity, being applicable even to samples that include many impurities [15]. In addition, liquid chromatography is the only technique that permits the separation and identification down to femtomolar components in complex matrices, but also allows for the isolation and purification of synthetic industrial products in ton quantities [13, 16]. With the high safety standards, optical-purity analysis of pharmaceutical agents and agrochemicals is nowadays strictly required, since the presence of “undesired” stereoisomers, even in small quantities, may sometimes lead to harmful side effects [17, 18]. In this context, HPLC-CD coupling—that is, the hyphenation of a CD dichrograph to an HPLC device—offers a unique potential for the stereochemical investigations on chiral analytes, even if occurring in trace concentrations and accompanied by further byproducts. Important information that can be obtained from an HPLC-CD experiment are the determination of the enantiomeric (or diastereomeric) excesses [19], the study of isomerization processes of stereochemically unstable analytes [5, 20], the determination of the elution order of stereoisomers, and the measurement of full CD spectra of even minor compounds from crude extracts or reaction mixtures. The current chapter focuses on the latter three issues.

10.3. THE HPLC-CD DEVICE In general the setup of common HPLC-CD interfaces as schematically depicted in Figure 10.1 has essentially remained unchanged since its first introduction in 1980 by Mason et al.: The outlet of a standard HPLC system is connected to a flow cell installed within a “normal” CD detector. In the early days of HPLC-CD coupling, only individually constructed instruments were used by a small number of experts in the field. But meanwhile the technique has become broadly available to nonspecialized end-users and some companies offer benchtop solutions for HPLC-CD applications at affordable costs. Usually an optical detector and a chiroptical one are connected in series to record UV and CD spectra simultaneously. Similar to usual UV detection in chromatography, a phase-sensitive CD trace can be obtained by monitoring the differential absorption of left- and right-circularly polarized light (A = AL − AR ) at a fixed wavelength λ, and the resulting output is thus a plot of A against time. The hyphenation of the chromatographic system to the spectrometer can be achieved by using a motor valve, permitting to stop the solvent flow through the measurement

357

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

Standard HPLC device Motor valve

Column UV detector

CD detector

Recorder output Waste UV signal A ‘On-flow ’ chromatograms at a fixed wavelength λ

A

ΔA

CD signal at λ1 t

UV

A

UV

Recorder ΔA

CD

λ

λ1

Full online spectra recorded in the λ “stopped-flow” mode

CD

λ1

λ λ

Figure 10.1. General schematic representation of a standard HPLC-CD device.

cell—for example, by redirecting the eluent from the HPLC pump directly into a waste flask. Consequently, full CD and UV spectra can be recorded in the stopped-flow mode with commercially available ECD detectors usually covering the spectral range from 200 to 850 nm. In addition, some HPLC pumps offer the possibility to keep the current eluent composition constant during the stopped-flow measurement thus permitting to subsequently proceed with the analysis of further substances within the same run. The huge potential and wide application range of HPLC-CD hyphenation is based on the simple experimental setup and, in particular, on the fact that CD itself is a quite sensitive method that can detect trace amounts of chiral compounds as long as they contain sufficiently UV-absorbing chromophores [21, 22]. In order to further improve the sensitivity of CD measurements also more specialized CD detection devices applying laser-beam sources [23, 24] or using phase-sensitive FDCD (fluorescence detection circular dichroism) [25–27] have been developed. Of these, however, only the laser-based CD detectors have so far been used in HPLC-CD hyphenation [23]. Beyond the outstanding potential of HPLC-CD alone, the additional hyphenation of high-performance liquid chromatography with further spectroscopic methods—that is, with tandem mass spectrometry (HPLC-MS/MS) and NMR spectroscopy (HPLCNMR)—has led to the “analytical triad” LC-MS/MS-NMR-CD [7]. This combined methodology has, during the past years, been applied to the full structural elucidation of—even complex—natural products right from crude extracts, as part of the strategy of a spectroscopy-guided search for structurally novel metabolites (Figure 10.2).

10.4. CHOICE OF THE CHROMATOGRAPHIC SYSTEM Mixtures of diastereomers can, in principle, be separated by HPLC on achiral phases. The chromatographic resolution of racemic mixtures, by contrast, requires a chiral auxiliary. It can be achieved by various approaches [13, 28], for example by the conversion of

358

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HPLC

HPLC

NMR

MS Fragmentation patterns by MS/MS or exact mass from HRMS - molecular formula - constitution HPLC Polarizer

Resolution of crude extracts or reaction mixtures by HPLC

CD

1D and 2D experiments (1H, 13C, TOCSY, COSY, NOESY, HMQC, ROESY)

NMR - constitution - relative configuration

'On-flow' monitoring of chiral analytes and recording of full CD spectra in the 'stopped-flow' mode - enantiomeric excesses (ee) Detector - absolute configuration - ...

CD Quantum-chemical calculations NMR and CD calculations absolute configuration

Full absolute stereostructures right from the peak of the chromatogram of a crude mixture!

Figure 10.2. The fruitful interplay of HPLC-MS/MS, HPLC-NMR, and HPLC-CD within the ‘‘analytical triad’’ combined with quantum-chemical calculations.

the enantiomers into diastereomeric derivatives using chiral reagents and subsequent separation on an achiral column, by using a chiral mobile phase, or by complex formation with chiral additives (like, for example, cyclodextrins [29]). The derivatization, however, requires an additional synthetic step prior to chromatography and can be hampered by different reaction rates of the enantiomers (leading to a kinetic racemic resolution). The use of chiral solvents or additives usually causes substantial disadvantages, especially in hyphenated HPLC-CD applications, since (depending on the detection wavelength) the chiral auxiliary itself may give its own CD response and may, thus, falsify the overall signal. This can imply even more serious drawbacks if a solvent gradient is applied—that is, with a varying composition of the mobile phase. Consequently, HPLC on a chiral stationary phase (CSP) is the most common and broadly applicable method for online HPLC-CD analysis. To date, a plethora of both normal- and reversed-phase CSPs have been developed, of which more than 100 are commercially available [30]. The chiral resolution is based on diastereomeric interactions of enantiomers with the CSP, namely their differential adsorption, resulting in different retention times for the two enantiomers. For a given resolution problem the separating capacity and recognition ability of a CSP can be evaluated qualitatively and quantitatively by two main characteristic values: The separation (or selectivity) factor α and the resolution factor R (t1 and t2 are the retention times of the faster and the more slowly eluting enantiomers, respectively; t0 is the retention time of a nonretained compound, that is, the dead time; k1 and k2 are the retention factors; w1 and w2 are the peak widths at their bases) [31,

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

32]: α=

k2 (t2 − t0 ) = , (t1 − t0 ) k1

R=

2(t2 − t1 ) , w1 + w2

k1 =

(t1 − t0 ) , t0

(t2 − t0 ) . t0

k2 =

The separation factor α reflects the selectivity of the CSP, namely, the affinity of the selected column for the individual enantiomers. The column performance is expressed by the plate number N ; thus the more efficient the column, the smaller will be the peak width w at a given retention time t for a component: 

t N = 16 w

2 .

The fundamental equation for optimizing HPLC separation conditions relates the resolution R to the number of theoretical plates N , the selectivity factor α, and the retention factor k2 : √   k2 N α−1 R= . α 1 + k2 4 Thus, R is a basic measure of the efficacy of the chromatographic system in separating two components in a mixture and in order to provide a good resolution, the three terms have to be maximized [32, 33]. Optimization of the experiment usually involves manipulation of column and mobile-phase parameters to alter the relative migration rates of the components in the mixture and to reduce peak broadening. While baseline separation between two peaks usually requires an R value >1.5, a resolution around R = 1.0 may sometimes be sufficient for HPLC-CD measurements. The reason for this seemingly higher resolution of LC-CD compared to LC-UV lies in the fact that CD spectra contain one (half) dimension more than UV spectra, in having positive and negative signals. Since enantiomers exhibit opposite Cotton effects, the CD signals of overlapping enantiomeric peaks partially compensate each other. Thus, in contrast to UV detection, two residual—opposite—signals may still remain at the edges of the seemingly unresolved peak in the UV chromatogram. Already during optimization of the chromatographic system the CD trace may provide a first hint at partial success, long before the resolution of the two peaks becomes visible by UV detection [34, 35]. Increasing N by lengthening the column leads to a longer retention time and augmented peak broadening, which may not be desirable. Alternatively, the number of theoretical plates can effectively be increased by reducing the size of the stationary-phase particles. In addition, separations may be improved by controlling the retention factor k . The retention factors should normally lie between 2 and 5, but for complex mixtures a larger range may be required to resolve all components. The value of the retention factor for a given compound depends on its chemical properties and the following experimental variables: 1. Flow rate 2. Composition of the mobile phase (including pH value)

359

360

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3. Column temperature 4. Stationary phase In many cases, the quality of the resolution is most efficiently improved by manipulating the selectivity factor α. If α is close to 1.0, optimization of k and increase of N may not be sufficient to achieve good separations with reasonable retention times. In these cases, k is optimized first and then α is increased, again by changing the mobilephase composition or the column temperature or by switching to a different stationary phase. Detailed procedures for optimizing the chromatographic system for an individual application have intensely been reviewed in various papers and book contributions [30, 36–38].

10.5. CHOICE OF APPROPRIATE DETECTION WAVELENGTHS The choice of a suitable detection wavelength is of utmost importance for HPLC-CD measurements, since it determines the response factor of the detector, which, in turn, affects both its sensitivity and selectivity. Concerning the selectivity, one may choose one of the following two “adjustment modes,” depending on the desired information: A nonselective detector will monitor the majority of different components of a mixture, while the optimization of the detection wavelength to a selected compound might preferentially record a response arising from one single species in a crude extract. Detection wavelengths below 250 nm are generally more suitable to reflect a plethora of different chiral substances in complex mixtures, since a large number of chromophores exhibit significant absorptions in this wavelength region. On the other hand, this also means that UV and CD detection are often hampered by interfering contributions of the solvents used as the mobile phase, especially if they are not completely devoid of absorbing contaminants. While the mobile phase has of course to be chosen primarily according to an optimum solubility of the analytes and the efficacy of the separation, a fine-tuning of the solvent composition may be advantageous or can even become necessary, especially for the measurement of full online CD spectra by HPLC-CD coupling: Then it may be advisable to substitute the eluent by a solvent with a similar polarity but different UV properties. Sometimes it also may be important to substitute a hydrogen-bond donating solvent like MeOH by a nonprotic one, to achieve spectra more similar to the gas-phase spectra and to the results of quantum-chemical predictions. For example, acetonitrile can often be used instead of methanol to minimize undesired absorptions by the solvent in the region around 200 nm. Table 10.1 lists approximate cutoff wavelengths below which the eluent absorbance may become unacceptable. Below the wavelength λ0 , the absorption of the solvent exceeds 0.05 absorbance units (relative to water) with a pathlength of 10 mm (i.e., A1cm > 0.05), while the absorption of the solvent is even 20 times higher at λ1 (A1cm > 1.0). Online CD measurements can be performed without problems down to λ0 , while CD curves should be interpreted with caution in the region between λ0 and λ1 , especially if the UV curve of the eluent rises steeply and/or if the observed/expected Cotton effects are small. Below λ1 CD measurements may yield ambiguous and sometimes badly reproducible results due to the strongly interfering UV absorption of the eluent. If at all possible, the detection of CD curves below λ1 should thus be avoided. In conclusion, the detection wavelength has to be adjusted carefully and sometimes a change of the mobile phase (if applicable) can offer an alternative to obtain CD spectra of highest possible quality.

361

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

TAB L E 10.1. Solvent Cutoff Wavelengths of Some Commonly Used HPLC Eluents [39] Eluent Dichloromethane Methanol or 2-propanol n-Hexane Acetonitrile Water

λ0 (nm)

λ1 (nm)

245 240 225 200 190

230 205 195 190 185

On the other hand, the sensitivity of the HPLC device is associated with the detection limit (LOD), which is strongly dependent on the spectroscopic properties of the analyte, but also on the stationary phase and the applied eluent. Since CD spectroscopy monitors the difference between the absorption of left- and right-circularly polarized light, the best signal-to-noise ratio (S/N) is usually obtained if ε is large and, at the same time, ε is comparatively small. Consequently, the most intense absorption may sometimes not provide the wavelength of choice for the detector setting. In general, CD detection affords the optimum S/N if the anisotropy factor g (see Section 10.6) is maximized. This is, however, possible only if the UV and CD properties of the compounds to be analyzed are known or can be estimated reasonably well. Frequently, the detection wavelength is more easily adjusted if the CD signal is known to arise from a specific transition like, for example, in the case of n → π * transitions in saturated ketones (around 300 nm). Furthermore, if the CD signal is expected to result from an exciton coupling—that is, the dipole–dipole interaction of locally excited states in adjacent, (ideally) identical chromophores—the CD detector is usually set to a wavelength that is red-shifted by about 10 nm as compared to the maximum absorption of the racemate. Since, by definition, the CD effect tends to zero at the UV maximum, this arbitrary shift of the CD detection wavelength usually fits the lowenergy extreme value of the respective couplet quite well. It is noteworthy that the CD response will switch sign if the detection wavelength is blue-shifted as compared to the corresponding UV absorption (for details see the chapter about the exciton chirality method). If two separate instruments are used for UV and CD detection, these can be adjusted to different wavelengths (as long as the g factor is not necessarily required). This might be advantageous since both sensitivity and selectivity can to some degree be tuned independently by the choice of the respective detection wavelengths of the UV (sensitivity) and CD (selectivity) detectors.

10.6. QUANTITATIVE ANALYSIS AND EE DETERMINATION USING HPLC-CD DETECTION: THE ANISOTROPY FACTOR G A common drawback in chromatography arises from the fact that the absolute quantity of a compound is hard to determine, especially at a single point of the chromatogram. However, with simultaneous UV and CD detection the so-called anisotropy or dissymmetry factor g can be derived from the ratio of the dichroic signal and absorbance [2, 34, 40, 41]: A . g= A

362

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The g factor does not depend on the sample concentration and is linearly related to the enantiomeric excess (ee), provided that the pathlength is the same for the UV and the CD measurement: ee . g = g max × 100 Thus, if the anisotropy factor g of the pure enantiomer (g max ) is known, the enantiomeric composition of the eluates can easily be determined at every single point of the chromatogram directly during the HPLC analysis: ee =

g g max

× 100

The direct measurement of the enantiomeric excess in a scalemic mixture is fundamental, for example, for the optimization of fraction collection in preparative liquid chromatography if two enantiomers can be resolved only partially on a given chiral stationary phase: The g factor of a pure enantiomer has a well-defined, constant value (g max ), which decreases when the first eluted peak in a partially resolved racemate becomes contaminated by the more slowly eluting peak—that is, by the oppositely configured analyte. Thus, monitoring of the anisotropy factor g permits to collect the largest-possible fractions of optically pure material from a partially resolved racemic or scalemic mixture [40].

10.7. GENERAL INTERPRETATION OF CD SPECTRA There are several strategies for the stereochemical analysis of experimental CD spectra, from merely empirical methods to quantum-chemical calculations. The most common (and merely experimental) approach to determine the absolute configuration of a novel chiral substance is the comparison of its CD spectrum with that of a structurally closely related, configurationally known compound. As simple and straightforward as this seems at first glance, the method implies hidden—and thus dangerous!—traps, since it is often difficult to judge whether the chosen reference structure is really ‘comparable’ or not. Of course the compared substances have to possess identical chromophores, with a similar stereo-orientation to each other. The pivotal effect of the conformation of the chromophores (e.g., of a phenyl substituent) on the overall CD can be seen in rocaglamide AE (1) versus its close, but cyclic, analogue cyclorocaglamide (rocaglamide AN, 2) [7, 42, 43], which shows a nearly opposite CD spectrum, despite the identity of the absolute configuration at all five stereocenters! The bridging in 2 stabilizes one particular conformational array, which is also present—but less populated—in 1, where chiroptically opposite conformers prevail. Furthermore, it is indispensable to know the influence of the different substituents of the chromophores on the CD spectrum. In most cases, simple substituents such as OH, OMe, or Me groups usually have no significant impact on a CD curve (Figure 10.3)—as long as their effects are overlayed by more dominant chromophores such as the naphthyl ring in dioncopeltine A (3) and habropetaline A (4) and if they do not have an effect to the conformation. By contrast, more strongly electron-withdrawing or -donating groups can effect a significant influence on the spectrum by altering the electronic structure and polarity of the subunits, thus changing the electron distribution in the chromophores, and the assignment of absolute configurations by a mere comparison of the CD spectra may become doubtful in such cases [44].

363

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

40 Cyclorocaglamide (2)

OMe OH OH

O

O HO

S R R R S

C

NMe2 O

Ph

CD [mdeg]

O

OMe

O OO

0 –20

MeO

O

20

–40 200

C

NMe2 O

Ph MeO

Rocaglamide AE (1)

OMe

OMe Rocaglamide AE (1)

OH

SRR RS

Cyclorocaglamide (2) 300

250

350

wavelength λ [nm] (a)

Me HO M MeO HO

R R

NH

OH Me 5'

CD [mdeg]

50 Dioncopeltine A (3)

25

Me HO M

0 Habropetaline A (4) –25

MeO MeO

Dioncopeltine A (3)

R

NH OH Me R

5'

Habropetaline A (4) –50 200

250

300

350

wavelength λ [nm] (b)

Figure 10.3. (a) Comparison of the CD spectra of rocaglamide AE (1) and cyclorocaglamide (2). Although they are constitutionally very closely related and possess the same absolute configuration, the CD spectra are nearly mirror-image like! (b) Dioncopeltine A (3) and habropetaline A (4). Changing the substituent at C5 from OH to MeO does not have any significant effect on the spectrum.

The use of semiempirical approaches (like the octant rule for saturated ketones) or of the nonempirical exciton chirality method may be a good alternative to derive absolute configurations from experimental CD curves. The octant rule and the exciton chirality approach are described in detail in other chapters of this volume. However, these methods are again limited to specific structures or to a detailed knowledge of transitions in the chromophores. The octant rule can only be applied to cyclic saturated ketones (or aldehydes) with known, rigid conformation and is not valid in the presence of an additional stronger chromophore, while the exciton chirality method has a broader range of application. In any case, however, the knowledge about the mutual orientation of the dipole moments of the chromophores is essential: Without knowing the exact and energetically relevant conformation(s) of the investigated structure, it is not possible to unambiguously assign absolute configurations. Thus, conformational analyses using, for example, DFT methods, are usually applied to give information about the different possible conformations of a new structure and then the exciton chirality method may become applicable [45].

364

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Chiral product of natural or synthetic origin

Experimental UV spectrum

+

Experimental CD spectrum 8

Comparison

exp.

0 8

–8

0

–16

Single UV spectra Starting geometry

Conformational analysis

Overall UV spectrum

Boltzmann weighting Single CD spectra

–24 200

calcd.

–8 250 16 24200 250 300 350

Absolute stereostructure

Comparison UV correction

Overall CD spectrum

Corrected overall CD spectrum

Figure 10.4. Flowchart of a general approach to calculate UV and CD spectra and of the determination of the absolute configuration by the comparison of experimental and computational results.

When completely new compounds with unknown chromophores are analyzed, or when there is a doubt in the applicability of certain helicity rules, the use of quantumchemical calculations is often the only way to determine the absolute stereostructure. This is achieved by comparing the experimental spectra with the ones quantum-chemically predicted for the respective stereoisomers (usually enantiomers). Which of the semiempirical, ab initio, or DFT methods will be the most appropriate has to be carefully decided and will be discussed in other chapters so that only some key facts of the basic, general approach (Figure 10.4) will be mentioned here: The first—and mandatory—step in any calculation of CD spectra has to be a solid conformational analysis with suited methods that yield reliable energies for the investigated class of compounds. CD spectra are very sensitive to even slightest conformational changes of the chiral molecule, and it can happen that two conformations with the same absolute configuration give rise to nearly mirror-image like CD curves. For example, the dihedral angle at the biaryl axis of a (P )-1,1 -binaphthyl has a drastic influence on the CD spectrum of the compound. With an angle between 50◦ and 100◦ a positive exciton couplet in the CD spectrum is produced, whereas an angle above 120◦ will give a negative one, although the absolute configuration of the chiral axis remains the same [45]. The observed experimental CD spectrum is the macroscopic result of the CD spectra originating from all individual molecules in a population—that is, the energy-weighted summation of the single CD spectra of all possible conformations of the measured structure according to their percental occurrence in the equilibrium mixture. The contribution of a single conformer to the overall CD (or UV) spectrum can be calculated based on a Boltzmann statistical weighting of the energies of the conformers found during the conformational analysis. In general, every conformation within an energy range of ∼12 kJ/mol above the global minimum may contribute to a significant degree, and so all conformations with an energy above this value can usually be disregarded for the subsequent calculation of the CD and UV spectra [46]. These CD computations will not give rise to full curves, but only to bar spectra (one single value for the energy of each of the excited states). To achieve a result that is optically comparable with the experiment, the single values have to be overlaid with

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

Gaussian-shape functions (sometimes Lorentz curves will be used instead) [47]. All of the single spectra thus obtained will then be summed up, taking into account the Boltzmann statistics of the conformational analysis, finally providing the full predicted CD and UV curves. In principle, it would be preferable to use the highest-level method available for the computation of the excited states of the chiral molecule. In practice, however, this is not feasible and one has to find a compromise between accuracy and computational costs.

10.8. APPLICATION OF THE ‘‘UV SHIFT’’ A simple empirical approximation, the so-called UV shift [46, 48], can sometimes help to keep the calculation time low. It is based on the following considerations: The most common software packages for quantum-chemical calculations will always compute rotational and oscillator strength values at certain excitations in one run. For both values the same method is used to obtain the energy of the excited state and thus the wavelength of the excitation. This means that any systematic error with respect to the wavelength determination will be the same for both spectra. Identification of the extent of this error is much easier for the UV than for the CD curve. By comparing the experimental UV spectrum with the calculated one, it is possible to determine the difference between the lambda values of the experimental and the calculated maxima—that is, the “UV shift” This empirical factor can then be applied to the CD spectrum, thus providing a better match between the experimental and the calculated data. If this match is not evident, the chosen calculation method is either unsuited or the absolute stereostructure (maybe even the constitution) for which the computation was done does not correspond to the analyzed compound. The use of the UV shift is, however, only recommended if the UV spectrum is already of sufficient quality, which, in simple words, means that the number of maxima, their relative intensities, and the energy differences between them should be the same in the experimental spectrum as in the computed one. Application of the UV shift often permits to tolerate larger systematic errors—for example, by using smaller basis sets, which can allow for significantly reduced computational time.

10.9. HPLC-CD IN PRAXIS In the following, the still underestimated potential of HPLC-CD analysis and its possible application areas will be demonstrated by a detailed description of two representative configurational assignments of chiral natural products, in particular the determination of the absolute configuration by a combination of HPLC-NMR, HPLC-CD, and quantumchemical calculations. Through these examples the reader will get an idea of how to perform a full configurational elucidation of chiral substrates by HPLC hyphenation techniques and will gain a feeling about possible difficulties encountered during the structural analysis and receive useful advice to overcome them. This provides a general guideline of how to proceed for the full stereochemical assignment of novel-type chiral compounds even from crude mixtures.

10.9.1. Ancistrocladium B—A Configurationally Semi-Stable, Axially Chiral Biaryl Ancistrocladinium B (5) is a novel-type metabolite isolated from an as yet not fully identified, possibly new Ancistrocladus species from the rainforest in the Democratic

365

366

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Republic of Congo [49]. It is one of the first of the recently discovered N , C -coupled representatives of the naphthylisoquinoline alkaloids. Its typical occurrence as a double peak in the HPLC chromatogram in combination with the observation that all isolation attempts always led to the same 45:55 mixture according to NMR and analytic HPLC indicated the presence of two diastereomers slowly interconverting at room temperature (Figure 10.5). That these were the respective atropo-diastereomers—that is, with different configurations at the (apparently configurationally semi-stable) novel N , C axis—became evident from the fact that both had the same configuration at the stereocenter at C3, S , as determined by oxidative degradation [50]. Due to the semi-stable configuration at the N , C axis, all offline spectra (NMR, CD) always recorded the diastereomeric mixtures, hampering an unambiguous elucidation of the absolute stereostructures with common methods, making ancistrocladinium B an excellent example of how to apply the so-called “analytical triad”, that is, the combination of HPLC-NMR, HPLC-MS, and HPLC-CD. 10.9.1.1. Optimizing the HPLC Separation Conditions. As for all HPLC hyphenation techniques, the first and most important step to the unequivocal elucidation of the constitution and the stereostructure of ancistrocladinium B (5) was the elaboration of a reliable method for the full resolution of this compound by HPLC. To get baseline-separated peaks, several columns (normal phase, RP-C18 and -C8 ) and solvents (acetonitrile, methanol, water) were tested. In addition, the pH value of the mobile phase and the temperature of the column had to be optimized to provide a good separation of the two presumed atropo-diastereomers, preferably with short retention times. After intense efforts, the best separation was achieved by using a Symmetry-C18 column (Waters, 4.6 × 250 mm; 5 μm) at 10◦ C with an isocratic solvent system of methanol and water (60:40, acidified with 0.05% TFA, flow rate: 0.8 mL/min). These conditions yielded a clear baseline separation with short retention times (Figure 10.5), giving one peak at about 17 minutes (Peak A) and a second one at about 19 minutes (Peak B). Still, a preparative separation of the two peaks to give fully pure diastereomers was not possible due to their slow interconversion at room temperature.

LC-UV 231 nm Peak A Peak B

MeO

*

Me OH

OMe

t [min]

N * OMe Me

17

19 LC-CD 235 nm

Me

Ancistrocladinium B (5) * stereogenic elements of initially unknown absolute configuration

Figure 10.5. HPLC-UV and HPLC-CD t [min] 17

19

(on-flow) chromatograms of ancistrocladinium B (5).

367

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

MeO

Me H S OH OMe N 6′

MeO

3

P

OMe Me H

Me H H N M 7′

3 S

Figure 10.6. Online-ROESY NMR

6′ 7′

MeO (P,3S)-5 (Peak A)

correlations diagnostically indicative for the relative configuration of the

OMe Me HO Me (M,3S)-5 (Peak B)

Me

atropo-diastereomers of ancistrocladinium B.

10.9.1.2. Information Obtained by HPLC-MS and HPLC-NMR Measurements. This interconversion was, however, slow enough to permit a complete stereochemical analysis online, right from the peak in the chromatogram, by applying the analytical triad for further structural investigations, using the optimized separation conditions described above. By HPLC-MS, each of the two peaks gave a mass of m/z = 406. HPLC-HRMS(ESI) experiments showed that both compounds were cationic, having the same molecular formula, C25 H28 NO4 + . This again confirmed that ancistrocladinium B was indeed a mixture of two isomers, probably atropo-diastereomers. For both peaks, 1 H, 13 C, COSY, ROESY, HMBC, and HMQC spectra were recorded by online HPLC-NMR measurements, using a Bruker Cryoprobe for higher 13 C sensitivity and a flow insert (CryoFit, Bruker) for HPLC-NMR hyphenation. The two compounds showed nearly identical NMR spectra, except for the fact that some of the signals displayed slightly different chemical shifts. The results corroborated the anticipated constitution of ancistrocladinium B (Figure 10.6); in addition, based on the online-ROESY correlations, even the relative configurations of the compounds were unambiguously assigned: The faster eluting Peak A displayed a diagnostically significant interaction between H7 and the proton at the stereocenter C3, showing these two protons to be both on the same side of the molecule—that is, both up (as drawn in Figure 10.6, left) or both down. The more slowly eluting Peak B, by contrast, had a strong correlation between H7 and the protons of Me3, hinting at the opposite relative configuration axis versus center (Figure 10.6, right). Therefore, the faster eluting atropo-diastereomer of ancistrocladinium B had to be (P , S )- or (M , R)-configured, while the more slowly eluting one had the (P , R)- or the (M , S )-configuration. Together with the (S )-configuration at C3 as already known from the degradation experiment, the two (3R)-configured stereoisomers were excluded. 10.9.1.3. HPLC-CD and Quantum-Chemical CD Calculations. To get an independent and unambiguous proof for the above assignments HPLC-CD measurements in combination with quantum-chemical calculations were performed. The full CD spectra of the two peaks were recorded online in the stopped-flow mode (Figure 10.7), and these spectra were nearly mirror-image like. It is known that the chiral axis in biaryls often dominates the CD spectrum and that additional stereocenters usually do have a negligible effect only (for exceptions due to the strong chromophores close to the stereogenic center, see reference 51). Thus, the CD spectra alone would already have provided a clear hint that the two peaks in ancistrocladinium B correspond to atropo-diastereomeric compounds. Assuming that the exciton chirality method (see Chapter 4 in this volume) was valid for ancistrocladinium B, the positive couplet around 350 nm would have predicted the (P )-configuration for Peak A and, vice versa, the (M )-configuration for Peak B with its negative couplet at 350 nm. However, without knowing the exact conformations of

368

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6

6 exp. Peak A

3 CD [mdeg]

CD [mdeg]

3

0 calcd. for (P,S)

–3

exp. Peak B

0

–3 calcd. for (M,S)

–6 200

–6 250 300 350 wavelength λ [nm]

MeO S

400

250 300 350 wavelength λ [nm]

200

Me + OH OMe N 6′

Me

MeO S

P

MeO

MeO Me

M

6′

MeO

Me Me

(P,3R)-5

Me

6 exp. Peak A

calcd. for (P,R)

exp. Peak B

3 CD [mdeg]

CD [mdeg]

+ OH OMe N 6′ P

MeO

6

3

Me

Me R

Me HO

(M,3R)-5

MeO

MeO

Me R

N+

6′

Me HO

(M,3S)-5

MeO

MeO

N+ M

Me

(P,3 S)-5

400

0

–3

0

–3 calcd. for (M,R) –6

–6 200

250 300 350 wavelength λ [nm]

400

200

250 300 350 wavelength λ [nm]

400

Figure 10.7. Assignment of the absolute configuration to the two—configurationally semistable—atropo-diastereomers of ancistrocladinium B (5) by comparison of the experimental LC-CD spectra (stopped-flow) of Peak A (left) and Peak B (right) with the spectra calculated for (P, 3S)-5, (M, 3R)-5, (P, 3R)-5, and (M, 3S)-5 by using TDDFT with subsequent UV-shift correction.

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

369

the diastereomers and the orientation of the dipole moments of the two chromophores, this was not a reliable assignment and thus computational calculations were done. To elucidate the absolute configuration of the diastereomers of 5 independently, the CD spectra of all possible configurations, (P , 3S )-5, (M , 3R)-5, (P , 3R)-5, and (M , 3S )5, were calculated and compared with the experimental results [46]. A conformational analysis was performed using B3LYP/6-31G* [52–55] yielding four relevant conformers for each diastereomer. For the subsequent TDAB3LYP/SVP [56, 57] calculations, solvent effects were taken into consideration by using the COSMO [58] approach, using an epsilon value of 56.52 and a refraction index of 1.33. By comparison of the calculated UV spectra with the experimental curves, a UV shift of 24 nm was determined and applied to the calculated CD curves. The CD spectrum computed for (P , 3S )-5 did fit quite well to the experimental one of Peak A, while the spectrum calculated for the (M , 3R)-enantiomer did not match the measured one, proving that Peak A was (P , 3S )configured. In the case of Peak B the comparison of the calculated spectrum of (M , 3S )-5 with the experiment showed a good agreement, while the curve calculated for (P , 3R)-5 again did not fit, corroborating the (M , 3S )-configuration of Peak B. 10.9.1.4. Further Investigations: Estimation of the Rotational Barrier. As already mentioned above, the iminium-aryl axis of ancistrocladinium B is configurationally semi-stable at room temperature. For an estimation of the barrier of rotation around the chiral axis, again HPLC experiments were carried out. For each of the two atropo-diastereomers the isomerization process was monitored by HPLC-UV measurements. The decrease of the diastereomeric excess of freshly purified fractions enriched in the respective (P )- or (M )-atropisomer was measured at three different temperatures

Peak B Peak A

t [min] 17

19

Isolation of Peak A

Isolation of Peak B

t = 330 min t

Figure 10.8. Determination of the axial isomerization

t = 180 min

t

rates of (P, S)-5 and (M, S)-5 by their chromatographic

t [min]

resolution and subsequent thermal equilibration over time (here exemplarily shown at 65◦ C), monitored by HPLC-UV on an achiral C18 phase.

t = 0 min 17

19 t [min]

17

19

370

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(55◦ C, 65◦ C, and 85◦ C, Figure 10.8), permitting determination of the respective rate constants of the isomerization process. By applying the Eyring equation, the Gibbs free energies of activation at room temperature (25◦ C) were calculated. The experimental values thus obtained were Grot = 105.8 kJ mol−1 for the isomerization from P to M and Grot = 105.7 kJ mol−1 for the conversion of M to P .1 In summary, the structure elucidation of ancistrocladinium B (5) impressively demonstrates the value of HPLC hyphenation techniques in modern natural products chemistry. Besides the establishment of the full constitution and relative configuration by HPLCNMR and HPLC-MS, also the absolute configurations of the two naturally occurring atropo-diastereomers of 5 were determined. Online HPLC-CD measurements were performed with trace quantities of crude material. Due to the unprecedented structure of the N , C -coupled chiral aryliminium cation, the CD spectra thus obtained could not be interpreted by empirical comparison with other compounds, but inevitably required quantum-chemical CD calculations. This finally permitted the unambiguous assignment of the full absolute stereostructures of the two diastereomers. Now knowing their elution order, HPLC techniques also served to determine the atropisomerization barriers by monitoring the decrease of the diastereomeric excess of freshly prepared samples enriched in one of the two rotational isomers. The absolute stereostructures of the two atropo-diastereomers of ancistrocladinium B (5) have meanwhile been confirmed by total synthesis [59].

10.9.2. The Absolute Axial Configuration of Knipholone and Knipholone Anthrone Knipholone (6) and knipholone anthrone (7) are well-known representatives of the class of naturally occurring phenylanthraquinones [60]. They are interesting biosynthetically (origin from eight plus four acetate units), pharmacologically (anti-infective and anti-tumoral properties), and, in particular, stereochemically, due to their rotationally hindered and thus configurationally stable biaryl axis (Figure 10.9) [60]. They have been initially discovered by Steglich, Dagne, and Yenesew in 1984 (knipholone) [61] and 1993 (knipholone anthrone) [62], but the elucidation of their correct absolute configurations succeeded in 2007 only [63]. Previous assignments using semiempirical calculations had deduced absolute configurations at the biaryl axis, which were in contrast to those expected from the results of a later stereoselective total synthesis using the lactone method [64, 65]. In nature, knipholone mostly occurs as a scalemic mixture (i.e., enantiomerically enriched, not enantiopure). In this chapter only, the stereostructure of the main enantiomer, (+)knipholone, will be discussed. In 2007 the correct absolute configuration was determined unambiguously and independently by renewed experimental work in combination with quantum-chemical calculations using higher-level methods [63]. The individual challenges arising in the course of these investigations will be described in the following paragraph, again demonstrating the value of the fruitful interplay of HPLC-CD measurements with computational work. 10.9.2.1. Remeasuring the CD Spectra of Knipholone and Knipholone Anthrone. As mentioned above, the initial assignment of the absolute configuration of knipholone (6) and knipholone anthrone (7) was in contrast to the results of the total Note that the two Grot values are not identical, although concerning the same transition state, due to the different energies of the two interconverting atropisomers, which in this case are diastereomers. This is also reflected by their experimentally determined isomeric ratio of 45:55 of M : P . 1

371

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

HO

O

HO

OH

O

OH

Me

Me X HO

M

X HO

OH

P OH

Me

Me MeO O (M)-6 X = O (M)-7 X = H2

O OMe (P)-6 X = O (P )-7 X = H2

Figure 10.9. Structures of knipholone (6) and knipholone anthrone (7).

(+)–6

4 Δε [cm2/mol]

Δε [cm2/mol]

5

0

–5

8

from B. capitata

(+)–7 from synthesis from K. foliosa

0 –4

from B. frutescens

–8 200

300 400 wavelength λ [nm]

500

200

300 400 wavelength λ [nm]

500

Figure 10.10. Experimental CD spectra of (+)-knipholone (6, left) and (+)-knipholone anthrone (7, right) from different origins (measured in methanol).

synthesis by applying the lactone concept [64].2 Several independent approaches were followed to clarify the origin of this discrepancy. First of all, the CD spectra of 6 and 7 isolated from different natural sources were remeasured to provide new and reliable experimental data (Figure 10.10). In addition, the influence of different solvents and solvent mixtures (n-hexane, methanol, acetonitrile:water 60:40) was investigated, since these may have a significant effect on a CD spectrum, due to the polarity and the hydrogen-donor or -acceptor properties of the solvent but also due to possible aggregation of the analyte in solution. In the case of knipholone, however, the spectra did not show any significant differences, regardless of whether the compounds were isolated from varying natural sources or measured in different solvents (Figure 10.11).

2

Note that in this experimental work the use of an S -configured catalyst in the lactone-cleavage reaction unexpectedly seemed to give the P -configuration (as erroneously deduced from the initially wrongly attributed absolute (P )-configuration of (+)-knipholone). This seeming contradiction was solved by the new—and correct—assignment of the absolute configuration of (+)-knipholone as being M -configured in 2007 [64].

372

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8

(+)–6

(+)–7

in n-hexane in n-hexane Δε [cm2/mol]

Δε [cm2/mol]

5

0

–5

in acetonitrile:water (60:40)

0

–8

300 400 wavelength λ [nm]

in methanol

–4

in methanol 200

4

200

500

in acetonitrile:water (60:40) 300 400 wavelength λ [nm]

500

Figure 10.11. Experimental CD spectra of (+)-knipholone (6, left) and (+)-knipholone anthrone (7, right) in various solvent systems.

8

8

(+) –7

(+) –7 Δε [cm2/mol]

Δε [cm2/mol]

fresh sample 4 0

4 0 –4

–4 partially decomposed

(+)–6 –8

–8 200

300

400

wavelength λ [nm]

500

200

300

400

500

wavelength λ [nm]

Figure 10.12. Experimental CD spectra of freshly prepared (+)-knipholone anthrone (7) and of a partially decomposed sample (left) and comparison of the CD spectra of (+)-knipholone and (+)-knipholone anthrone (right).

The newly recorded CD spectrum of knipholone anthrone (7), by contrast, obviously differed significantly from the spectra measured earlier (Figure 10.12): The freshly prepared sample of 7 showed a quite intense CD couplet at about 300 nm with a positive first Cotton effect, while this couplet was almost completely absent in the spectrum measured previously [65]. The question why this intense couplet had not been observed during the initial investigations was answered by the following experiment: Keeping the freshly purified 7 for 2 h in methanol at room temperature under ambient light led to a drastic change in the measured CD spectrum due to an intensity loss of the couplet around 300 nm. This clearly hinted at a partial decomposition of the compound to give a new chiral—and likewise chiroptically active—product, because the whole CD spectrum would just have decreased in intensity if simply racemization had occurred. The loss of only one couplet might be explained by the oxidation of knipholone anthrone to knipholone: While both compounds show a similar negative Cotton effect (CE) around 210 nm, their CEs around 300 nm are opposite. Consequently, these couplets will largely

373

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

1000 O

HO

OH

O

OH

HH HO

P

HO

800 Me O HO

600

Me

P OH

Me

Me 6

400

OH

O OMe

O OMe

7 *

unknown decomposition products

* 200 *

*

* *

0 3

4

5

6

32

7

8

9

10 t [min]

0h

CD [mdeg]

16 6h

24 h

Figure 10.13. HPLC-UV chromatogram of

0

a partially decomposed sample of knipholone anthrone (7) after 2 h at rt and ambient light (top) and full online HPLC-CD spectra of knipholone anthrone

−16 −32 200

300

400

wavelength λ [nm]

500

(7) after 0, 6, and 24 h in the flow cell (Chromolith column, solvent gradient of acetonitrile/water, both with 0.05% TFA).

compensate each other in a mixture of knipholone and knipholone anthrone resulting in a selective decrease of the intensity of the couplet at 300 nm. The above investigations exemplify that a detailed knowledge of the chemical stability of a given compound can be highly important for the measurements and interpretation of CD spectra: One may not only observe a decrease of the intensity of the entire CD curve, as in the case of a (partial) racemization, but also new signals may appear or genuine ones may disappear if new chiral compounds are formed in situ, thus changing the overall measured CD spectrum, which may easily lead to ambiguous or even wrong interpretations. Such falsifications of CD spectra by new (or existing) chiroptically active impurities can often be overcome by HPLC-CD: As an example Figure 10.13 shows the HPLC-UV chromatogram of partially decomposed knipholone anthrone (7) and full online CD spectra of 7. By HPLC-CD in the stopped-flow mode, the peak corresponding to knipholone anthrone was kept in the flow cell for several hours and the full CD curve was recorded after 0, 6, and 24 h. Since the sample was not exposed to light in the flow cell and air oxygen was largely excluded, no significant change of the CD spectrum was observed (Figure 10.13) even after 24 h, in contrast to the earlier presented offline measurement (cf. Figure 10.12), where decomposition may have significantly decreased the couplet at 300 nm already within a time span of 2 h.

374

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8 exp. for (+)–6

DFT/MRCI calcd. for (P)–7

4 Δε [cm2/mol]

Δε [cm2/mol]

5

0

0 –4

–5

exp. for (+)–7

DFT/MRCI calcd. for (P)–6 –8 200

300 400 wavelength λ [nm]

500

200

300 400 wavelength λ [nm]

500

Figure 10.14. Determination of the absolute configurations of (+)-knipholone (6) and (+)knipholone anthrone (7) by comparison of the experimental CD spectra with the CD curves calculated by DFT/MRCI.

10.9.2.2. Elucidation of the Absolute Configuration by QuantumMechanical Calculations. According to the exciton chirality method, the “newly found” couplet now also hinted at a (P )-configured biaryl axis of compound 7, again in contrast to the initial assignment of the absolute configuration as M [65]. Thus, in parallel to the renewed experiments, the computational investigations were rechecked leading to the conclusion that the previous semiempirical approaches had by far not been sufficient to unequivocally determine the absolute configuration 6 and 7. In the case of knipholone anthrone even TDDFT calculations led to ambiguous results and thus the DFT/MRCI approach developed by Grimme [66] was chosen for these phenylanthraquinones. Finally, the absolute configurations of the two compounds had to be revised: (+)-Knipholone (6) and (+)-knipholone anthrone (7) clearly do have the (P )-configuration (Figure 10.14), which is now in full accordance with the results of the total synthesis by using the lactone concept, and—in the case of 7—also with the attribution of the absolute configuration based on the exciton chirality method applied to the newly measured CD spectrum. 10.9.2.3. Stereochemical Correlation of 2 and 3 by Coelution Experiments. Additional support for the new assignment of the absolute configuration of the two compounds was expected from their stereochemically unambiguous interconversion by reduction of (P )-6 to give—if correctly assigned—(P )-7, and, vice versa, the oxidation of (P )-7 to deliver (P )-6, to be analyzed by chromatography on a chiral phase with online CD coupling (LC-CD). These investigations were, however, hampered by the partial racemization of the compounds under the applied conditions. Still, the enantiomers of the two compounds were easily resolved on a chiral OD-H column. Interestingly, they showed inverted elution orders: Thus, the (P )-enantiomer of knipholone (6) corresponded to the faster eluting peak of the two enantiomers, while for knipholone anthrone (7) the M -configured enantiomer had the shorter retention time (Figure 10.15). The described results highlight that elution orders of similar, but not identical, compounds do not necessarily provide a hint at their absolute configurations, because even two so closely related compounds—like 6 and 7—can show opposite elution orders of their identically configured enantiomers.

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

HO

OH

O

O

HO

Me

Me P

O HO

OH

OH

SnCl2, HOAc Me

Me 7

OH

KOH, air

P

HH HO

375

6

O OMe

O OMe

(b)

(a)

P

P

M LC-UV

LC-UV KOH, air

LC-CD

LC-CD

t [min]

t [min] 25 35 0:100

30 40 69:31 (c)

(d) M

P

P M

LC-UV

LC-UV

Figure 10.15. Proof of the stereochemical

SnCl2, HOAc

identity of (+)-knipholone as P (b), since it was obtained by oxidation of enantiomerically pure

LC-CD

LC-CD

t [min]

t [min] 25 35 48:52

30 40 75:25

knipholone anthrone (100:0 P to M) (a), and, vice versa, enantiomeric resolution of—mainly (P)-configured—knipholone anthrone (7) (d) obtained by reduction of authentic knipholone (75:25P to M) (c).

An authentic sample of enantiomerically highly pure synthetic (+)-knipholone anthrone (Figure 10.15a) was oxidized to knipholone showing only a ratio of 69:31P (rapid) to M (slow). And, vice versa, an authentic sample of knipholone (6) isolated from B. capitata, which had a ratio of 75:25P to M , was reduced to knipholone anthrone (7), whose now more slowly eluting main peak coeluted with the peak of the pure (P )-enantiomer of 7. The interconversion showed that the more rapidly eluting peak of knipholone anthrone (7) was (M )-configured, while in the case of knipholone (6) the faster eluting one was P . These results are in agreement with the fact that at the chosen wavelength of 290 nm (P )-configured 6 shows a positive Cotton effect, while the (P )-enantiomer of 7 displays a negative one! The example reveals that even seemingly marginal structural changes can be accompanied by an inversion of the CD response, at least at a given wavelength λ. Due to the near-racemic character of the knipholone anthrone sample obtained by reduction of (P )-knipholone (Figure 10.15d), additional evidence of the identity of the

376

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

P

P

M

P

t [min]

t [min]

t [min] 30 40

30 40

Near-racemate 45:55

Pure + ‘racemate’

30 40 Pure Knipholone anthrone

LC-UV

M

Figure 10.16. Spiking LC-CD t [min]

t [min] 30 40

t [min] 30 40

30 40

experiments to prove the identity of peaks, exemplarily for knipholone anthrone (7) of different enantiomeric ratios from different sources.

respective peaks, and thus of the reversed HPLC behavior of 7 in comparison to 6, was acquired by a spiking experiment by adding enantiomerically pure knipholone anthrone to that near-racemic sample of 7 with subsequent chromatographical analysis by using HPLC-UV and HPLC-CD (Figure 10.16). This experiment showed that after addition of pure (P )-configured knipholone anthrone to the racemate, the second-eluting peak increased in intensity. This clearly confirmed, once again, that the elution order for the diastereomers of 7 is inversed as compared to the chromatographic behavior of compound 6. 10.9.2.4. Knipholone and Knipholone Anthrone: Instructive Examples of the Manifold Aspects of HPLC-CD. The investigations on knipholone (6) and knipholone anthrone (7) exemplify several advanced aspects of HPLC-CD hyphenation: A general disadvantage might be that the solvent used for the HPLC-CD measurement may influence the experimental CD curves significantly, although such effects were shown to be negligible in the case of 6 and 7. In offline CD measurements, such aggregation phenomena can be evaluated by simply using different solvents. In online CD investigations, by contrast, hints at possible spectral changes originating from aggregated species can be obtained from a dilution series. Such a “dilution experiment” can sometimes be achieved by simply measuring the CD spectrum of the investigated compound at different positions of its peak in the UV-monitored chromatogram (e.g., left versus right slope). Another useful lesson one can learn from the racemate resolution and the online CD analysis of knipholone (6) and knipholone anthrone (7) is that even slight structural changes may lead to an inversion of the elution order of the respective enantiomers. Thus, the absolute configuration of structurally related analogues, and, in particular, of unknown compounds can usually not be determined for sure by simple comparison of relative retention times. Beyond the elucidation of the full absolute configuration by LC-CD in combination with empirical and nonempirical methods or quantum-chemical calculations, spiking (= coelution) experiments can help to establish the elution order of stereoisomers of known substances if stereochemically pure or enriched authentic

377

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

material (e.g., from a total synthesis) is available. The chemical stability of the analyte has to be considered when performing CD investigations since chiral decomposition products can largely contribute to the overall CD response, thus possibly leading to an ambiguous or even wrong assignment of absolute configurations. Again online HPLC-CD measurements in the stopped-flow mode can be helpful to overcome such problems by selectively measuring only the correct, still structurally intact analyte (even in a mixture of a plethora of other CD-active products) and, in addition, by minimizing reactions of the analyte by exclusion of light and air oxygen within the HPLC-CD flow cell.

10.10. FURTHER EXAMPLES The value of the method is emphasized by a broad variety of further stereochemically intriguing examples from most different classes of compounds, with stereogenic centers or elements of axial or planar chirality, whose absolute configurations were established by HPLC-CD in combination with quantum-chemical CD calculations and/or by applying empirical or nonempirical rules. A few selected examples are shown in Figure 10.17, among them the 3,8 -linked biflavonoid 14 from Gnidia involucrate [67], whose absolute configuration was determined using Gaffield’s isochroman helicity rule [68, 69], and flavanthrin (8), a 9,10-dihydrophenanthrene dimer isolated from Pholidota chinensis [70], which was structurally elucidated by HPLC-CD in combination with quantum-mechanical

Me

SMe CN

R

NH OH Me Phylline (9) R

OMe HO

OH M

OH Flavanthrin (8)

HO

A synthetic quateraryl, 10

MeO calcd. S NH N H CCl3 (S)-TaClo (11)

OH

calcd.

O R S

t

Ph

Ph

HO Ph

N

N

N

Ph

Ph

β,β '-Coupled bisporphyrins, 12

N

Zn

N N

Ph

R

P

P

M

R

H

R

R

Ph

P

H

H OH O O

N Zn

OMe

exp.

exp.

N

Me OMe

M

HO R

Ph R = CH2OH

OH R

O

OH

OH A biflavonoid, 14

A configurationally semi-stable bi[10]paracyclophane, 13

Figure 10.17. Selection of structurally most diverse compounds with one or more stereogenic centers, axes, and/or elements of planar chirality from different research groups, stereochemically investigated by HPLC-CD hyphenation techniques.

378

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CD calculations. Further configurational assignments succeeded with synthetic quateraryls like 10 [71], the neurotoxin TaClo (11) [72], and the anti-malarial drug artemisinin [73], and all were investigated through the hyphenation of HPLC and CD. Some additional selected applications of HPLC-CD hyphenation will be highlighted in the following pages, involving online CD investigations on axially chiral bisporphyrins like 12, likewise synthetic paracyclophanes of type 13, and the natural alkaloid phylline (9) as illustrative examples.

10.10.1. The Stereostructure of Intrinsically Axially Chiral β,β  -Bisporphyrins The utility of the online CD analysis in combination with quantum-chemical CD calculations for the fast configurational assignment of chiral compounds has also been succesfully applied to unnatural, merely synthetic compounds possessing unprecedented stereostructures—for example, to the stereochemical characterization of the first intrinsically axially chiral bisporphyrins with a rotationally hindered direct β,β  -linkage [74, 75]. Among a whole series of β,β  -linked porphyrin dimers of type 12, several racemic free-base representatives could not be resolved, although a variety of different chiral HPLC phases were tried under various conditions [75]. By contrast, a couple of fully metalated dimeric porphyrins like, for example, rac-12a, rac-12b, and rac-12c (Figure 10.14) gave a clear baseline separation of the respective atropo-enantiomers at room temperature after extensive optimization of the separation conditions (Chirex-3010 column, n-hexane/CH2 Cl2 60/40). Note that for the online LC-CD measurements the CD detection wavelength was set to 435 nm, that is, red-shifted by about 10 nm as compared to the UV maximum of the racemate (as outlined above), since the CD spectrum was expected to arise from an exciton coupling of the two adjacent identical chromophores. The full online CD spectra were perfectly mirror-imaged and showed a positive first Cotton effect around 450 nm for the faster enantiomer (Peak A) and a negative one for the more slowly (Peak B) eluting atropisomer (Figure 10.18). The absolute configurations at the biaryl axis of the enantiomers (P )-12a and (M )-12a were established by HPLCCD experiments in the stopped-flow mode in combination with quantum-chemical CD calculations (ZINDO/S-CI//BLYP-D/SVP), revealing the (P )-atropo-enantiomer to be the faster eluting one. As expected, the CD spectra of 12b and 12c (differing from 12a in the metalation pattern and/or the meso-substituents) were strongly related to those of 12a, permitting, in this case, configurational assignment by comparison of these curves with the ones calculated for the enantiomers of the parent dimer 12a, (P )-12a, and (M )-12a. This clearly evidenced that the chromatographically faster atropo-enantiomers of both, 12b and 12c, were (P )-configured and that, consequently, the slower ones had the M configuration at the porphyrin–porphyrin axis, here proving that the substitution pattern and, in this case, also the type of the central metals may have only minor effects on the overall CD spectrum. On the other hand, the fact that the palladium(II) and copper(II) derivatives of rac12a (structures not shown) revealed an inverse elution order (with the M -enantiomer being the faster eluting one), as compared to that of rac-12a, rac-12b, and rac-12c, again shows that even small structural differences can, within the same class of compounds, lead to substantial changes of the chromatographical behavior, highlighting that a configuration assignment cannot be based on elution orders alone, but must be assisted by, for example, online CD investigations.

379

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

LC-UV 425 nm

Resolution on a chiral HPLC phase (Chirex-3010®, rac-12a, CH Cl /n-hexane) 2 2 rac-12b, or rac-12c

Peak A Peak B

6

LC-CD t [min] 435 nm A

300

B

exp. for 12a 12b 12c

calcd. for (P)-12a

400

500

Online-CD (stopped flow)

6

CD [mdeg]

CD [mdeg]

Online-CD (stopped flow) 40 20 0 –20 –40 –60 –80

8 t [min]

8 t [min]

80 60 40 20 0 –20 –40

exp. for 12a 12b 12c calcd. for M-12a

600

300

wavelength λ [nm] A

Ar1 N

Ar1 N

Ar1

N

M1

Ar1 N

P

Ar2 Ar1 (P )-12a-c

Ar1

Ar2 N 2 N M N N

Ar2

400

500

600

wavelength λ [nm]

Ar2

N

N M1 N

B

N

Ar1 Ar2

Ar2

N 2N M N N

M Ar1

Ar2

Ar2

(M )-12a-c

Figure 10.18. Stereochemical characterization of the axially chiral β,β  -bisporphyrins 12a–c by online HPLC-CD measurements on a chiral phase and comparison of the experimental spectra with those obtained by quantum-chemical CD calculations [rac-12a (M1 = M2 = Zn, Ar1 = Ar2 = phenyl), rac-12b (M1 = Zn, M2 = Ni, Ar1 = Ar2 = Phenyl), rac-12c (M1 = M2 = Zn, Ar1 = Ar2 = 4methoxyphenyl)].

10.10.2. BI [10] Paracyclophane: An Axially Chiral, Yet Configurationally Semi-Stable Biphenyl Meso-compounds are constitutionally symmetric molecules that do possess pairs of stereogenic elements, but of opposite configurations each. They are, thus, achiral—at least on the time average: Even if possessing chiral conformations (maybe even exclusively), these may rapidly interconvert in flexible systems, so that the molecule will, macroscopically, appear as achiral above a certain temperature [76–78]. The bi[10]paracyclophane 13 (Figure 10.17) was synthesized by Tochtermann and co-workers [79, 80] since it was expected to constitute an unprecedented borderline case between an achiral meso-compound and an axially and planar chiral compound. It has two elements of planar chirality, whose absolute configurations were known to be opposite to each other from its synthesis so that one-half of the molecule was (pP )- and the other one (pM )-configured. In this remarkable molecule, rotation at the central axis will, despite the presence of the two planar-chiral elements, lead to enantiomers, (pM , aM , pP ) versus (pM , aP , pP ), not diastereomers. Different from the situation with similar, more hindered

380

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HPLC-UV at 210 nm

HPLC-CD at 210 nm

t [min] 8 10 12 14

HPLC-CD of Peak A

HPLC-CD of Peak B 10

12

calcd. for (pP,aM,pM)-13

6

Δε [cm2/mol]

8

exp. Peak B

4

2 0 –2

0 –4 –8

exp. Peak A –6

–12 200

250

300

wavelength λ [nm]

350

–10 200

calcd. for (pP,aP,pM)-13 250

300

350

wavelength λ [nm]

Figure 10.19. Resolution of an enantiomeric mixture of 13 by HPLC and determination of the absolute configuration by comparison of online CD spectra with the calculated CD curves.

(and thus C1 symmetric) analogues investigated earlier [80], the axis connecting these two moieties was configurationally semi-stable and, as a consequence, 13 should be a genuine meso-compound at higher and a “simple” C1 -symmetric compound at lower temperature. Due to the observed atropisomeric interconversion at room temperature, LC-CD was the method of choice for the investigation of the stereochemical properties of 13. Apparently due to the low rotational barrier at the central axis, the resolution was most unsatisfactory at ambient temperature, making it necessary to perform the separation at lower temperature: The best results were obtained using a Chiralcel ODRH column (Daicel) at 5◦ C and with acetonitrile/water (68:32) as the mobile phase. That the two peaks thus observed (Figure 10.19) indeed corresponded to the expected atropo-enantiomers of 13 was evidenced by their online CD analysis, which resulted in almost opposite CD spectra. It is noteworthy that although the two peaks of the enantiomeric mixture were not baseline separated, it was possible to obtain CD spectra of good quality. One reason is that, different from CD spectra of diastereomers, those of enantiomers are fully opposite and retain their qualitative appearance even if the sample is not enantiomerically pure (see also Section 10.4). Therefore, even enantiomers with nearly identical retention times, which thus give only one—seemingly unresolved—LCUV peak, may sometimes give full CD spectra by using the HPLC-CD method [34, 35].

381

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

The chromatographic separation of the two enantiomers of 13 and their online CD measurement permitted assignment of their absolute configurations by quantum-chemical CD calculations (CNDO-S-CI//AM1 method) and comparison of the computed CD spectra with the experimental curves. As can be seen in Figure 10.19, the experimental CD curve of the more rapidly eluting Peak A showed a good agreement with the spectrum calculated for the (pP , aM , pM )-enantiomer of 13, while that of the slower Peak B matched well with the CD spectrum predicted for (pP ,aP ,pM ), thus permitting an unambiguous assignment of the two peaks, A and B, to the corresponding enantiomers. As another powerful application of the LC-CD hyphenation, the decrease of the CD curve of Peak A was monitored directly after resolution for the determination of the half-life (t1/2 ) of the racemization process at the biaryl axis of 13. Because of the relatively unstable axial configuration and, therefore, fast vanishing of the CD effect, the time for scanning the CD spectrum had to be reduced dramatically. This was achieved by minimizing the spectral width from 200–230 nm down to 30 nm. On the basis of these experiments, a t1/2 of about 70 s at room temperature was roughly estimated, which fitted quite well with the value obtained from previous NMR experiments [79].

10.10.3. Phylline, Structure Elucidation Directly from the Crude Extract Phylline (9) is a nice example of the application of the analytical triad LC-MS/MS-NMRCD. The HPLC-UV chromatogram of an extract of the rare tropical liana Habropetalum dawei from Sierra Leone showed several peaks (Figure 10.20, left), some of them corresponding to known compounds such as dioncophylline A, but also hinting at the presence

Extract of H. dawei LC-UV at 266 nm

mV 30

new compound

20

mV positive CD effect

30

LC-CD at 266 nm

20 10

x2

10

0

0 0

10

20

30

t [min]

0

10

20

30

t [min]

LC-MS/MS-NMR Me trans NH OH

Me

Constitution, relative configuration

Me R 3 R NH 1

OH Me Natural Phylline (9)

Me S S NH

OH Me Synthetic ent-Phylline (1S,3S-9)

Absolute 1R,3R-configuration

Figure 10.20. Structure elucidation of natural phylline (9) directly from the crude extract by using the analytical triad LC-MS/MS-NMR-CD.

382

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of new natural products. According to HPLC-MS/MS and HPLC-NMR measurements, one of them was the naphthalene-devoid tetrahydroisoquinoline part of dioncophylline A, thus later simply named phylline [7]. Online ROESY experiments evidenced that the two methyl groups were trans to each other, so that natural phylline had to possess either the 1R, 3R- or the 1S , 3S -configuration. From a synthetic S , S -configured reference (ent-9) [81], it was known that the R, R-enantiomer should have a positive Cotton effect at 266 nm. Thus, already from the HPLC-CD chromatogram as detected at a single wavelength (266 nm), the 1R, 3R-configuration was deduced for the natural phylline since the corresponding peak showed a clear positive signal in the LC-CD trace (Figure 10.20, right).

10.11. CONCLUSIONS Due to the nowadays commercial availability and the reduced costs of HPLC-CD systems, the hyphenation of HPLC with CD becomes more and more interesting for industrial and academic research. While LC-CD is still mainly used for quality management in the pharmaceutical industry, it has several further substantial advantages in other research areas. In combination with HPLC-MSn and HPLC-NMR, it is a time- and work-saving method for the structure elucidation of novel natural products. Crude extracts of plants, bacteria, or fungi can be directly analyzed without the need of previous isolation work. Thus, already known products will easily be disregarded and no further time and money are wasted for unnecessary isolation work and one can focus on truly new substances or even new classes of compounds. In addition, chemically or configurationally unstable or semi-stable substances can more easily—that is, directly—be analyzed by these methods. Yet, CD spectroscopy is, in general, still largely neglected in chemical and biochemical studies and we, therefore, plead for the urgently needed change of mind here. Especially by the help of quantum-chemical calculations, it is nowadays quite easy to interpret the CD spectra of novel compounds, even with unknown, unprecedented chromophores. This has further enhanced the value of CD as an efficient tool for the elucidation of absolute configurations, which is true not only for ECD but also for vibrational circular dichroism (VCD). Unfortunately, due to detection limits and problems with eluent absorption, it is still not possible to hyphenate HPLC with VCD measurements, which would be a very desirable additional tool for the determination of absolute configurations of new structures, especially for substances that lack suitable UV-active chromophores. Another promising combination of techniques could be HPLC-ORD [82–84], but here are again some problems that have to be solved before. The quality of these measurements in many cases is not sufficient (i.e., because of the still too high detection limit). An additional problem is the unknown concentrations of substances during HPLC runs. Of course, there is the possibility to measure HPLC-CD instead of HPLC-ORD, and convert the spectra by the Kramers–Kronig transformation, but without knowing the concentrations the results are not very helpful. Furthermore, the rotatory power α is known to be strongly influenced by the eluent composition, which hampers applications of HPLC-ORD hyphenation [85].

10.12. SOFTWARE RECOMMENDATIONS There are several possibilities to analyze and visualize the results of experiments and calculations in this field. Normally, every software delivered from the vendor for HPLC

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

systems or spectrometers is capable of exporting chromatograms or spectra as pure ASCII files (often with the *.txt file ending). These files can be imported either into commercially available software or can be processed with free software. Among the numerous commercially available software for the Windows® OS, the authors only have experience with Excel from Microsoft® and Origin from OriginLab® and thus cannot give a full overview over all available program packages. Comparing Excel and Origin, we would recommend Origin since its main focus lies on the data analysis of work of scientists and engineers. Alternatively, in the following a few examples of freely available software will be given, showing that the preparation of publishable graphics does not necessary require expensive software tools. For the processing of statistical or scientific results Gnuplot (http://www.gnuplot.info) is one of the best-known free tools and is available for several platforms, including Linux and Windows. Unfortunately, it is not an easy-to-use software, because it is command-line driven, but there are third-party tools that make use of the high-quality plotting capabilities of Gnuplot. Knowing that experimental data can be plotted with Gnuplot, there is still the need to get spectra from the output of quantum-chemical calculations in a format that is usable for plotting software. Thus, one needs a program that can do a Gauss or Lorentz curve generation and afterwards gives readable files. The quantum-chemical package ORCA (http://www.thch.uni-bonn.de/tc/orca) delivers a tool (orca_mapspc) to generate these curves and to produce files that can be used for further plotting. Using Gaussview, curves of UV and CD spectra can be obtained from Gaussian calculations (http://www.gaussian.com). Gaussum (http://gausssum.sourceforge.net) or Gabedit (http://gabedit.sourceforge.net) are two examples that can do this work as well and will then export the curves as text files that can be used for plotting with other software. Both are available for different platforms (e.g., Windows or Linux) and can handle the results of Gaussian, ADF (http://www.scm.com), and GAMESS (http://www.msg.chem.iastate.edu/gamess). In addition, Gabedit can in parts read the results of ORCA and QChem (http://www.q-chem.com) calculations. However, the results here have to be prepared “manually” for Gnuplot, and thus it is quite laborious to get publishable graphics of comparisons of experimental and calculated UV or CD curves. An alternative is the freely available software SpecDis as developed in our lab (http://www-organik.chemie.uni-wuerzburg.de/lehrstuehlearbeitskreise/bringmann/ specdis). It can directly compare results from UV/CD calculations performed with Gaussian, TURBOMOLE (http://www.turbomole.com), ORCA, or DFT/MRCI with experimental results in the ASCII format. The only drawback of SpecDis is that it is currently only available for Windows. However, it can be used to determine a UV shift or to do a Boltzmann weighting, and it generates files that can be processed with Gnuplot to get publishable graphics in the eps format. SpecDis can also handle chromatogram data files (in ASCII format). All graphics in this chapter that show spectra or chromatograms were initially prepared with SpecDis and Gnuplot.

REFERENCES 1. A. Gergely, J. Pharm. Biomed. Anal . 1989, 7 , 523–541. 2. A. F. Drake, J. M. Grould, S. F. Mason, J. Chromatogr. A 1980, 202 , 239–245. 3. P. Salvadori, C. Rosini, C. Bertucci, J. Org. Chem. 1984, 49 , 5050–5054.

383

384

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4. G. Brandl, F. Kastner, R. Fritsch, H. Zinner, A. Mannschreck, Monatsh. Chem. 1992, 123 , 1059–1069. 5. A. Mannschreck, Chirality 1992, 4 , 163–169. 6. G. Bringmann, C. G¨unther, J. Schlauer, M. R¨uckert, Anal. Chem. 1998, 70 , 2805–2811. 7. G. Bringmann, K. Messer, M. Wohlfarth, J. Kraus, K. Dumbuya, M. R¨uckert, Anal. Chem. 1999, 71 , 2678–2686. 8. G. Bringmann, K. Messer, K. Wolf, J. M¨uhlbacher, M. Gr¨une, R. Brun, A. M. Louis, Phytochemistry 2002, 60 , 389–397. 9. G. Bringmann, M. Wohlfarth, H. Rischer, J. Schlauer, R. Brun, Phytochemistry 2002, 61 , 195–204. 10. K. Iwasa, W. Cui, M. Sugiura, A. Takeuchi, M. Moriyasu, K. Takeda, J. Nat. Prod . 2005, 68 , 992–1000. 11. W. Cui, K. Iwasa, M. Sugiura, A. Takeuchi, C. Tode, Y. Nishiyama, M. Moriyasu, H. Tokuda, K. Takeda, J. Nat. Prod . 2007, 70 , 1771–1778. 12. K. Iwasa, W. Cui, T. Takahashi, Y. Nishiyama, M. Kamigauchi, J. Koyama, A. Takeuchi, M. Moriyasu, K. Takeda, J. Nat. Prod . 2010, 73 , 115–122. 13. Y. Zhang, D.-R. Wu, D. B. Wang-Iverson, A. A. Tymiak, Drug Discov. Today 2005, 10 , 571–577. 14. O. McConnell, A. B. II, C. Balibar, N. Byrne, Y. Cai, G. Carter, M. Chlenov, L. Di, K. Fan, I. Goljer, Y. He, D. Herold, M. Kagan, E. Kerns, F. Koehn, C. Kraml, V. Marathias, B. Marquez, L. McDonald, L. Nogle, C. Petucci, G. Schlingmann, G. Tawa, M. Tischler, R. T. Williamson, A. Sutherland, W. Watts, M. Young, M.-Y. Zhang, Y. Zhang, D. Zhou, D. Ho, Chirality 2007, 19 , 658–682. 15. A. M. Krstulovic, Chiral Separations by HPLC: Applications to Pharmaceutical Compounds, Ellis Horwood, Chichester, 1989. 16. K. K. Unger, R. Ditz, E. Machtejevas, R. Skudas, Angew. Chem., Int. Ed . 2010, 49 , 2300–2312. 17. D. E. Drayer, Clin. Pharmacol. Ther. 1986, 40 , 125–133. 18. T. Andersson, Clin. Pharmacol . 2004, 43 , 279–285. 19. O. Zerbinati, R. Aigotti, P. G. Daniele, J. Chromatogr. A 1994, 671 , 281–285. 20. P. Salvadori, C. Bertucci, G. Ascoli, G. Uccello-Barretta, E. Rossi, Chirality 1997, 9 , 495–505. 21. H. Jiang, X. Huang, K. Nakanishi, N. Berova, Tetrahedron Lett. 1999, 40 , 7645–7649. 22. A. Kawamura, N. Berova, V. Dirsch, A. Mangoni, K. Nakanishi, G. Schwartz, A. Bielawska, Y. Hannun, I. Kitagawa, Bioorg. Med. Chem. 1996, 4 , 1035–1043. 23. Z. Rosenzweig, E. S. Yeung, Appl. Spectrosc. 1993, 47 , 2017–2021. 24. R. E. Synovec, E. S. Yeung, Anal. Chem. 1985, 57 , 2606–2610. 25. T. Nehira, C. A. Parish, S. Jockusch, N. J. Turro, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 1999, 121 , 8681–8691. 26. D. H. Turner, I. Tinoco, Jr., M. Maestre, J. Am. Chem. Soc. 1974, 96 , 4340–4342. 27. I. Tinoco, Jr., D. H. Turner, J. Am. Chem. Soc. 1976, 98 , 6453–6456. 28. T. J. Ward, B. A. Baker, Anal. Chem. 2008, 80 , 4363–4372. 29. A. Le´on, A. Olives, M. Mart´ın, B. del Castillo, J. Inclusion Phenom. Macrocyclic Chem. 2007, 57 , 577–583. 30. G. Subramanian, Chiral Separation Techniques, 2nd ed., Wiley-VCH, Weinheim, 2001. 31. J. J. Kirkland, Modern Practice of Liquid Chromatography, Wiley, New York, 1971. 32. C. Horv´ath, W. R. Melander, J. Chromatogr. Libr. 1983, 22A, 27–135. 33. D. Song, J. Wang, J. Pharm. Biomed. Anal . 2003, 32 , 1105–1112.

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

34. P. Salvadori, C. Bertucci, C. Rosini, Chirality 1991, 3 , 376–385. 35. G. Bringmann, S. Tasler, H. Endress, J. Kraus, K. Messer, M. Wohlfarth, W. Lobin, J. Am. Chem. Soc. 2001, 123 , 2703–2711. 36. T. Tokunaga, M. Okamoto, K. Tanaka, C. Tode, M. Sugiura, Anal. Chem. 2010, 82 , 4293–4297. 37. G. Liu, D. M. Goodall, A. T. Hunter, P. R. Massey, Chirality 1994, 6 , 290–294. 38. A. Berthod, S. C. Chang, D. W. Armstrong, Anal. Chem. 1992, 64 , 395–404. 39. Kaye & Laby Online, Tables of Physical & Chemical Constants. 2.1.4 Hygrometry. Version 1.0 (2005), www.kayelaby.npl.co.uk. 40. C. Bertucci, E. Domenici, P. Salvadori, J. Pharm. Biomed. Anal . 1990, 8 , 843–846. 41. P. Horv´ath, A. Gergely, B. Nosz´al, Talanta 1997, 44 , 1479–1485. 42. P. Salvadori, L. Di Bari, G. Pescitelli, in Circ. Dichroism, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 797–817. 43. G. Bringmann, J. M¨uhlbacher, K. Messer, M. Dreyer, R. Ebel, B. W. Nugroho, V. Wray, P. Proksch, J. Nat. Prod . 2003, 66 , 80–85. 44. U. Kiehne, T. Bruhn, G. Schnakenburg, R. Fr¨ohlich, G. Bringmann, A. L¨utzen, Chem. Eur. J . 2008, 14 , 4246–4255. 45. N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. 46. G. Bringmann, T. Bruhn, K. Maksimenka, Y. Hemberger, Eur. J. Org. Chem. 2009, 2717–2727. 47. P. J. Stephens, N. Harada, Chirality 2010, 22 , 229–233. 48. G. Bringmann, S. Busemann, in Natural Product Analysis: Chromatography, Spectroscopy, Biological Testing, P. Schreier, M. Herderich, H.-U. Humpf, and W. Schwab, eds., Vieweg, Wiesbaden, 1998, pp. 195–211. 49. G. Bringmann, I. Kajahn, M. Reichert, S. E. H. Pedersen, J. H. Faber, T. Gulder, R. Brun, S. B. Christensen, A. Ponte-Sucre, H. Moll, G. Heubl, V. Mudogo, J. Org. Chem. 2006, 71 , 9348–9356. 50. G. Bringmann, R. God, M. Schaeffer, Phytochemistry 1996, 43 , 1393–1403. 51. F. Bracher, W. J. Eisenreich, J. M¨uhlbacher, M. Dreyer, G. Bringmann, J. Org. Chem. 2004, 69 , 8602–8608. 52. A. D. Becke, J. Chem. Phys. 1993, 98 , 1372–1377. 53. C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37 , 785–789. 54. P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28 , 213–222. 55. M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M. S. Gordon, D. J. DeFrees, J. A. Pople, J. Chem. Phys. 1982, 77 , 3654–3665. 56. A. Dreuw, M. Head-Gordon, Chem. Rev . 2005, 105 , 4009–4037. 57. A. Sch¨afer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97 , 2571–2577. 58. A. Klamt, COSMO-RS: From Quantum Chemistry to Fluid Phase Thermodynamics and Drug Design, Elsevier, Amsterdam, 2005. 59. G. Bringmann, T. Gulder, B. Hertlein, Y. Hemberger, F. Meyer, J. Am. Chem. Soc. 2010, 132 , 1151–1158. 60. G. Bringmann, J. Mutanyatta-Comar, M. Knauer, B. M. Abegaz, Nat. Prod. Rep. 2008, 25 , 696–718. 61. E. Dagne, W. Steglich, Phytochemistry 1984, 23 , 1729–1731. 62. E. Dagne, A. Yenesew, Phytochemistry 1993, 34 , 1440–1441. 63. G. Bringmann, K. Maksimenka, J. Mutanyatta-Comar, M. Knauer, T. Bruhn, Tetrahedron 2007, 63 , 9810–9824. 64. G. Bringmann, D. Menche, Angew. Chem., Int. Ed . 2001, 40 , 1687–1690.

385

386

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

65. G. Bringmann, J. Kraus, D. Menche, K. Messer, Tetrahedron 1999, 55 , 7563–7572. 66. S. Grimme, M. Waletzke, J. Chem. Phys. 1999, 111 , 5645–5655. 67. J. Ferrari, C. Terreaux, T. Kurt´an, A. Szikszai-Kiss, S. Antus, Jerome D. Msonthi, K. Hostettmann, Helv. Chim. Acta 2003, 86 , 2768–2778. 68. W. Gaffield, Tetrahedron 1970, 26 , 4093–4108. 69. G. Kerti, T. Kurt´an, T.-Z. Illy´es, K. E. K¨ov´er, S. S´olyom, G. Pescitelli, N. Fujioka, N. Berova, S. Antus, Eur. J. Org. Chem. 2007, 2007 , 296–305. 70. S. Yao, C.-P. Tang, Y. Ye, T. Kurt´an, A. Kiss-Szikszai, S. Antus, G. Pescitelli, P. Salvadori, K. Krohn, Tetrahedron: Asymmetry 2008, 19 , 2007–2014. 71. A. Goel, F. V. Singh, V. Kumar, M. Reichert, T. A. M. Gulder, G. Bringmann, J. Org. Chem. 2007, 72 , 7765–7768. 72. G. Bringmann, D. Feineis, R. God, K. Maksimenka, J. M¨uhlbacher, K. Messer, M. M¨unchbach, K.-P. Gulden, E.-M. Peters, K. Peters, Tetrahedron 2004, 60 , 8143–8151. 73. J. Chen, J. Gu, R. Zhao, R. Dai, J. Wang, Chromatographia 2009, 69 , 361–363. 74. G. Bringmann, S. R¨udenauer, D. C. G. G¨otz, T. A. M. Gulder, M. Reichert, Org. Lett. 2006, 8 , 4743–4746. 75. G. Bringmann, D. C. G. G¨otz, T. A. M. Gulder, T. H. Gehrke, T. Bruhn, T. Kupfer, K. Radacki, H. Braunschweig, A. Heckmann, C. Lambert, J. Am. Chem. Soc. 2008, 130 , 17812–17825. 76. K. Mislow, Introduction to Stereochemistry, W.A. Benjamin, New York, 1966. 77. E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Organic Compounds, Wiley, New York, 1994. 78. R. W. Hoffmann, Angew. Chem., Int. Ed . 2003, 42 , 1096–1109. 79. G. Bringmann, T. A. M. Gulder, K. Maksimenka, D. Kuckling, W. Tochtermann, Tetrahedron 2005, 61 , 7241–7246. 80. W. Tochtermann, D. Kuckling, C. Meints, J. Kraus, G. Bringmann, Tetrahedron 2003, 59 , 7791–7801. 81. G. Bringmann, R. Weirich, H. Reuscher, J. R. Jansen, L. Kinzinger, T. Ortmann, Liebigs Ann. Chem. 1993, 877–888. 82. D. M. Goodall, Trends Anal. Chem. 1993, 12 , 177–184. 83. Z. Wu, D. M. Goodall, D. K. Lloyd, J. Pharm. Biomed. Anal . 1990, 8 , 357–364. 84. D. K. Lloyd, D. M. Goodall, Chirality 1989, 1 , 251–264. 85. J. L. Sims, J. A. Carreira, R. A. Bragg, Chirality 1997, 9 , 122–125.

11 DETERMINATION OF THE STRUCTURES OF CHIRAL NATURAL PRODUCTS USING VIBRATIONAL CIRCULAR DICHROISM Prasad L. Polavarapu

11.1. INTRODUCTION Circular dichroism in electronic transitions, referred to as electronic circular dichroism (abbreviated as ECD, and most often simply as CD), has dominated the chemical sciences in much of the twentieth century for chiral molecular structural determination [1]. The measurement of circular dichroism associated with molecular vibrational transitions is referred to as vibrational circular dichroism (VCD).1 VCD is supported by all chiral molecules, just as ECD. However, the number of molecular vibrations that can be studied in the accessible infrared region is far greater than the number of electronic transitions that can be studied in the accessible ultraviolet–visible region. Furthermore, vibrational transitions are expected to be more sensitive than electronic transitions to the conformational details of molecules. Therefore the information content that is derived through VCD spectra is anticipated to be much more detailed. Advances in quantum chemical methods (see Chapter 24 by Kenneth Ruud in Volume 1) provided reliable approaches for predicting the VCD spectra of chiral molecules [2]. Subsequent development of quantum chemical software packages [3], for calculating the VCD spectra, provided a convenient means to analyze the experimental VCD spectra. For chiral compounds of known absolute configurations, the quantum chemically predicted VCD spectra provided impressive agreements with corresponding experimentally measured VCD spectra. These findings revolutionized the outlook and applications for 1

Although circular dichroism in the infrared region was reported in 1972, the first reference to circular dichroism in vibrational transitions appeared in 1974. See: G. Holzwarth, E. C. Hsu, H. S. Mosher, T. R. Faulkner, A. Moscowitz, J. Am. Chem. Soc. 1974, 96 , 252–253.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

387

388

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O

O

O

H

COOH COOH

HO

OR1

COOH

O

OR2

H

OH

HOOC

(2S,3S)-1

H

O

O

H

O

O

(2S,3R)-2

(R)-3

(R)-4

(1R,5R)-5 H

O O O

O

O

OR H

(1R)-6

(1S)-7

(S)-8

(4S)-9

(4S)-10

(1R,2S,4R)-11 H

H 7 C N B 2 E N D H 21 20

O O

O

F O

H

A

O

O

N N

H

O

O

12

13

(R)-14

H

H N

(2R,7S,20S,21S)-18

H H N N

H3CO

O

O

(2R,7S,20S,21S)-17

7

HO

6

N

5

4

2 3

A

O 2’ R1 = bzl =

H

OH

R2O

7 6

2 H CN 21 E N D 20

O

N

20

5

CH3 1 4

O 2

R1 = R2 = Acl =

3 OR1

O

21

7

B

H3CO OH

(2R,19R,20S,21S)-19

CH3

1

H

N H

O

N H

H3CO

(2S,7R,20R,21R)-16 N

N

O

H3CO

O

H3CO

N H

H3CO

(2R,7S,20S,21S)-15

22

23

Scheme 11.1. Structures of compounds 1–23 discussed in the text.

VCD spectroscopy in the twenty-first century [4]. As a result, the number of laboratories adapting VCD spectroscopy for structural applications of chiral molecules has been increasing in recent years. This chapter will first provide a tutorial for chiral molecular structure determination using VCD and then provide a comprehensive review of the applications of VCD to chiral natural products. An earlier review on VCD spectroscopy of natural products was published in 2008 [5]. Some of the natural products discussed here are shown in Schemes 11.1–11.4.

11.2. DETERMINATION OF CHIRAL MOLECULAR STRUCTURES USING VCD: A TUTORIAL The procedure for determining the chiral molecular structures using VCD spectra is dependent on one basic criterion: If the experimentally observed VCD spectrum for an enantiomer of a chiral compound is satisfactorily reproduced by quantum chemically

389

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

O R2O

N

Sen= O

OH

Tgl=

O

OMe

H

OR1

N

O

O

(1R,3R,5S,6R)-24 [R1 = H; R1 = H]

Cl

OMe O

Ang=

O

H

N

Haemanthamine 29

(2S,3S,4aS,11S)-30

(1R,3R,5S,6R)-25 [R1 = Sen; R1 = H] (1R,3R,5S,6R)-26 [R1 = H; R1 = Tgl] (1R,3R,5S,6R)-27 [R1 = H; R1 = Sen] (1R,3R,5S,6R)-28 [R1 = H; R1 = Ang] O H

OCH3

O H

OCH 3 O H 5 C HD 89 O O B 1 A 10 O H H O 14

O

O

O O H

H

O

OCH3 H

H

H O

O H

OCH3

O

O H

O

H

O O H O

H

O

OH (1R,5S,8S,9S,10S)-32

(1R,5S,8S,9S,10S)-31

CH3

O

(1R,5S,8S,9S,10S)-33

(1R,5S,8S,9S,10S)-34 O

O O H AcO

3

OAc

RO

H 11 H

HO 1

O O (1S,11S,12S)-39

OH

HO

H

HO

O

H

H O

O

OAc

H

(1R,2R,8R,11R)-41 O

H 3 1 OH 5 10 H 6 8 O OH 11

O

O

H

(1R,2R,5S,8R,11R)-40

H H

H H

H H

2 5

OR

RO

(3S,5R,8R,9R,10S,13S,14S)-37, R = H (3S,5R,8R,9R,10S,13S,14S)-38, R = Ac

36

8

H

H

(R)-35

H

H

OAc

H

2 3 4

14

(1R,2R,5S,8R,11R)-42

14

11

1

2

10 9 5 6 78

3 4

15 O

13

12

15

11 1 10

5 6

7

13

9 8 12

O (1R,2R,4R,8R,11R)-43

44

(4R,9R,10R)-45

(4R,9S,10R)-46

Scheme 11.2. Structures of compounds 24–46 discussed in the text.

predicted VCD spectrum for only one of all possible absolute configurations, then the absolute configuration used to calculate that satisfactorily predicted VCD spectrum can be assigned to the enantiomer used for experimental measurements. It is imperative that the calculated VCD spectra be obtained using reliable quantum chemical methods.

11.2.1. Experimental VCD Spectra Historically, VCD measurements began in the near-infrared region (10,000–4000 cm−1 , where overtone vibrations appear) and then extended to the O–H, N–H and C–H stretching vibrational region (4000–2500 cm−1 ). The experimental and theoretical aspects of near-infrared VCD have been presented by Sergio Abbate and his co-workers in

390

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Br

Br 9

10

3 O H 11

CI Br

OH

Br

Br

O

OH

48

Br

14

O O

R 1O O

O OR

(S)-53 [R1 = H, R2 = CH3]

14

(S)-54 [R1 = Ang, R2 = CH3]

52

OAng OH

57 [R = H] 58 [R = Ac]

(S)-55 [R1 = Sen, R2 = CH3] (S)-56 [R1 = H, R2 = H]

HO

1

O O

5 H

AcO 59

Me

Me

OH

OH

15 Me

HO

H

10

H Me

Me 13

11 H 15 O 12 O 60

Me

Me Me

Me

1

11 15

HO

3

Me Me

5

13

HO

SCH3

R O

O

7

H H 12

12

NH

OR2

9 5

H

8

14 63

62 S

OAc

11

1

Me

61

O H

9 H

HO

14

O

6

OAc

OO 12 H 5 11 10 7 1 6 9 8

13

13

OR

15

4

2

15

(7S,10S)-50 (R = H) (7S,10S)-51 (R = Ac)

(2R,5R,5aR,8R,9aS)-49 OR2

3

4

5

1

7 8

10 9

13 (2R,5R,5aR,7S,8S,9aS)-47

2

H

12

Br

8

7

5a OH

5

Br

O H

O H

2

OR1

OAc

(4R,5S,7R,9R,10R,11R)-64

N H

H N

(4R,5S,7S,8S,10R,11R)-65 R1 = Ang; R2 = i-val (4R,5S,7S,8S,10R,11R)-66

67

68 [R = CH2CN] 69 [R = CH3]

R1 = R2 = Ac H3CS

SCH3

N

N

S

N H

OCH3 (S)-70

H

(R)-71

CH3

O OCH3

O N

O

S

S O

SCH3

N

N OCH3 (2R,3R)-72

S

O

N H (S)-73

Scheme 11.3. Structures of compounds 47–73 discussed in the text.

Chapter 10 of Volume 1. The measurements subsequently were extended to cover the heavy atom stretching (–C=O, –C–C, –C–N, etc.) and hydrogen (–X–H, X=C, N, O, etc.) bending vibrations that absorb infrared radiation in the 2000- to 900-cm−1 region. The details of VCD measurements and instrumentation have been provided by Laurence Nafie in Chapter 5 of Volume 1. With the realization that the vibrational bands in the 2000- to 900-cm−1 region are better resolved, and can be predicted reliably using quantum chemical methods, most of the structural determination studies using VCD continue to be conducted in the 2000- to 900-cm−1 region. The experimental VCD measurements are usually conducted for liquid solutions in a suitable infrared transmitting solvent. The solvents usable in the 2000- to 900-cm−1 region include CCl4 , CH2 Cl2 , CHCl3 , CH3 CN,

391

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

OCH3

O R1 O HO R3

O

O

O

O

OCH3

OCH3

R2

H3CO

81-(a) 74 [R1= H, R2 = R3 = CH3] 75 [R1 = R2 = R3 = CH3] 76 [R1 = Ac, R2 = CH3, R3 = CH3] 77 [R1 = H,R2 = C2H5,R3 = CH3] 78 [R1 = H,R2 = CH3, R3 = C2H5]

OCH3 O

HO O

O

79 [R1 = CH3,R2 = C2H5, R3 = CH3] 80 [R1 = CH3, R2 = CH3, R3 = C2H5]

HO

OCH3 81-(b)

O

81-(c)

CH2

O

O

8

OH

O

O

9 O

O

O

O

HO

7 O

5 O 18

O

84

85

(1S,2S,3R)-83 82 CH3

CH3

CH3

OH H 3C

H3C

CH3

CH3 HO

OH

HO

H3C

H2C

CH3 H 3C

H3C

(P)-87 CH3

O HO CH3

HO

CH3

O

O

H H OH HO

OH

OH

OH OH

HO

(M)-88 CH3

OH H3C

CH3

HO

H3C

86

OH

CH3

O

OH HO

O

H3C OH (P)-89

(M)-90

OH

O

(aS)-91

Scheme 11.4. Structures of compounds 74–91 discussed in the text.

(CH3 )2 SO, CH3 OH, and/or their deuterated analogues. Measurements can also be done for water solutions [6], but in such cases high concentrations and lower pathlengths are required to minimize the interference from strong H2 O absorption at 1650 cm−1 . The use of D2 O clears up the region around 1650 cm−1 , but strong absorption of D2 O at ∼1250 cm−1 will preclude the measurements around that region. It is preferable to use low sample concentrations to avoid solute aggregation effects (such as formation of dimers) and to avoid solvents that form hydrogen bonds with solute molecules. However, practical considerations such as lower solubility and smaller vibrational molar extinction

392

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

coefficients of the sample being studied may not leave much choice, in some cases, for the solvents and concentrations to be used. The enantiomers are usually labeled by the signs of their experimental optical rotations at 589 nm as (+)589 or (−)589 . The wavelength subscript is not often included, and by default these labels are written simply as (+) or (−). If both enantiomers of a chiral compound are available, then it is advisable to measure the VCD spectra for both of them. Using one-half of the difference between the VCD spectra of enantiomers, as [(+)-(−)]/2, provides better signal to noise and reduces the spectral and baseline artifacts for the resulting spectrum. As an alternative, if only one enantiomer is available along with its racemic mixture, then subtraction of the VCD spectrum of racemic mixture from that of enantiomer will reduce the spectral and baseline artefacts. If the racemic mixture is also not available, then the VCD spectrum of the solvent may be subtracted from that of the enantiomer solution, but one should be wary of the residual artifacts in such cases. Measurements on films [7], mulls, and KBR pellets [8] are also possible, but one should take extreme care in minimizing the spectral artifacts. Unless the investigators can recognize and minimize the spectral artifacts, VCD measurements in solid state are not recommended. To undertake the VCD measurements, the needed sample concentration and cell pathlength should be determined by measuring the infrared (IR) vibrational absorption (VA) spectrum of the sample first. A spectral band, for which VCD measurement is needed, should not have absorbance greater than ∼1.0. The optimal range for the absorbance of a band of interest is ∼0.3–0.6. The presence of a VCD band where there is no corresponding absorption band may indicate a possible spectral artifact, so the experimental VCD spectrum should be presented along with the corresponding VA spectrum, preferably one above the other, both on the same wavenumber scale. Representative experimental VCD and absorption spectra are shown in Figure 11.1.

11.2.2. Quantum Chemical VCD Predictions Density functional theoretical (DFT) method is the preferred choice for VCD predictions [9], and the B3LYP functional [10] is the most often used functional. The smallest basis set to be preferred is 6-31G(d) (also labeled as 6-31G*) [11]. The use of any other smaller basis set must be calibrated against molecules with known absolute configurations. Quantum chemical calculations of VCD yield vibrational band positions and integrated VCD intensities. These numbers have to be morphed into spectra, for a comparison with the experimentally observed spectra, by associating a band shape at each of the calculated vibrational band positions. Lorentzian (or Gaussian) band shapes are normally assumed with a constant width at all band positions. Since experiments are generally performed for liquid solutions, gas-phase calculations are not true representatives of the solution-phase experimental measurements. If the solute does not form hydrogen bonds, or complexes, with solvent, then solution-phase calculations using a reliable solvation model are to be investigated. Polarizable Continuum Model (PCM) [12] has been the most widely used solvation model. The older implementations of PCM suffered [13] from discontinuities and approximations in the definition and the implementation of the free energy functional in solution and of the corresponding analytical derivatives. The recently developed continuous surface charge PCM [14] overcomes these limitations [15]. If the solute forms hydrogen bonds, or complexes, with solvent, then solute–solvent clusters are to be investigated (vide infra).

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

600 noise

1138

1254

400

Δε x 104

1119 1203

200

1065 1801 VCD

0

1034

1755

1300

–200

1227

1092

(+)-Garcinia acid dimethyl ester in CD2Cl2

–400 1900

1700

1500

1300

1100

900

cm–1 1000 (+)-Garcinia acid dimethyl ester in CD2Cl2 800

1805

1751

600 ε 400

1254 12111153 1439

200 Absorption 0 1900

1700

1308 1350

1500

1300

1068 1119 1092 1100

900

cm–1

Figure 11.1. Experimental VA (bottom panel) and VCD (top panel) spectra for (+)-garcinia acid dimethyl ester in CD2 Cl2 solvent. The noise trace at the top represents the reproducibility level in the VCD spectrum collected for 1 h.

11.2.2.1. Conformational Analysis. A reliable prediction of VCD requires a thorough conformational analysis. The signs of predicted VCD bands can vary among different conformers (with the same absolute configuration), and therefore it is important that all predominant conformations are identified. In the earlier days, conformational analysis used to be conducted manually by constructing the structures of different possible conformers and optimizing their geometries. This laborious task can now be automated, although without certain risk, using commercial conformational analysis programs, which include SPARTAN [16], CONFLEX [17], HYPERCHEM [18], and MACROMODEL [19]. These programs generate conformers by rotating atoms around different bonds or by wagging chosen atoms, and they use molecular mechanics force constants to optimize their structures. These rotations can be selected systematically (using rules predefined within the software) or randomly (Monte Carlo method). When these programs are used, it is important to be aware of the restrictions built into these programs, which vary among the programs. In addition, molecular mechanics force constants are associated with classically defined bonds (single, double, etc.), and rotations can occur only around single bonds. Therefore the starting structural definitions that are entered as input into

393

394

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the molecular mechanics-based conformational analysis programs are to be identified with classical bonds.2 The energies calculated with molecular mechanics methods are not accurate, but as long as a large energy window is used to identify the possible conformations, the conformers generated with these commercial programs will serve as a reasonable set of starting conformers for further geometry optimizations using DFT theory. However, it should be noted that some conformational analysis programs do not recognize, by default, the variation of five-member-ring puckering angles as a viable option for the conformational search. The default use of such programs must therefore be supplemented with manual construction of additional conformers and/or relaxed potential energy scans (PESs) at least at the B3LYP/6-31G∗ level, to examine if any conformations are missed by the conformational analysis programs. During a relaxed PES, a desired dihedral (or torsion) angle can be changed in certain increments and the rest of the structural parameters optimized to minimize the energy. A trough in the plot of energy versus dihedral angle will indicate a stable conformer. It is recommended that the options employed in the use of a chosen conformational analysis program be stated clearly, so that the reader can assess the completeness of the reported conformational search. 11.2.2.2. Molecules with a Single Source of Chirality. If the molecule under investigation has a single source of chirality (only one chiral center or only one helical segment), then one of the two possible absolute configurations is chosen for the theoretical studies (because calculated VCD spectra for the two possible configurations are mirror images of each other). For that chosen configuration, a set of low-energy conformations are identified as described above. Then the geometries of these conformers need to be re-optimized at least at the B3LYP/6-31G(d) level of theory. Based on the internal energies at these optimized energies, a better idea of low-energy conformers in the gas phase can be gained. Then a VCD calculation for the gas-phase predicted conformers, at these optimized energies, is performed. The absence of imaginary vibrational frequencies indicates that the conformers used are stable. These stable conformers may then be used for simulating the VCD spectrum. On the other hand, the presence of one or more imaginary vibrational frequencies indicates that the conformation used is an intermediate or a transition state. Such conformations are not used for VCD spectral simulation. The relative populations of stable conformers are calculated using Gibbs free energies obtained in the VCD calculations. Then a population-weighted VCD spectrum is obtained from the individual conformer VCD spectra and populations of stable conformers. The predicted VCD spectrum is referred to as the gas-phase predicted VCD spectrum. If the solvent used for experimental measurements does not form hydrogen bonds, or react, with solute molecules, then quantum chemical geometry optimizations mentioned in the previous paragraph are to be repeated with an appropriate solvation model to account for solvation influences and are referred to as solution-phase predicted conformers. VCD predictions for the solution-phase predicted conformers, using a solvation model, are used to generate population-weighted solution-phase predicted VCD spectrum. Note that the relative energy ordering of conformers predicted for the gas phase and the solution phase can be different, so population of a given conformer can change from the gasphase prediction to the solution-phase prediction. Most of the literature studies are based 2

The geometry files created by quantum mechanical programs do not necessarily identify the chemical bonds as single, double etc; or may identify the bonds as single, double etc using predefined standards. Such files have to be edited before importing into conformational search programs.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

on the contention that the gas-phase predictions can be compared to the solution-phase experimental measurements. This choice is clearly inadequate, and this practice should be avoided . If the solvent used for experimental measurements forms hydrogen bonds, or reacts, with solute molecules, then quantum chemical geometry optimizations mentioned in the previous paragraphs are to be undertaken for solute–solvent clusters [20a] or for resulting reaction complexes. For such cases the use of solvation models is inappropriate. 11.2.2.3. Molecules with Multiple Chiral Centers or Multiple Sources of Chirality. For a molecule with n chiral centers (or a mixture of chiral centers and helical segments), there will be 2n diastereomers, one-half of which are the enantiomers of the other half. Therefore one must undertake theoretical investigations for 2n−1 diastereomers. The conformational analysis, population-weighted gas-phase, and/or solution-phase VCD spectral predictions (as mentioned for molecules with single source of chirality) must now be undertaken for all 2n−1 diastereomers.

11.2.3. Absolute Configuration Determination The VCD spectra calculated for different sets of configurations (that describe all chiral centers and helical segments) are compared with the experimental VCD spectrum of an enantiomer. Calculated VCD spectrum, which satisfactorily reproduces the experimental VCD spectrum, identifies the absolute configuration of the enantiomer as that employed in that calculated spectrum. To emphatically state that a chiroptical spectroscopic method unequivocally determined the correct absolute configuration, it is necessary to predict the appropriate spectra for all possible diastereomers (each with appropriate low-energy conformations) and demonstrate that all but one configuration do not yield spectra that can agree with the experimental spectra. Such studies, although computationally demanding, are not emphasized in the literature but should be required. The abovementioned VCD spectral comparisons should be accompanied by the corresponding VA spectral comparisons for many reasons. If the predicted absorption spectrum does not match the observed absorption spectrum, then the corresponding VCD comparison carries less merit. Representative predicted spectra are shown in Figure 11.2.

11.3. APPLICATIONS TO CHIRAL NATURAL PRODUCTS 11.3.1. Carboxylic Acids Carboxylic acids, and molecules with O–H functional groups, often exist as hydrogenbonded complexes, due to the possibility of intermolecular hydrogen bonding between solute molecules or with solvent molecules. Theoretical prediction of chiroptical spectra for such compounds becomes very labor-intensive due to the possibility of formation of intermolecular aggregates [20b]. To avoid these aggregates and associated conformational issues, corresponding esters (or appropriate derivatives) are used for experimental studies. A few chiral carboxylic acids with biological function are found in nature. 11.3.1.1. Garcinia Acid (1) and Hibiscus Acid (2). Garcinia Cambogia (tamarind fruit) is used as a medicine for weight loss and for controlling the appetite in South Asian countries. The skin of the tamarind fruit is rich in 2-hydroxycitric acid, which was isolated as (2S ,3S )-tetrahydro-3-hydroxy-5-oxo-2,3-furandicarboxylic acid

395

396

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

600

400

(2S,3S)-Garcinia acid dimethyl ester B3LYP/aug-cc-pVDZ-PCM

Δε x 104

1264

1141

200

1196

1813 0

–200

1029 1094

1311

1758

1235 –400 1900

1700

1500

1300

1100

900

cm–1 1000

1816 1757 (2S,3S)-Garcinia acid dimethyl ester

800

B3LYP/aug-cc-pVDZ-PCM

600 ε

1257 400

1096 1066

1300 200

0 1900

1453

1700

1215

1359

1500

1159

1124 1300

1100

900

cm–1

Figure 11.2. The conformer population weighted calculated VA (bottom panel) and VCD (top panel) spectra for (2S,3S)-garcinia acid dimethyl ester at the B3LYP/aug-cc-pVDZ level. The incorporation of CD2 Cl2 solvent influence, using the PCM model as implemented in Gaussian 09 program, makes the two predicted carbonyl stretching band positions at 1816 and 1757 cm−1 match well with the corresponding experimental band positions at 1805 and 1751 cm−1 .

or garcinia acid, 1 (see Scheme 11.1). Its diastereomer, (2S,3R)-tetrahydro-3-hydroxy-5oxo-2,3-furandicarboxylic acid, or hibiscus acid, 2, is extracted from the Roselle plant, which is a common ingredient in many herbal tea blends. The known absolute configurations of these acids can be used to test the predictive abilities of VCD spectroscopy. For this purpose, the combined experimental and theoretical VCD spectral investigations were undertaken on the corresponding dimethyl esters. Each of these dimethyl esters can exist in eight different conformations. The experimental VCD spectrum of garcinia acid dimethyl ester (GADE) (Figure 11.1), with positive optical rotation at 589 nm, was reproduced by the population-weighted predicted VCD spectrum of garcinia acid dimethyl ester with known (2S ,3S ) configuration [15] at the B3LYP/aug-cc-pVDZ-PCM level (Figure 11.2). Similarly, the experimental VCD spectrum of the dimethyl ester of hibiscus acid (HADE), with positive optical rotation at 589 nm, was reproduced by the population weighted predicted VCD spectrum of hibiscus acid dimethyl ester with known (2S ,3R) configuration [20c] at B3LYP/aug-cc-pVDZ-PCM level. However,

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

as mentioned earlier, the experimental VCD spectra have to be compared to the calculated VCD spectra for all stereoisomers. When this was done, it was noted [20c] that the experimental VCD spectra for (+)-GADE can be correlated to the calculated VCD spectra of both (2S,3S ) and (2R,3S ) stereoisomers. Based on the comparison of experimental and calculated absorption spectra, however, only (2S,3S ) could be correlated with (+)-GADE. Similarly, the experimental VCD spectra for (+)-HADE can be correlated to the calculated VCD spectra of both (2S,3R) and (2R,3R) stereoisomers. Again, based on the comparison of experimental and calculated absorption spectra, only (2S,3R) could be correlated with (+)-HADE. Since chemical conversion of acids into esters does not influence the absolute configuration, the absolute configurations of the parent acids are inferred to be same as those of corresponding dimethyl esters. 11.3.1.2. Hexylitaconic Acid and Its Methyl Esters [21]. Although most natural products exist as single enantiomers, both enantiomers of hexylitaconic acid (3 with R1 = R2 = H) are found in nature: (+)-Hexylitaconic acid was isolated from Aspergillus niger, and (−)-hexylitaconic acid was derived from marine endophytic fungus Apiospora montagnei . A comparison of the experimental VCD spectra of both enantiomers of hexylitaconic acid methyl ester (3 with R1 = R2 = CH3 ) (in CDCl3 ) with the population-weighted predicted VCD spectrum at the B3PW91/6-311+G(d , p) level indicated the absolute configuration to be (R)-(−). In these predictions it was assumed that the C–C–C–C dihedral angles in alkyl carbon chain are 180◦ . The populationweighted spectrum was generated from 32 low-lying conformers that were identified using a conformational analysis program [17]. To reduce the conformational uncertainties, the diester was converted to a cyclic lactone, hexyl-3-methylenedihydrofuran-2(3H )-one (4), with a more rigid structure. A similar analysis of the experimental and predicted VCD spectra of this lactone suggested the absolute configuration of lactone to be (+)-(R). Stereochemical transformation from (−)-hexylitaconic acid mono methyl ester to (−)hexylitaconic acid dimethyl ester and from (−)-hexylitaconic acid mono methyl ester to (+)-hexyl-3-methylenedihydrofuran-2(3H )-one was used to confirm the assignment of (R)-(−) hexylitaconic acid dimethyl ester and, as a consequence, of the parent acid as (R)-(−) hexylitaconic acid.

11.3.2. Monoterpenes and Derivatives VCD spectra of some monoterpenes have been used to test the instrumental performance and to investigate normal mode versus local mode behavior [22, 23]. VCD spectra of some other monoterpenes have been used as test cases to assess the predictive abilities of, or for determining their molecular structures with, VCD spectroscopy. 11.3.2.1. α-Pinene (5), Camphor (6), and Fenchone (7) [24–28]. α-Pinene and camphor have been the widely used chemicals for VCD studies, starting from very early stages. Because of the large VCD signals observed for these molecules, they have been used for checking the correct function of the VCD instruments and for calibration purposes. Since the absolute configurations of these molecules are known for a long time, these molecules have also been used as standards for evaluating the reliability of quantum theoretical levels for VCD predictions. Based on the comparisons between experimental and quantum chemical predictions of VCD spectra, Hartree–Fock theoretical level was considered to be inadequate for VCD predictions. Similarly, the use of DFT (with functional such as B3PW91 or B3LYP) with 6-31G∗ basis set was considered to be

397

398

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

minimal level for placing any reliability on VCD predictions [24, 25]. Excellent agreement between B3PW91/6-31G∗ -predicted VCD spectra and experimental VCD spectra for (1R)-α-pinene and (+)-α-pinene, (1R)-camphor and (+)-camphor, and (1S )-fenchone and (+)-fenchone indicated, in as early as 1996, that the absolute configurations of chiral natural products can be predicted reliably using combined experimental and quantum chemical VCD studies. 11.3.2.2. Carvone (8) [29]. Conformational analysis of 8, at the B3LYP/cc-pVDZ level, indicated that there are six conformers for this molecule with three conformations being predominant. These dominant conformations have isopropenyl group in equatorial position and differ in the relative orientation of the isopropenyl group with respect to the cyclohexenone ring. The population-weighted predicted VCD spectrum for (S )-carvone, at the B3LYP/cc-pVDZ level, was considered to match well with the experimental VCD spectrum of (+)-carvone liquid, thereby suggesting that VCD can be used for determining the absolute configurations. It was also pointed out that, while ordinary IR and Raman spectra may not be able to distinguish the conformers, VCD spectral analysis helps in identifying the conformations present in liquid solutions. 11.3.2.3. Limonene Oxide (9) [30]. 9 is an atmospheric pollutant resulting from the oxidation of other terpenes (limonene, α-pinene, etc.). Conformational analysis at the B3LYP/cc-pVDZ level indicated that there are a total of 12 conformers, with six of them containing isopropenyl group in equatorial position that account for 97% population. Three of them have the oxirane ring in trans, and the other three in cis, orientation with respect to the hydrogen atom at the chiral carbon (to which isopropenyl group is attached). The lowest-energy conformer has the oxirane ring in trans orientation. By comparing the experimental IR, Raman, and VCD spectra of the liquid sample with the corresponding population-weighted gas-phase predicted spectra, at the B3LYP/cc-pVDZ level, it was concluded that five of the abovementioned conformers are present in the liquid state, with two trans conformers accounting for ∼50% population. 11.3.2.4. Perilladehyde 10 [31]. This monoterpene with aldehyde functional group has 12 different conformations. Isopropenyl group can be in equatorial or axial position, with three orientations of isopropenyl group in each position. Furthermore, aldehyde group can take two different orientations with respect to the cyclohexene ring. Three equatorial conformers are the most stable ones among the conformers investigated. The population-weighted gas-phase predicted VCD spectrum at the B3LYP/cc-pVDZ level for (S )-configuration was compared with the experimental VCD spectrum of (−)perillaldehyde liquid sample. Based on this comparison, the perillaldehyde molecules in liquid sample are suggested to exist in three conformers with (S )-(−) absolute configuration. 11.3.2.5. endo-Borneol (11 with R = H) and Its Derivatives [32]. While the absolute configuration of 11 (with R = H) has been known for a long time, the issue of how many predominant conformations this molecule has could only be answered through a reliable conformational analysis. The O–H group can rotate around its connecting bond to C atom, which yields at least three possible conformers. This conformational freedom can be restricted by chemically converting the O–H to methoxy (O–CH3 ), acetyloxy (O–COCH3 ), t-butoxy (O-C(CH3 )3 ), or trimethyl silane (O–Si(CH3 )3 ) derivative. The influence of this derivatization on structural predictions using VCD has been investigated

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

[32]. The experimental IR and VCD spectra of the O–H, acetyloxy, t-butoxy, or trimethyl silane derivatives were considered to compare well with the corresponding predicted spectra at the B3LYP/TZ2P and B3PW91/TZ2P levels of theory, for the conformations identified. Through this comparison, the (1R, 2S , 4R)-(+)-configuration was confirmed, and it was further concluded that chemical derivatization of hydroxyl groups increases the conformational rigidity and eliminates intermolecular hydrogen bonding. 11.3.2.6. Myrtenal 12 [33]. This molecule has a rigid cyclic structure with conformational mobility coming from the aldehyde group. Even though two orientations of aldehyde group are possible, the trans orientation has nearly 99% population. The small molecular size and conformational rigidity prompted VCD spectroscopic investigations with a view to evaluate the predictions with different size basis sets. The experimental VCD spectra were measured for (−)-12 in CCl4 (6.1 mg/150 μL). Theoretical VCD predictions were obtained for (1R)-12 using a B3LYP functional and 6-31G*, 6-31+G ∗∗ , 6-311+G∗∗ , and DGDZVP basis sets. Another calculation using B3PW91 functional and DGTZVP basis set was also undertaken. After a comparison of all these predictions with experimental VCD spectra, it was concluded that the absolute configuration of (−)-12 is (1R) and that there were no significant differences of concern among these different levels of predictions and that computationally less intensive calculation at the B3LYP/DGDZVP level provides computationally economical choice without compromising the predictive abilities. 11.3.2.7. 3-Oxo-1,8-cineole or 1,3,3-Trimethyl-2-oxabicyclo[2.2.2]octan5-one (13) [34]. The pure enantiomers, (−)-13 and (+)-13, were prepared by oxidation of (+)- and (−)-3-hydroxy-1,8-cineole. The absolute configuration of (−)-3-hydroxy-1,8cineole was determined by preparing Mosher esters. The oxidation of (−)-3-hydroxy-1,8cineole and (+)-3-hydroxy-1,8-cineole with pyridinium chlorochromate yielded, respectively, (1R, 4R)-(+)-13 and (1S , 4S )-(−)-13. VCD spectroscopy was used to obtain independent evidence for the absolute configuration of 3-oxo-cineole. Conformational search for (1S , 4S )-13 indicated a single conformer. The VCD spectrum predicted for (1S , 4S )-13 at the B3LYP/DGDZVP level was considered to compare well with the experimental VCD spectrum of (−)-13, leading to the assignment of (1S , 4S )-(−)-13, which is in agreement with that obtained from Mosher ester analysis. 11.3.2.8. β-Pinene, Nopinone, Menthene, Limonene, and Menthenol, and Menthol. VCD spectra of these compounds [35, 36], except limonene, were investigated before the development of DFT methods for VCD and are not discussed here. A recent investigation [37] using experimental IR, Raman and VCD spectra of (R)-(+)-limonene, along with B3LYP/aug-cc-pVDZ calculations, suggested that three different equatorial conformers exist for this molecule. 11.3.2.9. Camphor–CDCl3 Complex [38]. For (R)-(+)-6 dissolved in achiral CDCl3 solvent, a negative VCD band was observed at 2254 cm−1 , which is associated with the C–D stretching vibration of CDCl3 . Opposite-signed VCD band was seen at the same position for (S )-(−)-6 in CDCl3 . VCD observed for the C–D stretching vibration of CDCl3 was referred to as induced VCD or chirality transfer (from chiral solute to achiral solvent). The VCD induced in CDCl3 was rationalized as due to D. . .O bond formation between D atom of CDCl3 and O atom of 6. Induced VCD however was not visible in CD2 Cl2 solvent. The optimized geometry of the 1:1 camphor–CDCl3 complex, using

399

400

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

B3LYP functional and cc-pVDZ basis set, indicated a stable structure with D. . .O distance of 2.21 A. The predicted VCD spectrum for (R)-(+)-6-CDCl3 complex reproduced the experimental observation of negative VCD for the C–D stretching mode. 11.3.2.10. Pulegone (14)–CDCl3 Complex [39, 40]. For (R)-(+)-14 dissolved in achiral CDCl3 solvent, a positive VCD band was observed at 2250 cm−1 , which is associated with the C–D stretching vibration of CDCl3 . Opposite-signed VCD band was seen at the same position for (S )-(−)-14 in CDCl3 . The IR absorption for C–D stretching vibration itself exhibited enhanced absorption intensity in the presence of 14, which indicates the presence of interaction, possibly D. . ..O bond formation between D atom of CDCl3 and O atom of 14. Induced VCD, however, was not visible in CD2 Cl2 solvent, which was associated with less acidic character of D atom in CD2 Cl2 that does not favor D. . .O bond. To explain the induced VCD in the C–D stretching vibration of CDCl3 , 1:1 complex of pulegone–CDCl3 and solvation influence (incorporated via PCM) were considered. Incorporating the 1:1 complex in solvated model was considered to yield most satisfactory predictions. When induced VCD can be measured in an otherwise achiral solvent, one would think that the relative orientations of chiral solute and achiral solvent molecules can be inferred from the experimental observations. A detailed theoretical analysis by Nicu et al. [40] suggested that such information cannot be deduced with confidence because the interaction between solute and solvent molecules is very weak. They found that the angle between electric and magnetic dipole transition moments, for the C–D stretching vibration of CDCl3 , as well as for the C=O stretching vibration of pulegone, is close to 90◦ . As a result, the VCD signs predicted for these vibrational modes depended on computational variations (such as tight geometry versus assumed or default geometry, basis set, functional, etc.), and therefore one cannot place much reliance on the predicted VCD signs of these bands.

11.3.3. Alkaloids 11.3.3.1. Cinchonidine [41]. Three different conformers were suggested for cinchonidine based on B3LYP/6-31G∗ calculations and solvent-dependent NMR studies. The population-weighted calculated VCD spectrum at the B3LYP/6-31G∗ level was considered to reasonably match the experimental VCD spectrum of cinchonidine in CDCl3 , leading to the suggestion that the findings from NMR and VCD are consistent. 11.3.3.2. Schizozygane Alkaloids [42]. The group of alkaloids isolated from the East African plant, Schizozygia caffaeoides, are referred to as schizozygane alkaloids. The structure of schizogygine 15, but not its absolute configuration, was deduced from chemical reactions and spectroscopic properties. The absolute configuration of a closely related alkaloid, strempeliopine (16), isolated from Cuban plant Strempeiopsis strempelioides, was determined to be (−)-(2S , 7R, 20R, 21R). 15 and 16 differ only in the absence of ring F and the hydrogenation of C14 –C15 bond. Since the [α]D of 15 in CHCl3 is +15.5, which is of opposite sign to that of 16, one could speculate the absolute configuration of 15 to be (2R, 7S , 20S , 21S ), but such correlation can be dangerous. Therefore the absolute configuration of 15, either (2R, 7S , 20S , 21S ) or (2S , 7R, 20R, 21R), remained undetermined until VCD spectroscopic analyses were undertaken. The experimental VCD spectra (at concentrations 0.127–0.425 M) and optical rotatory dispersion (ORD) spectra (0.0116 M) were measured in CDCl3 solvent, while ECD (12.5 mM) spectra were measured in CHCl3 . The conformational analysis indicated only two conformations, one with

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

ring C in chair conformation and another with ring C in boat conformation. Both conformations had ring F in near planar conformation (this was probably an artifact because SPARTAN program [16] used here does not pucker the five-member rings unless forced by the user). Re-optimization of these structures at the B3LYP/6-31G∗ level indicated that ring F is puckered. A potential energy scan varying the ring F puckering angle indicated the existence of two puckering angles for ring F. Thus a total of four conformations were found and their VCD spectra predicted at B3LYP/TZ2p and B3PW91/TZ2p levels. The population-weighted predicted VCD spectrum for (2R, 7S , 20S , 21S ) configuration was considered to match well with the experimental VCD spectrum while that for (2S , 7R, 20R, 21R) was not. Therefore the absolute configuration was suggested to be (2R, 7S , 20S , 21S )-(+)-15. This assignment was confirmed by comparing the predicted ECD and ORD spectra with the corresponding experimental spectra. Two more alkaloids, schizogaline (17) and schizogamine (18), also isolated from Schizozygia caffaeoides, have structures similar to that of 15. Assuming that the biosynthetic pathway for these compounds is identical to that of 15, the absolute configurations were suggested to be (2R, 7S , 20S , 21S )-(+)-17, and (2R, 7S , 20S , 21S )-(−)-18. In the same manner, the absolute configuration of naturally occurring 6,7-dehydro-19β-hydroxyschizozygine (19) was also suggested to be (2R, 19R, 20S , 21S )-19. 11.3.3.3. Iso-schizozygane Alkaloids [43]. Iso-schizogaline (20) and Isoschizogamine (21) had [α]D values of -260 and 262, respectively. To determine their absolute configurations, VCD spectral predictions were made for (2R, 7R, 20S , 21S )-20 and (2R,7R,20S ,21S )-21. It is not clear how this particular configuration, among many other possible diastereomers, was chosen. The conformational search for hexa-cyclic core indicated three conformations originating from changes in the conformations of rings C and E. The conformational analysis of methoxy benzene indicated two stable conformers with C1 C2 OC3 dihedral angles of 0◦ (cis) and 180◦ (trans). As a result, a total of six conformers were investigated for (2R, 7R, 20S , 21S )-20 and found to be stable at B3LYP/6-31G∗ level. Only three of these conformers were considered to be significantly populated in the gas phase. The conformational search done for orthodimethoxy benzene indicated seven stable conformations, which, in combination with three conformations of rings C and E, yielded 21 conformations for the case of (2R,7R, 20S , 21S )-21. The 14 lowest-energy conformations were further optimized at B3LYP/6-31G∗ , B3LYP/TZ2P, and B3PW91/TZ2P levels. Three of these conformers account for > 90% population. The population-weighted predicted VCD spectrum of (2R, 7R, 20S , 21S )-20 was considered to compare well with the experimental VCD spectrum (0.255 M in CDCl3 ) of naturally occurring (−)-20. In the same manner, the population-weighted predicted VCD spectrum of (2R, 7R, 20S , 21S )-21 was considered to compare well with the experimental VCD spectrum (0.225 M in CDCl3 ) of naturally occurring (−)-21. Based on these observations, the absolute configurations were assigned as (2R, 7R, 20S , 21S )-(−)-20 and (2R, 7R, 20S , 21S )-(−)-21. These assignments were confirmed with a comparison of predicted ECD and ORD spectra with corresponding experimental spectra.

11.3.4. Tropane Alkaloids 11.3.4.1. 6β-Hydroxyhyoscyamine (22). Two diastereomers of 22 with positive and negative optical rotations have been isolated from the plant materials. The literature assignments of absolute configurations of these compounds were tentative, and

401

402

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

this uncertainty led to VCD studies on these two diastereomers for determining their absolute configurations [44]. Also, it was of interest to find the spectral signature that will differentiate the two diastereomers. Using a B3LYP functional and a 6-31G∗ basis set, it was suggested that eight different conformers of each of the diastereomers contribute to the conformationally averaged VCD and VA spectra. The VCD bands in the 1300–1200 cm−1 exhibited significant dependence on the conformations, while those in the 950–1100 cm−1 showed little variation with conformation. But the predicted VCD bands in the 950- to 1100-cm−1 region gave opposite signs for the two diastereomers. These bands were suggested to originate from the tropane ring and provide differentiation between the diastereomers. The comparison between experimental and predicted VCD spectra was used to suggest the absolute configuration assignments as (−)-(3S , 6S , 2 S )-22 and (+)-(3R,6R, 2 S )-22. 11.3.4.2. 3α,6β-Diacetoxytropane (23). (−)-23 is a semisynthetic product obtained from (−)-(3S ,6S ,2 S )-22. The absolute configuration of (−)-23 was not known until VCD spectroscopy was used [45] to determine its absolute configuration as (3S , 6S )-(−)-23. Theoretical analysis was conducted with B3LYP functional and DGDZVP basis set. Because a tropane ring is conformationally rigid, the conformational freedom in this molecule originated from the disposition of N -methyl group and two acetyl groups. The equatorial orientation of the N -methyl group was found to be preferred. The C6 acetyl group was found to have more stable orientation when its C=O group was oriented outside the molecule, while C3 acetyl group was found to be iso-energetic for outside and inside orientations. A similarity in the VCD spectra of this molecule with that of (−)-(3S , 6S , 2 S )-22 led the authors to suggest the existence of several bands that reflect the absolute configuration of molecular skeleton. 11.3.4.3. 3α,6β-Tropanediol (24), and Its Mono Esters. The absolute configuration and conformational analyses for 24 (see Scheme 11.2) and four monoesters were reported, using experimental and theoretical VCD studies [46]. These monoesters, 6β-hydroxy-3α-senecioyloxytropane (25), 3α-hydroxy-6β-trigloyloxytropane (26), 3α-hydroxy-6β-senecioyloxytropane (27), and 3α-hydroxy-6β-angeloyloxytropane (28), were extracted from leaves of S . grahamii and S . pinnatus, but their optical rotation data were not provided to label them as (+)/(−) compounds. From experimental and theoretical VCD studies, the absolute configurations of investigated monoesters were deduced as (1R, 3R, 5S , 6R). For theoretical predictions, the initial conformers were screened with the Monte Carlo search method, for both axial and equatorial positions of the N–Me group. For (1R,3R,5S ,6R)-25, this resulted in finding 16 conformations with the axial N–Me group and 22 conformations with the N–Me equatorial group. These conformations were re-optimized at the B3LYP/6-31G* level to select 10 lowest-energy conformations within 2.5 kcal/mol energy difference. Of these conformations, two accounting for 72.8% population have an axial N–Me group and the remaining have an equatorial N–Me group. These conformers were re-optimized with a B3LYP functional and a DGDZVP basis set, and VCD predictions were undertaken at the same level. The two axial N–Me conformers have intramolecular hydrogen bonding between the hydroxyl group at C6 and the nitrogen atom of N–Me group. Solvation may have significant influence on the experimental spectra, which was not incorporated in the calculations. This could be one reason for significant differences between predicted and observed VCD spectra in the 1500- to 1100-cm−1 region. Nevertheless, VCD bands in 1100- to 950-cm−1 region showed similarities, and it was concluded that the 950- to 1100-cm−1 region is critical in assigning the absolute configurations for 3,6-tropanediol

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

derivatives. Based on the comparison in this region for (1R, 3R, 5S , 6R)-25 and experimental spectra, the absolute configurations of the investigated compound was suggested as (1R, 3R, 5S , 6R). For 26 and 27, since the O–H group at C6 is now esterified, hydrogen bonding with N–Me is not present; as a result, conformers with equatorial N–Me orientation (>80%) become more populated than axial N–Me conformers. The sample used for VCD experiments was identified via NMR as a 69:31 mixture of 26 and 27, and therefore the calculated VCD spectrum for this mixture was generated by combining individual calculated spectra in that ratio. A satisfactory agreement between the predicted and experimental VCD spectra was used to suggest the absolute configurations of both compounds as (1R,3R,5S ,6R).

11.3.5. Montanine-Type Alkaloids When hemanthamine (29) was reacted with SOCl2 , a compound (30) of the formula C17 H18 NO3 Cl was formed which was identified using 2D NMR to have montanine-type structure [47]. Relative stereochemistry of 30 was determined using ROESY spectrum, while the absolute configuration of 30 was determined from combined analysis of experimental and theoretical VCD spectra. Monte Carlo conformational search revealed two low-energy conformations accounting for 99% population. The geometries of these two conformations were optimized first at B3LYP/6-31G∗ level and subsequently at the B3LYP/DGDZVP level. VCD spectra were calculated for the optimized conformers at both B3LYP/DGDZVP and B3PW91/DGDZVP levels. The population-weighted predicted VCD spectra for (2S , 3S , 4aS , 11S )-30 were considered to be in good agreement with the experimental VCD spectra in CDCl3 (5.3 mg/150 μL) of (+)-30. Therefore the absolute configuration was assigned as (+)-(2S , 3S , 4aS , 11S )-30.

11.3.6. Iridoids 11.3.6.1. Plumericin (31) and Isoplumericin (32) [48]. The relative configurations of 31 and its isomer, 32, were suggested [49] based on their chemical and spectroscopic properties. The X-ray diffraction data confirmed the relative configurations assigned previously, but semiempirical analysis of experimental ECD spectra suggested [50] the opposite. The utilization of theoretical methods for interpreting the experimental VCD and ORD spectra resolved this controversy in favor of the originally suggested configurations. The conformational analysis at the MMFF94 level for 31 and 32 indicated four conformations each. These conformations were also supported by B3LYP/6-31G*, B3LYP/TZ2P, and B3PW91/TZ2P level optimizations. These conformations differ in the puckering angle of ring B and in the orientation of OCH3 group. In each case, two conformations account for 90% of the populations, so only two conformers are major contributors to the predicted spectra. The experimental VCD spectra of (+)-31 in CDCl3 (0.069 M) were considered to match well with the corresponding population-weighted predicted spectrum of (1R, 5S , 8S , 9S , 10S )-31, but not with that of (1S , 5R, 8R, 9R, 10R)-31. Therefore the absolute configurations were suggested as (1R, 5S , 8S , 9S , 10S )(+)-31 and (1R, 5S , 8S , 9S , 10S )-(+)-32. These assignments are in agreement with the original assignments [49] and were further supported by the analysis of predicted and experimental specific rotations. 11.3.6.2. Prismatomerin (33). In terms of chemical structures, 33 differs from 31 in a p-phenol group replacing the methyl group of 31 at the C14 position. The [α]D of 33 is −136 (EtOH), opposite in sign to that of 31. If it is assumed that the substitution

403

404

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of phenol group for methyl group in 31 does not change the sign of optical rotation, then one would assign 33 the opposite configuration to that of 31. But such assumptions can be dangerous because there is no way to reliably know the influence of chemical substitution on [α]D of chemically related molecules. For this reason, VCD spectroscopy was used [51] to determine the absolute configuration of 33. Since the presence of O–H groups can result in inter- or intramolecular hydrogen bonding, which makes the theoretical analysis complex, the acetate derivative, prismatomerin acetate 34, was investigated. In addition to the four conformers, as were found for 31, the rotation of phenyl acetate group results in four more conformers. Thus a total of 16 conformers were found for 34. Twelve of these conformers accounted for 99% of the population, so populationweighted predicted VCD spectra were obtained from these 12 conformers. The predicted VCD spectra at the B3LYP/TZ2P and B3PW91/TZ2p levels for (1R, 5S , 8S , 9S , 10S )-34 were considered to agree well with the experimental VCD spectra of (−)-34, leading to the assignment (1R, 5S , 8S , 9S , 10S )-(−)-34. Since the conversion of phenol group to its acetate does not change the absolute configuration, the absolute configuration of 33 is also (1R, 5S , 8S , 9S , 10S )-(−). Note that naturally occurring 31 and 33 have the same absolute configuration but oppositely signed [α]D values. Therefore it was suggested not to assign the absolute configurations based on the signs of [α]D values of chemically related molecules.

11.3.7. Meroditerpenoids 11.3.7.1. Sargaol Acetate (35) [52]. 35 is an optically active substance with positive optical rotation at 589 nm in chloroform solvent. To determine its absolute configuration, a negative–positive couplet centered at 1200 cm−1 in the experimental VCD spectrum in CCl4 (5.9 mg/150 μL) was used. Theoretical VCD spectra were predicted for 35 with (R)-configuration, using a B3LYP functional and a DGDZVP basis set. However, the tris-isoprenyl chain of this molecule introduces a large number of conformations. With a view to simplify the computational problem, the tris-isoprenyl chain was replaced with ethyl and isoprenyl, one substitution at a time. These modelmolecule predictions were considered to reproduce the major negative–positive feature found in the experimental spectrum, but weaker bands in the 1000- to 1100-cm−1 region were not. With an objective to improve the predictions, the authors used di-isoprenyl group substitution and claimed better agreement with experimental VCD spectrum on the lower-energy side of the spectrum. Finally, the authors investigated the complete structure without any model substitution. The conformer-averaged spectrum was no better than the model structure predictions, although all calculations correctly predicted the experimentally observed negative–positive couplet, suggesting the absolute configuration (+)-(R)-35. 11.3.7.2. Isoepitaondiol Diacetate (36). To prevent possible epimerization of parent diol, its diacetate derivative (36), which was extracted from S.flabelliforme, was used instead. Its structure was proposed based on various NMR spectral data for 36, and it was confirmed using X-ray diffraction data. In light of this new structure, the reported structure of parent diol was corrected. The absolute configuration of 36 could not be established from X-ray diffraction data, so VCD spectroscopy was used for this purpose [53]. Theoretical VCD predictions were obtained at the B3LYP/DGDZVP level. The Monte Carlo conformational search identified 13 low-energy conformers. This number was reduced to two, based on single-point energy calculations at the B3LYP/DGDZVP level.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

The geometries of these two conformers were optimized and VCD calculations undertaken. The resemblance between experimental VCD spectrum in CDCl3 (6.6 mg/150 μL) and conformer-averaged predicted VCD spectrum in the 1700- to 1000-cm−1 region was used to propose the absolute configuration for 36 sample as depicted in Scheme 11.2. 11.3.7.3. Stypotriol (37). The absolute configuration of 37 was deduced from that of its triacetate derivative, 38. Theoretical VCD calculations for 38 were done at the B3LYP/DGDZVP level and compared to the experimental VCD spectra of 38. Based on this comparison, the absolute configuration was concluded [54] to be (3S , 5R, 8R, 9R, 10S , 13S , 14S )-(−)-37.

11.3.8. Verticillane Diterpenoids 11.3.8.1. Verticilla-3E,7E-dien-12-ol (39). The absolute configuration of 39 derived from x-ray analysis of its p-iodobenzoate derivative was at odds with that proposed earlier using the octant rule associated with ECD. To resolve this controversy experimental VCD spectrum of (+)-39 was measured in CCl4 (10 mg/200 μL) and compared to the predicted VCD spectrum [55]. For theoretical predictions, the structure of (1S , 11S , 12S )-39 was subjected to Monte Carlo conformational search, which identified six conformations. Further optimizations of geometries of these conformers at the B3LYP/6-31G*level indicated that three lowest-energy conformers accounted for 97% population. In these three conformers, the six-member ring is in chair conformation and cyclooctadodecane moiety in chair–chair–chair–chair conformation. All three conformers have the same hydrocarbon skeleton, which was verified with 1 H NMR coupling constants. The population-weighted predicted VCD spectrum at the B3LYP/6-31G∗ level was considered to have a good correspondence with the experimental VCD spectrum for (+)-39. The same level of agreement was also noted for the predicted VCD spectrum with a B3LYP/DGDZVP basis set leading to the suggested absolute configuration, (+)-(1S , 11S , 12S )-39. This assignment is in accord with that derived from X-ray structure analysis of p-iodobenzoate derivative.

11.3.9. Sesquiterpenes 11.3.9.1. Quadrone (40) [56]. The total synthesis of quadrone enantiomers led to the determination of the absolute configuration [57] of quadrone 40 as (1R, 2R, 5S , 8R, 11R)-(−). This assignment was independently confirmed, using VCD, ECD, and ORD spectroscopic methods, in order to evaluate the applicability of these methods for structurally related sesquiterpenes, subersenone (41), subersanone (42), and seberosenol A acetate (43). The conformational analysis of 40 indicated the presence of three stable conformers. But the relative energies of these conformers at the B3LYP/6-31G∗ level indicated that only one conformer has predominant population. This predominant conformation has a cyclohexane ring in a chair conformation, a lactone ring in a boat conformation, and a cyclopentane ring in a non-planar conformation. The VCD spectrum predicted for this stable conformer with (1R, 2R, 5S , 8R, 11R) configuration at the B3PW91/TZ2P level was considered to match well with the experimental (0.14 M and 0.24 M in CDCl3 ) VCD spectrum of (−)-40, which led to confirming the previously established absolute configuration as (1R, 2R, 5S , 8R, 11R)-(−)-40. This assignment was further confirmed using predicted and experimental ECD spectra and also the [α]D values. The absolute configurations of related sesquiterpenes were suggested based

405

406

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

on the comparison between experimental and predicted [α]D values, as (1R, 2R, 8R, 11R)-(+)-41, (1R, 2R, 5S , 8R, 11R)-(−)-42, and (1R, 2R, 4R, 8R, 11R)-(−)-43, but it was suggested that a further analysis using VCD and ECD spectroscopies would be required for confident assignments. 11.3.9.2. 8-Epiisolippidiol-3-O-β-D-glucopyranoside (3β,8β,-Dihydroxy1α, 4β, 5α, 6β, 7α, 11β-guai-10(14)-ene-6,12-olide-3-O-β-D-glucopyranoside) (44). The structure of this compound was characterized on the basis of 1 H NMR spectra data (assuming α-configuration of H7) and X-ray diffraction and ECD data of its dehydrogenated derivative. The known absolute stereochemistry of its dehydrogenated derivative did not favor the diastereomeric structure with opposite configurations of aglycone. VCD spectral analysis of this compound was undertaken [58] assuming that the absolute configuration previously suggested [59] was correct. Conformational analysis involved the rotation of hydroxyl group at the C8 position, as well as rotation around the glycosidic bonds and hydroxymethyl group in the glucose moiety. The rotations around glycosidic bond were found to have major influence on energy. The orientation of hydroxyl hydrogen atom at C8 favored interaction with π -electrons of exomethylene double bond. The rotations of hydroxymethyl group of glucose moiety had smaller influence on energies as well as on spectra. A total of six conformers were investigated at the B3LYP/6-31G∗ level. The population-weighted predicted VCD spectra, after scaling the band positions, was considered to match the experimental VCD spectrum, and therefore the literature-suggested configuration was considered to be correct. 11.3.9.3. Africanane and Lippifoliane. African-1(5)-ene-2,6-dione (45) and lippifoli-1(6)-en-5-one (46), both functionalized tricyclic systems, are derived from a shrub, Lippia integrifolia. X-ray diffraction analysis of a derivative, 4,10,11tribromo-10,11-seco-lippifoli-1(6)-en-5-one, along with a combined experimental and quantum chemical analysis of ECD spectra, was used previously to derive the absolute configuration of 46 as (4R, 9S , 10R), while that of 45 was proposed based on the basis of its relationship to other natural products derived from Lippia integrifolia. An independent confirmation of the absolute configuration of 46 and determination of the absolute configuration of 45 using VCD spectroscopy were undertaken [60]. The experimental VCD spectra of (+)-45 were obtained in CCl4 solvent (5.2 mg/100 μL). For theoretical VCD predictions, conformational analysis was undertaken using the Monte Carlo search method, which resulted in two conformations for 45 and four conformations for 46. Further geometry optimizations of these conformers at the B3LYP/6-31G∗ and B3LYP/DGDZVP levels indicated that only one conformer each accounts for > 90% population in these molecules. Additional support for these conformers was derived from a comparison with the X-ray structure of 4,10,11-tribromo-10,11-seco-lippifoli-1(6)-en-5-one and 1 H NMR coupling constants. For (4R, 9R, 10R)-45, VCD spectra were predicted at three different levels of theory (B3LYP/6-31G∗ , B3LYP/DGDZVP, and B3PW91/DGDZVP2) and all of them were considered to provide good correspondence with the experimental VCD spectra of (+)-45, thereby suggesting the possible configuration as (+)-(4R, 9R, 10R)-45. For (4R, 9S , 10R)-46, VCD spectra were predicted at two different levels of theory (B3LYP/6-31G∗ and B3LYP/DGDZVP), and both of them were considered to provide good correspondence with the experimental VCD spectra of 46 [the sample studied was not identified with (+)/(−) optical rotation].

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

11.3.10. Halogenated Sesquiterpenes 11.3.10.1. Pacifenol-Related Compounds. The structure and absolute configuration of pacifenol {2,7-dibromo-8-chloro-2,5,7,8,9,9a-hexahydro-5,8,10,10tetramethyl-6H-2,5a-methano-1-benzoxepin-5-ol} (47) (see Scheme 11.3), has been documented in the literature as (2R, 5R, 5aR, 7S , 8S , 9aS )-(−). When treated with sodium hydroxide, 47 yields dibrominated chamigrene (48). When 48 is treated with mchloroperbenzoic acid, it is transformed into a α-bromoketone (49) {(−)-(2R, 5R, 5aR, 8ζ , 9aS )−2,8-dibromo-2,5,9,9a-tetrahydro-5-hydroxy-5,8,10,10-tetramethyl-6H-2,5a-methano-1-benzoxepin-7(8H )-one}, whose absolute configuration at the 8 position could be either (S ) or (R). Although the absolute configuration of 49 could not be determined from X-ray diffraction data, the two diastereomers, (2R, 5R, 5aR, 8S , 9aS )-49 and (2R, 5R, 5aR, 8R, 9aS )-49, could be distinguished using combined experimental and theoretical VCD investigations [61]. The experimental VCD spectra were obtained in CDCl3 (8.5 mg/150 μL). For the theoretical studies, Monte Carlo conformational search, done separately for the two diastereomers, (2R, 5R, 5aR, 8S , 9aS )-49 and (2R, 5R, 5aR, 8R, 9aS )-49, indicated three conformations each. Further optimization of the geometries of these conformers at the B3LYP/DGDZVP level reduced the number of low-energy conformers to two each. Theoretical VCD spectra, obtained as population-weighted spectra of the conformers, for each of the two diasteromers, were compared to the experimental VCD spectrum. In addition to the visual comparison of VCD plots, experimental rotational strengths were plotted against predicted rotational strengths for each of the two diastereomers. These analyses suggested that the absolute configuration of the reaction product 49 is (2R, 5R, 5aR, 8R, 9aS ). An interesting point to note is that in the 8R diastereomer, the cyclohexanone exists in a predominantly boat conformation, which is contrary to the notion of the chair form being more stable than the boat form. 11.3.10.2. Majapolene B (50) and acetylmajapolene B (51). These two brominated sesquiterpenes have moderate antibacterial activity against some marine bacteria. The presence of bromine atom in the chemical structure makes these compounds suitable for the X-ray determination of absolute configuration, but it was not easy to obtain the good-quality crystals of these compounds. The experimental VCD spectra (∼0.1 M in CCl4 ) measured for (−)-50 and (−)-51, when analyzed [62] with corresponding theoretical VCD spectra, revealed their configurations to be (7S , 10S ). To arrive at this conclusion, conformational search was used to identify the low energy conformers, which were then optimized at the B3PW91/6-31G(d , p) level and their VCD spectra were calculated. The population-weighted predicted VCD spectrum for (7S , 10S ) configuration was considered to compare well with the experimental VCD spectra of (−)-enantiomer. Therefore it was suggested that the configurations of the both sesquiterpenes studied are (−)-(7S ,10S ).

11.3.11. Endoperoxides 11.3.11.1. Acetylmajapolene A (52). The absolute configuration of endoperoxides is usually derived by converting them to diol derivatives via reductive cleavage and using the established practice for determining the configurations of secondary alcohols. Such reactions may sometimes produce tertiary alcohols. These difficulties were avoided [63] for 52. This natural product, isolated from red alga Laurencia, contained a disatereomeric mixture, and these diastereomers could be separated on a CHIRALPAK column. The first eluted diastereomer (A) and second eluted diastereomer (B) have the

407

408

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

same sign for optical rotation. The NMR spectra of the diastereomers are similar, so their stereochemical assignments were not possible. The absolute configurations of these diastereomers were assigned [63] using VCD spectral analysis. The experimental VCD spectra of the two diastereomers (0.15 M in CCl4 ) appeared similar except that the sign associated with a band at ∼1050 cm−1 is opposite for the two samples. A comparison of the experimental VCD spectra of the two diastereomers with predicted VCD spectra for (1R, 4R, 7S , 10S ) and (1S , 4S , 7S , 10S ) configurations indicated that the first eluted diasteromer has the (1R, 4R, 7S , 10S ) configuration and second eluted diasteremoer has the (1S , 4S , 7S , 10S ) configuration. Absolute configurations at the 7 and 10 positions were selected as S based on those established for related metabolites 50 and 51.

11.3.12. Furochromones 11.3.12.1. 5-O-Methylvisamminol (53). The structure of 53 was obtained from X-ray diffraction analysis, while its absolute configuration was derived from the experimental VCD spectrum in CDCl3 (5.3 mg in 150 μl) by comparing it with theoretical VCD predictions [64]. For the theoretical VCD spectra, initial conformational analysis included Monte Carlo search, which yielded 14 conformations. The geometries of these conformations were further optimized at the B3LYP/DGDZVP level and VCD spectra calculated at the same level. The population-weighted predicted VCD spectrum, at the B3LYP/DGDZVP level, for (S )-53 provided good correspondence with the experimental spectra of (+)-53, thereby establishing the configuration as (+)-(S )-53. 11.3.12.2. 5-O-Methylvisamminol Derivatives. Two new dihydrofurochromones, (+)-4 -O-angeloyl-5-O-methylvisamminaol, (+)-54, and (+)-4 -O-senecioyl5-O-methylvisamminaol, (+)-55, were extracted from the roots of Prionosciadium thapsoides, a medicinal plant, found in Mexico. Reaction of (+)-53 with angeloyl and senecioyl chloride provided (+)-54 and (+)-55, and therefore the absolute configurations of these compounds were assigned as (S ) by chemical correlation to the parent molecule (+)-53. In the same chemical correlation approach the absolute configuration of visamminol 56 was assigned [64] as (+)-(S )-56.

11.3.13. Eremophilanoids 11.3.13.1. 6-Hydroxyeuryopsin (57) and Its Acetyl Derivative 58 [65]. 57, isolated from Senecio toluccanus, and 58 have anti-feedant activity against certain insects. Experimental VCD spectrum for 57 was measured in CCl4 (5 mg/100 μL). The optical rotation sign was not provided to label this natural product as (+)/(−). For VCD spectral predictions, six conformers were identified via Monte Carlo search, and the geometries of these conformers were further optimized at the B3LYP/6-31G* level of theory. The energies of optimized structures indicated two conformers accounting for ∼95% population. The population-weighted predicted VCD spectrum from these two conformers with (4S , 5R, 6S ) configuration was considered to be in good agreement with the experimental spectrum, except in the 1200-cm−1 region, which was attributed to intermolecular hydrogen bonding of the hydroxyl group. To eliminate this hydrogen bonding effect, the VCD spectrum of 58 was measured in CCl4 (5.7 mg/100 μL). The predicted VCD spectrum, following the conformational analysis similar to that of parent 6-hydroxy euryopsin, for the structures with (4S , 5R, 6S ) configuration was considered to be in good agreement with the experimental spectrum including the 1200-cm−1 region.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

Therefore the absolute configuration of the natural product, 57, and of its acetylated derivative 58, were suggested [65] to be (4S , 5R, 6S ). Additional calculations at the B3PW91/DGDZVP level confirmed this assignment. 11.3.13.2. Eremophilanolide (59). The chemical name [66] of 59 is 1αangeloyloxy-8β,10β-dihydroxyeremophil-7(11)-en-8α,12-olide. The experimental VCD spectra for this compound were measured in CCl4 (4.8 mg/100 μL). The theoretical VCD spectra were predicted in the same way as for 57, but only a single conformer was found to be stable. The predicted VCD spectra for the (1S , 4S , 5R, 6S , 8S , 10S ) configuration, at both the B3LYP/6-31G∗ and B3PW91/DGDZVP levels of theory, were considered to be in good agreement with the experimental VCD spectrum. Hence the absolute configuration of the natural product 59 was suggested [65] to be (1S , 4S , 5R, 6S , 8S , 10S ).

11.3.14. Eudesmanolides 11.3.14.1. 1-Hydroxy-15-acetoxyeudesm-11(13)-en-6,12-olide (60) [67]. This was a new compound isolated from Milkania. The molecular formula was deduced from a high-resolution mass spectrum, and its structure and relative configurations were deduced from 1 H, 13 C NMR, and coupling constants. The conformation in solid state, deduced from X-ray diffraction, was found to be essentially same as that obtained from molecular modeling. Monte Carlo conformational search identified 40 conformers, and this number was reduced to 8 on further optimization at the B3LYP/6-31G∗ level. These eight conformers were further optimized at the B3LYP/6-31G∗∗ level, and their VCD spectra were calculated. The population-weighted predicted VCD spectrum at the B3LYP/6-31G∗ level was compared to the experimental VCD spectrum, which was obtained at a concentration of 10 mg/20 μL CDCl3 . The predicted VCD spectrum for (1S , 4S , 5S , 6S , 7S , 10R)-60 was considered to be in good agreement with the experimental VCD spectrum of (+)-60. This assignment, (1S , 4S , 5S , 6S , 7S , 10R)-(+)-60, was considered [67] to be in agreement with those known for many eudesamanolides.

11.3.15. Presilphiperfolanes 11.3.15.1. 9-Epi-presilphiperfolan-1-ol, (61) [68]. A natural product constituent of the oil of Anemia tomentosa var. anthriscifolia was identified [69], from a comparison of its 13 C NMR chemical shifts with those of presilphiperfolan-1-ol (62), whose crystal structure was known, as (−)-epi-presilphiperfolan-1-ol (63). A further evaluation, based on X-ray diffraction data and VCD analysis, of the natural product led to a change in the identity of its structure. The experimental VCD spectra of the natural product were obtained in CCl4 (22.0 mg/150 μL). Using the relative configuration and structure from X-ray diffraction data of the natural product, Monte Carlo conformational search was conducted for the natural product with (1S , 4S , 7R, 8R, 9S ) configuration. This search yielded nine low-energy conformers and this number of conformers was reduced to six based on single-point energy calculations at the B3LYP/6-31G∗ level. The geometries of these six conformers were optimized at the B3LYP/DGDZVP level followed by VCD calculation. The population-weighted predicted VCD spectrum for (1S , 4S , 7R, 8R, 9S ) configuration was considered to have a good correlation to the experimental VCD spectrum of the natural product, which led to the reassignment [68] of the identity of natural product, from (−)-63, to (−)-(1S , 4S , 7R, 8R, 9S )-61.

409

410

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

11.3.16. Longipinane Derivatives 11.3.16.1. 7,9-Diacetyloxylongipin-2-en-1-one (64) [70]. The absolute configuration of this compound was established [71] as (4R, 5S , 7R, 9R, 10R, 11R), from a comparison of ECD spectra with its parent (without acetyl groups) and vulgarone B. This assignment was confirmed using combined experimental and theoretical VCD spectral investigation. The experimental VCD spectra were obtained in CCl4 (7 mg/150 μL). Theoretical VCD spectra were obtained for (4R, 5S , 7R, 9R, 10R, 11R) configuration. Geometries of two low-energy conformers identified via Monte Carlo conformational search were optimized at the B3LYP/6-31G∗ level. These optimized geometries were further optimized at the B3LYP/DGDZVP level, and VCD predictions were made at the B3LYP/DGDZVP and B3PW91/DGDZVP levels of theory. The predicted VCD spectra at both levels were considered to be in good accordance with the experimental VCD spectra. This observation was used to confirm the literature assignment. VCD spectra for two additional diasteromers, (4R, 5S , 7R, 9S , 10R, 11R) and (4R, 5S , 7S , 9R, 10R, 11R), were also predicted to compare the sensitivity of VCD to epimeric stereoisomers. It was stated that significant differences are evident in the 1250- to 950-cm−1 region for the C7 epimer and in the 1100- to 950-cm−1 region for the C9 epimer. 11.3.16.2. 7β –Angeloyloxy-8α-isovaleroyloxylongipin-2-en-1-one, (65) and 7β,8α-diacetyloxylongipin-2-en-1-one (66). The methanolic extract of the plant Stevia monardifolia yielded 65 along with two known compounds, 7β,8α–diangeloyloxy longipin-2-en-1-one and 7β,8α–diangeloyloxylongipinan-1-one. The mass spectral data, 1 H and 13 C NMR data of 65, along with crystal structure of 7β,8α–dihydroxylongipin-2-en-1-one, were used to determine the relative stereochemistry. The absolute configuration was suggested [72] from the comparison of experimental and predicted VCD spectra. To obtain the predicted spectra, a Monte Carlo conformational search yielded 83 conformers. This number of conformers was reduced to 18 and 13 when optimized at the B3LYP/6-31G∗ level the B3LYP/DGDZVP level, respectively. The population-weighted VCD spectra of these 13 conformers at the B3LYP/DGDZVP level was then compared to the experimental VCD spectrum. The experimental VCD spectrum of (+)-65 was considered to compare well with that predicted for (4R, 5S , 7S , 8S , 10R, 11R)-65 at the B3LYP/DGDZVP level, leading to the assignment of natural product as (4R, 5S , 7S , 8S , 10R, 11R)-(+). The diacetyl derivative, 66, was used to further confirm this assignment, where the experimental VCD spectrum of (+)-66 was considered to compare well with the population-weighted predicted VCD spectrum of (4R, 5S , 7S , 8S , 10R, 11R)-66.

11.3.17. Cruciferous Phytoalexins 11.3.17.1. Dioxibrassinin (67), and 3-Cyanomethyl-3-hydroxyoxindole (68). These compounds were synthesized [73] in racemic form and the (−)-enantiomer separated using chiral HPLC. There are no established spectra-structure correlations that can determine the absolute configuration at C3 , containing the tertiary alcohol group, of oxindole ring. Therefore, (−)-67 and (−)-68 were subjected to VCD spectroscopic investigations. A model compound 3-hydroxy-3-methyloxindole, 69 was also investigated for the sake of facilitating spectral interpretations. Experimental VCD spectra were measured in DMSO-d6 (∼2 M). For obtaining the theoretical VCD predictions, several low energy conformers were selected from conformational analysis and their geometries re-optimized at B3LYP/6-31G(d , p) level. Although the

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

experimental VA spectra showed a simple pattern in the 1900- to 1400-cm−1 region, with three bands at 1730, 1620, and 1470 cm−1 , the associated VCD band signs differed among the three molecules. The 1730-cm−1 band originated from the C=O stretching vibration and 1620-cm−1 band originated from the aromatic ring stretching vibration. The population-weighted predicted VCD spectra were considered to compare well with the corresponding experimental spectra for (−)-(S )-67, (−)-(S )-68 and (−)-(S )-69. 11.3.17.2. Spirobrassinin (70), 1-Methoxyspirobrassinin (71), and 1-Methoxyspirobrassinin methyl ether (72). The absolute configurations of 71 and 72 were not known until they were determined using VCD spectroscopy [74]. These compounds were synthesized in racemic form, and the enantiomers separated using chiral HPLC. VCD spectra were measured in CDCl3 at a concentration of 0.15 M. Most of the VCD bands measured for (+)-71 were found to have signs opposite to those found for (S )-(−)-70. Therefore the absolute configuration of 71 was established as (R)-(+). This assignment was supported by the observation of opposite signed ECD bands for (+)-71 and (S )-(−)-70. In the case of 72, such correlation could not be used due to the differences in their chemical structures. Therefore, theoretical VCD predictions were undertaken. The 11 conformations, identified using a conformational analysis program [17], were further optimized at B3LYP/6-31G(d , p) level, and four lowest-energy conformers were used to obtain theoretical VCD at the same level. The population-weighted predicted VCD spectrum for (2R, 3R)-72 was considered to provide a reasonable correlation to the experimental VCD spectrum of (−)-72, thereby assigning the configuration of 72 as (2R, 3R)-(−). This conclusion was further supported by stereochemical transformation reactions of (−)-72 and (+)-72. 11.3.17.3. Brassicanal C (73) [75]. The chirality of this natural product arises from the sulfinate group attached at the C-2 position. The racemic form of 73 was synthesized and its enantiomers separated using chiral HPLC. The experimental VCD spectra were measured in CDCl3 (0.05 M). The geometries of 36 conformers of 73 identified by manually changing the dihedral angles were first optimized at the B3LYP/631G(d ) level, and the resulting nonidentical conformations were further optimized at the B3LYP/6-311+G(2df , 2p) level. The four lowest-energy conformers obtained were used to calculate VCD. The population-weighted predicted VCD spectrum for (S )-73 was considered to compare reasonably well with the experimental VCD spectrum of (−)-73, resulting in the assignment of its absolute configuration as (−)-(S ). This conclusion was further supported by comparing the experimental and theoretical predictions of optical rotation and ECD for (S )-(−)-73.

11.3.18. Furanones [76–78] Most of the naturally occurring furanones are often isolated as racemic compounds due to racemization of the enantiomers caused by keto–enol tautomerism. But it was possible to separate the enantiomers by HPLC. The absolute configuration of 2,5-dimethyl-4hydroxy-3-(2H )-furanone, 74 (see Scheme 11.4), was difficult to determine because of its rapid racemization through keto–enol tautomerism. This problem was overcome by carefully converting it to its methyl ether, 2,5-dimethyl-4-methoxy-3(2H )-furanone, 75, determining the absolute configuration of the later and relating it back to the parent. The absolute configuration of 75 was determined by comparing the experimental VCD spectrum (0.17 M in CCl4 ) with the predicted VCD spectrum at B3PW91/6-31G(d , p)

411

412

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

level. There was only one stable conformer for this molecule. The predicted spectrum for this conformer with (R)-configuration was found to be identical to the experimental VCD spectrum of (+)-enantiomer, so the absolute configuration of 75 was assigned as (+)-(R). Since careful reaction of (+)-74 with diazomethane yielded (+)-(R)-75, the absolute configuration of the former was also deduced as (+)-(R). In the case of 4-acetoxy-2,5-dimethyl-3(2H )-furanone, 76, conformational analysis [17] yielded four conformations and optimization of their geometries at B3PW91/6-31G(d , p) level resulted in two low-energy conformers. The population-weighted predicted VCD spectrum for (R)-configuration was considered to compare well with the experimental VCD spectrum (0.08M in CCl4 ) of (+)-enantiomer, suggesting the absolute configuration as (R)-(+)-76. The absolute configurations of 5-ethyl-4-hydroxy-2-methyl-3(2H )-furanone (77) and 2ethyl-4-hydroxy-5-methyl-3(2H )-furanone (78) are also difficult to determine because of their rapid racemization through keto–enol tautomerism. Just as in the case of 74, this problem was overcome by determining the absolute configuration of their methyl ethers 79 and 80 using VCD spectroscopy and relating them back to the parent molecules. The absolute configurations determined in this manner were (R)-(+) for all of these compounds.

11.3.19. Furanocoumarins [79] Furanocoumarins contain a furan ring fused with coumarin. The furan ring may be fused in different ways, resulting in linear and bent structures. (+)-5,8-Dimethoxymarmesin, referred to as (+)-alternamin, a furanocoumarin isolated from the plant, Murraya alternans, was found to have antidote activity against snake venom. The chemical structure of alternamin, 81, was deduced from high-resolution mass spectra and NMR. While three different possibilities exist for fusing the furan ring (two linear (81-(a), 81-(b)) and one bent (81-(c)) structural forms), the NMR data suggested the linear form 81-(a), but the absolute configuration was not known. The absolute configuration of (+)-alternamin was established using VCD spectroscopy. The experimental VCD spectra measured in CDCl3 solvent were compared to the predicted VCD spectra for the three possible structural isomers, (a)–(c), with (S )-configuration. In each case, conformational analysis [17] yielded multiple conformations. The low-energy conformers were subjected to geometry optimizations at B3PW91/6-311++G(d , p) level. The population-weighted predicted VCD spectrum for the (S )-configuration of structure (a) was considered to yield the best agreement with the experimental VCD spectrum of (+)-81. Therefore the absolute configuration of alternamin was assigned as (S )-(+)-81-(a).

11.3.20. Klaivanolide [80] Although the structure of 82 has been established from NMR spectra, its absolute configuration was not. VCD spectroscopy was used [80] to establish the absolute configuration of 82. The experimental IR and VCD spectra, in the 1900- to 800-cm−1 range, were obtained for (+)-82 in CDCl3 (0.03 M). Initial conformational analysis revealed 33 conformers. Further geometry optimization of these conformations using B3LYP functional and 6-31G∗ basis set yielded 24 stable conformations, of which seven have populations greater than 4%. Subsequent geometry optimizations with B3PW91 functional and TZ2P basis set identified five lowest energy conformers, all resulting from rotations around C7 -C8 , C8 -O9 and C5 -O18 bonds. The VA and VCD spectra were predicted for the five lowest–energy conformers of (7S )-82 using the B3PW91 functional and the TZ2P basis

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

set and the population-weighted spectra compared to the experimental spectra. The major features in the experimental VCD spectrum are one broad positive VCD band in the 1700 to 1800-cm−1 range (which is associated with C=O stretching vibrations) and a bisignate couplet with positive VCD at 1275 cm−1 and negative VCD at 1232 cm−1 (associated with C–O bond stretching vibrations). These features were considered to be reproduced well in the predicted VCD spectrum for (S )-82 with one notable difference. The predicted VCD spectrum in the 1700 to 1800-cm−1 range has resolved positive VCD bands, while the corresponding region in the experimental spectrum was broad and unresolved. Nevertheless, the satisfactory, yet qualitative, agreement between experimental VCD spectrum of (+)-82 and predicted VCD spectrum of (7S )-82 was used to suggest that the absolute configuration of the studied natural product is (+)-(7S )-82.

11.3.21. Pheromones 11.3.21.1. 1-Acetoxymethyl-2,3,4,4-tetramethylcyclopentane (83). The relative configuration of 83, a sex pheromone, was known from NMR spectra, but the absolute configuration was not known until VCD spectral investigations were undertaken [81]. The experimental VCD spectrum (0.90 M in CDCl3 ) of (+)-83 was compared to the predicted VCD spectra for 83 with (1S , 2S , 3R)- and (1R, 2R, 3S )-configurations. A conformational search indicated 15 conformations, and population-weighted predicted VCD spectrum of (1S , 2S , 3R)-83 and experimentally measured VCD spectrum of (+)-83 were considered to be in good agreement leading to the absolute configuration assignment as (1S , 2S , 3R)-(+)-83. 11.3.21.2. Frontalin (1,5-dimethyl-6,8-dioxabicyclo[3.2.1]octane), 84. The Southern pine beetle, Dendrooctoonus frontalis, uses (1S , 5R)-(−)-84 as the active pheromone. Since the absolute configuration of this molecule was already known [82], this compound was useful to test the predictive capabilities of VCD. The experimental VCD spectra were measured for (+)-84, and theoretical spectra were calculated [83] for (1R,5S )-84. The conformer with a six-member ring in a chair conformation and a seven-membered ring in a boat conformation was found to be energetically favored. The VCD spectrum predicted at the B3LYP/6-31G∗ level for this stable conformation of (1R, 5S )-84, but not that of (1S , 5R)-84, was considered to be in good agreement with the experimental VCD spectrum observed for (+)-84. Also, the observed rotational strengths correlated well to the predicted rotational strengths of (1R, 5S ), but not to those of (1S , 5R). Thus VCD spectroscopy confirmed the previously known configuration of 84 as (1R, 5S )-(+).

11.3.22. Norlignan Norlignan is a class of phenylpropanoids. Hinokiresinol, belonging to this class, can have an E or Z double bond in its molecule. The Z -isomer is named nyasol, 85, and the E -isomer as hinokiresinol (sometimes as E -hinokiresinol), 86. There were some uncertainties regarding the absolute configurations of these compounds. To resolve these uncertainties, VCD spectroscopic investigation was undertaken [84] for 85. The experimental VCD spectra of (+)-85 were measured in DMSO-d6 solvent (5 mg/100 μL). Although the experimental VCD measurements were also done for KBR pellet, artifacts were suspected in the measured spectra. Conformational search and subsequent geometry optimizations at B3LYP/6-31G∗∗ level of theory identified eight lowest-energy conformers. Gas-phase predicted conformations did not differ significantly from solution-phase

413

414

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

predicted conformations obtained with PCM. The population-weighted predicted VCD spectra in the gas phase at B3YP/6-31G∗∗ , and B3LYP/aug-cc-pVDZ levels of theory, and in the solution phase at B3LYP/6-31G*∗ (using PCM) were compared to the experimental spectra. The solvent absorption restricted the measurements to the 1800- to 1100-cm−1 region where only one dominant positive VCD band was found at 1510 cm−1 . This experimentally observed positive VCD band was considered to reproduce the corresponding band in the predicted spectrum for (S )-85 leading to the assignment of configuration as (S )-(+). Based on this conclusion, and the chemical conversions, the absolute configuration of 86 was suggested as (S )-(−).

11.3.23. Taxol Paclitaxel (Taxol) is a natural product derived in a very low yield from the Yew tree, Taxus brevifolia, and is used as an anticancer drug. The total synthesis of taxol involves complex reactions, and therefore commercial production was not economically feasible. For this reason, another precursor natural product, baccatin III, readily available in large quantities attracted much attention for the synthesis of paclitaxel. It was found that the VCD spectra of taxol and baccatin III (both obtained at 0.029 M in CDCl3 ) are quite similar [85]. The conformational analysis of baccatin III indicated 23 lowest-energy conformations, which upon further geometry optimizations yielded three lowest-energy conformers that account for 97% of the population. The population-weighted predicted VCD spectrum of baccatin III was considered to reproduce the experimental VCD spectrum quite well. Therefore the three lowest-energy conformers used for predicting the VCD spectrum were considered to be the predominant conformations of baccatin III in solution. The electron micrograph of the crystal structure of the paclitaxel–tubulin complex, which is considered to be a bioactive structure, indicated a conformation that is among the unstable conformations noted for baccatin III in solution. Based on this observation, it was suggested that a conformational change occurs in paclitaxel during its binding with tubulin.

11.3.24. Ginkgolides Ginkgo biloba is one of oldest living trees. Extract from the Ginkgo plant leaves and roots contains different ginkgolides: ginkgolide A (GA), ginkgolide B (GB), ginkgolide C (GC), ginkgolide J (GJ), and ginkgolide M (GM). This extract is often used to prevent the development of dementia or to improve focus. 11.3.24.1. Nature of Interactions Between Ginkgolides and Amyloid Peptide. The interactions with amyoid peptides were investigated [86] using amyloid Aβ(25–35) peptide, as a model for the full-length peptide. GA, GB, GC, and bilobalide (BB) and two ethers, GA-monoether and GA-diether, were investigated using VCD spectra. The experimental VA and VCD spectra of GA, GB, GC, BB, GA-monoether, and GA-diether were analyzed in conjunction with density functional theoretical predictions. The time-dependent experimental spectra of Aβ(25–35) peptide and the corresponding experimental spectra in the presence of ginkgolides indicated that the influence of ginkgolides in modulating the aggregation of Aβ(25–35) peptide is relatively minor. Such minor effects may have been due to the absence of a specific interaction with Aβ(25–35) peptide. It was suspected that the therapeutic effect of Ginkgo biloba extract may have been more complex.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

11.3.24.2. Ginkgolide B. The experimental infrared absorption and VCD spectra of GB, in combination with quantum mechanical calculations, were used to suggest [87] the preferred conformations of GB in solution. The experimental data were obtained separately for GB in dimethyl sulfoxide-d6 (3.9 mg/20 μL), in CD3 CN (1.8 mg/50 μL) solvents, and as KBr pellet (0.5 mg in 250 mg KBr). Quantum chemical calculations were undertaken at the B3LYP/6-31G∗ level for GB and also for GA, GC, GJ, and GM to assess the possibility of distinguishing ginkgolides form their spectra. Although a number of diastereomers are possible for ginkgolides, two diastereomers with inversion at C1 (labeled GB-C1 i) and C10 (labeled GB-C10 i) were considered for GB. Two conformers within 7–8 kJ/mol were identified each for GB, GC, and GM. Only the first lowest-energy conformer was used in further calculations, for GA, GJ, GBC1 i, and GBC10 i, while two lowest-energy conformers were used for GB, GC, and GM. Based on the comparison of predicted VCD spectra for different ginkgolides and assuming that the experimental spectra can be closely approximated by the corresponding predicted VCD spectra, it was suggested that VCD spectroscopy might allow discrimination between GB and other ginkgolides. Furthermore, the predicted VCD spectra for GB markedly differed from those of GBC1 i and GBC10 i, and this observation prompted the suggestion that the experimental VCD spectra might be able to provide discrimination among the diastereomers of GB. The structure of lowest-energy conformer of GB optimized with 6-31G∗ basis set agreed with its crystal structure. There are similarities between experimental VCD spectra of GB and theoretical VCD spectra predicted for the lowest-energy conformer of GB. But there are also differences between them, and these differences appeared to correspond to the predicted features for the second lowest-energy conformer. Based on this observation, it was suggested that the two lowest-energy conformers of GB are probably present in solution.

11.3.25. Peptides Although peptides represent a large class of molecules, a few peptides are also natural products. Among these natural products, pexiganan and cyclosporin have been studied using VCD spectroscopy. 11.3.25.1. Pexiganan. Pexiganan is an analogue of the magainin family of antimicrobial peptides and contains 22 amino acids (Gly-Ile-Gly-Lys-Phe-Leu-Lys-Lys-Ala-LysLys-Phe-Gly-Lys-Ala-Phe-Val-Lys-Ile-Leu-Lys-Lys). This peptide is present in the skin of the African clawed frog. Conformational analysis of pexiganan was conducted [88] using the experimental VCD studies in different solvents (trifluoroethanol, methanol, D2 O and DMSO-d6 ) and supplemented with IR and ECD measurements. All these spectroscopic measurements suggested that the pexiganan peptide has the tendency to adopt different structures in different environments: an α-helical conformation in TFE, a sheet-stabilized β-turn structure in methanol, a random coil with β-turn structure in D2 O, and a solvated β-turn structure in DMSO-d6 . 11.3.25.2. Cyclosporins. These are a group of cyclic peptides of 11 amino acids isolated from fungi. They differ from each other in the identity of one amino acid and contain an unnatural d-amino acid. To determine the solution state conformations of cyclosporin A (CsA), cyclosporin D (CsD), cyclosporin G (CsG), and cyclosporin H (CsH), VCD spectral investigations were undertaken [89]. The experimental IR and VCD spectra in the amide I region were measured for free cyclosporins (20 mM) and

415

416

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

their Mg2+ complexes (20 mM with 10-fold excess magnesium perchlorate) in CD3 CN. The spectra in the N–H stretching region for free cyclosporins were measured at 2 mM in CDCl3 . IR and VCD spectra were predicted for three different fragments, with side chains reduced to methyl groups and adding terminal methyl groups. The geometry optimizations and IR and VCD spectra were predicted at the BPW91/6-31G∗ level. Based on the comparison of experimental spectra with the predicted IR and VCD spectra of fragments, it was concluded that the conformers of CsA, CsD, and CsG in solution are similar to those found in crystal structures. The presence of additional conformers with decreased intramolecular hydrogen bonding was suggested for CsH in solution. Evidence for disruption of β-sheet structure for CsC and CsH in CDCl3 and for CsA in CD3 CN was suggested. The spectral alterations seen in the presence of Mg2+ ions indicated strong interactions between metal ions and cyclosporin.

11.3.26. Axially Chiral Natural Products 11.3.26.1. Dicurcuphenol B (87) and Dicurcuphenol C (88). The curcuphenol dimers contain both axial chirality (due to hindered rotation around the bond between two phenyl rings) and central chiraity (due to chiral centers in the alkyl side chain). 87 and 88 have the same alkyl side chains and differ only in their axial chirality. They both exhibit positive specific rotations at the sodium D line. The experimental VCD spectra were obtained at 0.23M for 87 and at 0.11 M for 88, both in CDCl3 . Some simplifications were made for determining the absolute configuration using VCD spectroscopy [90]. (a) The VCD spectrum for structural analysis was generated by taking one-half of the experimental VCD spectral difference, (Dicurcuphenol B - Dicurcuphenol C)/2, with the assumption that this difference reflects the VCD spectrum due to axial chirality. (b) Since (S )-(+)-curcuphenol itself did not exhibit any VCD features in the 1600- to 1425-cm−1 region, and the VCD features seen in this region for Dicurcuphenol B and Dicurcuphenol C overlapped with each other, VCD features in this region are considered to be arising from axial chirality. (c) To avoid conformational mobility associated with an alkyl side chain, a model compound 89 that represents axial chirality but contains no side chain was used for predicting VCD spectra at the B3LYP/6-31G∗ level. The predicted VCD spectrum for the model compound with P chirality was considered to match well with the above-mentioned difference spectrum in the 1600- to 1425-cm−1 region. Therefore 87 and 88 were assigned P and M configurations, respectively. 11.3.26.2. Gossypol (90). 90 exists as an aldehyde in CDCl3 solvent and as an equilibrium mixture of tautomers in some other solvents [91]. Based on the ECD spectral interpretations using approximate theoretical methods, the absolute configuration of gossypol enantiomers were suggested [92] in the literature to be (M )-(−) and (P )(+). This assignment has been confirmed by VCD spectroscopy [93] and, in addition, predominant conformations of 90 have been identified. The experimental VCD spectra were measured for both enantiomers in CDCl3 solvent (5 mg of (+)-90 and 4 mg of (−)-90 each in 100 μL of CDCl3 ). For conformational analysis, at the B3LYP/6-31G∗ level, the rotational freedom associated with the aldehyde group and isopropyl group were considered. In addition, different intramolecular hydrogen bonding patterns among neighboring O–H groups and between O–H groups and oxygen of aldehyde group were considered. The population-weighted VCD spectrum, obtained from three low-energy conformers of (M )-90, was considered to correlate with the experimental VCD spectrum of (−)-90 and confirmed the literature assignment.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

11.3.26.3. Cephalochromin (91). Isolated from the culture broth of the endophytic fungus Pseudoanguillospora sp, 91 possesses the axial chirality of the atropisomeric naphthopyranone systems, in addition to central chirality elements as two tertiary stereogenic centers. This results in four diastereomers: (aS , 2R, 2 R), (aS , 2S , 2 S ), (aR, 2R, 2 R), and (aR, 2S , 2 S ). The axial chirality of (+)-91 was determined from ECD studies to be (aS ). Then two possibilities exist for the absolute configuration of (+)-91: (aS , 2R, 2 R) and (aS , 2S , 2 S ). It was not possible to determine the nature of central chirality in the presence of overwhelming influence from axial chirality on its chiroptical properties. This problem was resolved [94] using VCD spectroscopy because vibrations attributable to naphthapyranone moieties (whose relative orientation determines the axial chirality) and stereogenic centers appear spatially separated at different vibrational frequencies. Based on the comparison between the experimental VCD spectra of (+)-91 and B3LYP/6-311G∗ predicted VCD spectra for (aS , 2R, 2 R)-91 and (aS , 2S , 2 S )-91, three VCD bands in the 1050- to 900-cm−1 region provided a means to distinguish the two diastereomers of 91, thereby allowing the determination of both types of chirality elements. The sign of the VCD band at 1038 cm−1 was used to label the absolute configurations of two stereoisomers of 91 as (+)589 -[(−VCD)1038cm−1 ]-(aS , 2R, 2 R) and (+)589 -[(+VCD)1038cm−1 ]-(aS , 2S , 2 S ).

11.4. CONCLUDING REMARKS VCD spectroscopy has evolved from a curiosity in 1970s into an accepted structural tool of the twenty-first century for the determination of three-dimensional molecular structures of chiral molecules. A comparison of quantum chemical predictions of VCD with corresponding experimental observations is required for this purpose. In this process, however, VCD spectral calculations for all possible diastereomers, and all possible conformations in each case, must be undertaken. A conclusion on the absolute configuration can be reached if the calculated VCD spectrum for only one of all possible absolute configurations matches the experimental VCD spectrum for an enantiomer of a given chiral compound. With the availability of commercial VCD instruments and increasingly advanced computational resources, the task of determining the absolute configurations using the combined experimental and quantum chemical VCD spectra is becoming a routine process.

ACKNOWLEDGMENTS I am grateful to Drs. Joseph-Nathan, K. Monde, and M. Urbanova for providing a list of their VCD publications on natural products. Funding from NSF (CHE-0804301) is gratefully acknowledged.

REFERENCES 1. N. Berova, K. Nakanishi, R. W. Woody, eds., Circular Dichroism: Principles and Applications, John Wiley & Sons, New York, 2000. 2. P. J. Stephens, J. Phys. Chem. 1985, 89 , 748–752; A. D. Buckingham, P. W. Fowler, P. A. Galwas, Chem. Phys. 1987, 112 , 1–14. 3. (a) Gaussian 09, Gaussian Inc., www.gaussian.com. (b) Turbomole, www.turbomole.com. (c) Amsterdam Density functional program; www.scm.com. (d) DALTON program; www.kjemi. uio.no/software/dalton.

417

418

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

P. J. Stephens, F. J. Devlin, J. J. Pan, Chirality 2008, 20 , 643–663. L. A. Nafie, Nat. Prod. Comm. 2008, 3 , 451–466. T. A. Keiderling, Nature 1986, 322 , 851–852. G. Shanmugam, P. L. Polavarapu J. Am. Chem. Soc. 2004, 126 , 10292–10295. L. A. Nafie, R. K. Dukor, Vibrational optical activity in chiral analysis, in Chiral Analysis, K. W. Busch and M. A. Busch, eds., Elsevier, Amsterdam, 2006. pp 505–544. J. R. Cheeseman, M. J. Frisch, F. J. Devlin, P. J. Stephens, Chem. Phys. Lett. 1996, 252 , 211–220. A. D. Becke, J. Chem. Phys. 1993, 98 , 5648–5652. R. Ditchfield, W. J. Hehre, J. A. Pople, J. Chem. Phys. 1971, 54 , 724–728. J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev . 2005, 105 , 2999–3094. P. L. Polavarapu, N. Jeirath, S. Walia, J. Phys. Chem. A 2009, 113 , 5423–5431. G. Scalmani and M. J. Frisch, J. Chem. Phys. 2010, 132 , 114110–114124. P. L. Polavarapu, G. Scalmani, E. K. Hawkins, C. Rizzo, N. Jeirath, I. Ibnusaud, D. Habel, D. S. Nair, S. Haleema, J. Nat. Prod . 2011, 74 , 321–328. Spartan ’08, Wavefunction, Inc., Irvine, CA; www.wavefun.com/products/spartan.html Conflex, CONFLEX Corporation, AIOS Meguro 6F, 2-15-19, Kami-Osaki, Shinagawa-ku, Tokyo 141-0021, Japan; www.conflex.us Hyperchem, Hypercube Inc., 1115 NW 4th street, Gainesville, FL 32601; www.hyper.com F. Mohammadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M. Lipton, C. Caufield, G. Chang, T. Henderickson, W. C. Still, J. Comput. Chem. 1990, 11 , 440–467. (a) P. Zhang, P. L. Polavarapu, J. Phys. Chem. A 2007, 111 , 858–871; (b) J. He, P. L. Polavarapu, J. Chem. Theory Comput. 2005, 1 , 506–514; (c) P. L. Polavarapu, E. A. Donahue, G. Shanmugam, G. Scalmani, E. K. Hawkins, C. Rizzo, I. Ibnusaud, G. Thomas, D. Habel and D. Sebastian J. Phys. Chem A 2011, 115 , 5665–5673. A. Nakahashi, N. Miura, K. Monde, S. Tsukamoto, Bioorg. Med. Chem. Lett. 2009, 19 , 3027–3030. S. Abbate, G. Longhi, L. Ricard, C. Bertucci, C. Rosini, P. Salvadori, A. Moscowitz, J. Am. Chem. Soc. 1989, 111 , 836–840. E. Castiglioni, F. Lebon, G. Longhi, S. Abbate, Enantiomer 2002, 7 , 161–173. F. J. Devlin, P. J. Stephens, J. R. Cheeseman, M. J. Fritsch, J. Phys. Chem. A 1997, 101 , 9912–9924. F. J. Devlin, P. J. Stephens, J. R. Cheeseman, M. J. Fritsch, J. Am. Chem. Soc. 1996, 118 , 6327–6328. C. Guo, R. D. Shah, R. K. Dukor, T. B. Freedman, X. Cao, L. A. Nafie, Vibrational Spectrosc. 2006, 42 , 254–272. E. Castiglioni, F. Lebon, G. Longhi, S. Abbate, Enantiomer 2002, 7 , 161–173. M. Urbanova, V. Setnicka, K. Volka, Chirality 2000, 12 , 199–203. J. R. A. Moreno, F. P. Urena, J. J. L. Gonzalez, Vibrational Spectrosc. 2009, 51 , 318–325. J. R. A. Moreno, F. P. Urena, J. J. L. Gonzalez, Phys. Chem. Chem. Phys. 2009, 11 , 2459–2467. F. P. Urena, J. R. A. Moreno, J. J. L. Gonzalez, J. Phys. Chem. A 2008, 112 , 7887–7893. F. J. Devlin, P. J. Stephens, P. Besse, J. Org. Chem. 2005, 70 , 2980–2993. E. Burgue˜no-Tapia, L. G. Zepeda, P. Joseph-Nathan, Phytochem. 2010, 71 , 1158–1161. M. del H. Loandos, M. B. Villecco, E. Burgue˜no-Tapia, P. Joseph-Nathan, C. A. N. Catal´an, Nat. Prod. Comm. 2009, 4 , 1537–1545. (a) R. D. Singh, T. A. Keiderling, J. Am. Chem. Soc. 1981, 103 , 2387–2394; (b) P. Bour, J. McCann, H. Wieser, J. Phys. Chem. A 1998, 102 , 102–110. F. Maurer, H. Wieser, Proceedings of SPIE-The International Society for Optical Engineering 1992, 1575 (Int. Conf. Fourier Transform Spectrosc., 8th, 1991), 410–411. F. P. Urena, J. R. A. Moreno, J. J. L. Gonzalez, Tetrahed. Asym. 2009, 20 , 89–97.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

38. E. Debie, L. Jaspers, P. Bultinck, W. Herrebout, B. van der Veken, Chem. Phys. Lett. 2008, 450 , 426–430. 39. E. Debie, L. Jaspers, P. Bultinck, W. Herrebout, B. van der Veken, Phys. Chem. Chem. Phys. 2008, 10 , 3498–3508. 40. V. P. Nicu, E. Debie, W. Herrebout, B. van der veken, P. Bultinck, E. J. Baerends, E. J. Chirality, 2010, 21, E287–E297. 41. T. Burgi, A. Vargaa, A. Baker, J. Chem. Soc. Perkin Trans. 2002, 2 , 1596–1601. 42. P. J. Stephens, J. J. Pan, F. J. Devlin, M. Urbanova, J. Hajicek, J. Org. Chem. 2007, 72 , 2508–2524. 43. P. J. Stephens, J. J. Pan, F. J. Devlin, M. Urbanova, O. Julinek, J. Hajicek, Chirality 2008, 20 , 454–470. 44. M. A. Mu˜noz, O. Mu˜noz, P. Joseph-Nathan, J. Nat. Prod . 2006, 69 , 1335–1340. 45. M. A. Mu˜noz, O. Mu˜noz, P. Joseph-Nathan, Chirality 2010, 22 , 234–241. 46. M. Reina, E. Burgue˜no-Tapia, M. A. Bucio, P. Joseph-Nathan, Phytochem. 2010, 71 , 810–815. 47. J. C. Cedr´on, A. Est´evez-Braun, A. G. Ravelo, D. Guti´errez, N. Flores, M. A. Bucio, N. P´erez-Hern´andez and P. Joseph-Nathan, Org. Lett. 2009, 11 , 1491–1494. 48. P. J. Stephens, J. J. Pan, F. J. Devlin, K. Krohn, T. Kurtan, J. Org. Chem. 2007, 72 , 3521–3536. 49. G. Albers-Schonberg, H. Schmid, Helv. Chim. Acta 1961, 44 , 1447–1473. 50. B. Elsasser, K. Krohn, M. N. Akhtar, U. Florke, S. F. Kouam, M. G. Kuigoua, B. T. Ngadgui, B. M. Abegaz, S. Antus, T. Kurtan, Chem Biodiversity 2005, 2 , 799–808. 51. P. J. Stephens, J. J. Pan, K. Krohn, J. Org. Chem. 2007, 72 , 7641–7649. 52. M. A. Mu˜noz, C. Areche, J. Rovirosa, A. San-Mart´ın, P. Joseph-Nathan, Heterocycles 2010, 81 , 625–635. 53. C. Areche, A. San-Mart´ın, J. Rovirosa, M. A. Mu˜noz, A. Hern´andez-Barrag´an, M. A. Bucio, P. Joseph-Nathan, J. Nat. Prod . 2010, 73 , 79–82. 54. M. A. Mu˜noz, C. Areche, A. San Mart´ın, J. Rovirosa, P. Joseph-Nathan, Nat. Prod. Comm. 2009, 4 , 1037–1040. 55. C. M. Cerda-Garc´ıa-Rojas, H. A. Garc´ıa-Guti´errez, J. D. Hern´andez-Hern´andez, L. U. Rom´anMar´ın, P. Joseph-Nathan, J. Nat. Prod . 2007, 70 , 1167–1172. 56. P. J. Stephens, D. M. McCann, F. J. Devlin, A. B. Smith, J. Nat. Prod . 2006, 69 , 1055–1064. 57. A. B. Smith, J. P. Konopelski, B. A. Wexler, P. A. Sprengeler, J. Am. Chem. Soc. 1991, 113 , 3533–3542. 58. Michalski, W. Kisiel, K. Michalska, V. Setnicka, and M. Urbanova, J. Mol. Struct. 2007, 871 , 67–72. 59. W. Kisiel, B. Barszcz, Phytochem. 1995, 39 , 1395. 60. C. M. Cerda-Garc´ıa-Rojas, C. A. N. Catal´an, A. C. Muro, P. Joseph-Nathan, J. Nat. Prod . 2008, 71 , 967–971. 61. M. A. Mu˜noz, C. Chamy, A. Carrasco, J. Rovirosa, A. San Mart´ın, P. Joseph-Nathan, Chirality 2009, 21 , E208–E214. 62. K. Monde, T. Taniguchi, N. Miura, C. S. Vairappan, M. Suzuki, Chirality 2006, 18 , 335–339. 63. K. Monde, T. Taniguchi, N. Miura, C. S. Vairappan, M. Suzuki, Tetrahed. Lett. 2006, 47 , 4389–4392. 64. J. M. Torres-Valencia, O. E. Ch´avez-R´ıos, C. M. Cerda-Garc´ıa-Rojas, E. Burgue˜no-Tapia and P. Joseph-Nathan, J. Nat. Prod . 2008, 71 , 1956–1960. 65. E. Burgue˜no-Tapia and P. Joseph-Nathan, Phytochem. 2008, 69 , 2251–2256. 66. E. Burgue˜no-Tapiaa, B. Hern´andez-Carlosb and P. Joseph-Nathan J. Mol. Struct. 2006, 825 , 115–123. 67. M. Krautmann, E. C. de Riscala, E. Burgue˜no-Tapia, Y. Mora-P´erez, C. A. N. Catal´an and P. Joseph-Nathan, J. Nat. Prod . 2007, 70 , 1173–1179.

419

420

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

68. P. Joseph-Nathan, S. G. Leit˜ao, S. C. Pinto, G. G. Leit˜ao, H. R. Bizzo, F. L. P. Costa, M. B. de Amorim, N. Martinez, E. Dellacassa, A. Hern´andez-Barrag´an and N. P´erezHern´andez, Tetrahedron Lett. 2010, 51 , 1963–1965. 69. S. C. Pinto, G. G. Leitao, H. R. Bizzo, N. Martinez, E. Dellacassa, F. Santos Jr., F. L. P, Costa, M. B. Amorim, S. G. Leitao, Tetrahedron Lett. 2009, 50 , 4785–4787. 70. C. M. Cerda-Garc´ıa-Rojas, E. Burgue˜no-Tapia, L. U. Rom´an-Mar´ın, J. D. Hern´andezHern´andez, T. Agull´o-Ortu˜no, A. Gonz´alez-Coloma and P. Joseph-Nathan, textitPlanta Med. (J. Med. Plant Res.), 2010, 76 , 297–302. 71. P. Joseph-Nathan, C. M. Cerda-Garcia-Rojas, R. E., del Rio, L. U. Roman-Marin, J. D. Hernandez, J Nat prod . 1986, 49 , 1053–1060. 72. R. E. Rojas-P´erez, E. Cedillo-Portugal, P. Joseph-Nathan and E. Burgue˜no-Tapia, Nat. Prod. Comm. 2009, 4 , 757–762. 73. K. Monde, T. Taniguchi, N. Miura, S. I. Nishimura, N. Harada, R. K. Dukor, L. A. Nafie, Tetrahedron Lett. 2003, 44 , 6017–6020. 74. K. Monde, T. Taniguchi, N. Miura, P. Kutschy, Z. Curillov´a, M. Pil´atov´a, J. Mojˇziˇs, Bioorg. Med. Chem. 2005, 13 , 5206–5212. 75. T. Taniguchi, K. Monde, K. Nakanishi, N. Berova, Org. Biomol. Chem. 2008, 6 , 4399–4405. 76. K. Monde, A. Nakahashi, N. Miura, Y. Yaguchi, D. Sugimoto, M. Emura, Chirality 2009, 21 , S110–S115. 77. M. Emura, Y. Yaguchi, A. Nakahashi, D. Sugimoto, N. Miura, K. Monde, J. Agric. Food Chem. 2009, 57 , 9909–9915. 78. Y. Yaguchi, A. Nakahashi, N. Miura, D. Sugimoto, K. Monde, M. Emura, Org. Lett. 2008, 10 , 4883–4885. 79. H. M. Min, M. Aye, T. Taniguchi, N. Miura, K. Monde, K. Ohzawa, T. Nikai, M. Niwa, Y. Takaya, Tetrahedron Lett. 2007, 48 , 6155–6158. 80. F. J. Devlin, P. J. Stephens, B. Figad`ere, Chirality 2009, 21 , E48–E53. 81. B. Figadere, F. J. Devlin, J. G. Millar, P. J. Stephens, Chem Comm. 2008, 9 , 1106–1108. 82. K. Mori, Tetrahedron 1975, 31 , 1381–1384. 83. C. S. Ashvar, P. J. Stephens, T. Eggimann and H. Wieser, Tetrahedron: Asymmetry 1998, 9 , 1107–1110. 84. P. R. Lassen, D. Mondrup Skytte, L. Hemmingsen, S. Feldbæk Nielsen, T. B. Freedman, L. A. Nafie, S. B. Christensen, J. Nat. Prod . 2005, 68, 1603–1609. 85. H. Izumi, A. Ogata, L. A. Nafie, R. K. Dukor, J. Org. Chem. 2008, 73 , 2367–2372. 86. J. He, A. G. Petrovic, S. V. Dzyuba, N. Berova, K. Nakanishi, P. L. Polavarapu, Spectrochim. Acta Part A 2008, 69 , 1213–1222. 87. N. H. Andersen, N. J. Christense, P. R. Lassen, T. B. Freedman, L. A. Nafie, K. Stromgaard, L. Hemmingsen, Chirality 2010, 22 , 217–223. 88. G. Shanmugam, P. L. Polavarapu, D. Gopinath, R. Jayakumar, Biopolymers: Peptide Science 2005, 80 , 636–642. 89. L. A. Bodack, L. A. Freedman, B. Z. Chowdhry, L. A. Nafie, Biopolymers 2004, 73 , 163–177. 90. R. H. Cichewicz, L. J. Clifford, P. R. Lassen, X. Cao, T. B. Freedman, L. A. Nafie, J. D. Deschamps, V. A. Kenyon, J. R. Flanary, T. R. Holman, P. Crews, Bioorg. Med. Chem. 2005, 13 , 5600–5612. 91. B. Brzezinski, K. Olejnik, S. Paszyc, J. Mol. Struct. 1990, 220 , 261–268. 92. L. Huang, Y. K. Si, G. Snatzke, D. K. Zheng, J. Zhou, Coll. Czech Chem Comm. 1988, 53 , 2664–2666. 93. T. B. Freedman, X. Cao, R. Oliveira, Q. B. Cass, L. A. Nafie Chirality 2003, 15 , 196–200. 94. P. L. Polavarapu, N. Jeirath, T. Kurt´an, G. Pescitelli, K. Krohn. Chirality 2009, 21 , E202–E207.

12 DETERMINATION OF MOLECULAR ABSOLUTE CONFIGURATION: GUIDELINES FOR SELECTING A SUITABLE CHIROPTICAL APPROACH1 Stefano Superchi, Carlo Rosini, Giuseppe Mazzeo, and Egidio Giorgio

12.1. INTRODUCTION The knowledge of the absolute configuration (AC) of a chiral molecule is a fundamental prerequisite for the understanding and prediction of its interactions with chiral systems both at molecular and supramolecular level. The tight relationship between AC and bioactivity of chiral drugs and metabolites as well as the importance of chiral recognition in catalysis and materials is in fact well known [1]. The AC assignment is therefore an important problem commonly faced by researchers involved in medicinal chemistry, natural products chemistry, asymmetric catalysis, and materials chemistry. Many different methods to solve this fundamental problem are available to the chemist [2], but, to date, no one can be considered fully general. The classical chemical correlation procedure requires a long process, and often it involves complex stereo-controlled chemical reactions for transforming the compound under analysis into another having known AC. The anomalous X-ray scattering method [3], although being a very reliable method, presents some limitations, requiring the presence of “heavy” atoms on the molecule, crystalline products, and the availability of single crystals, as well as expensive and complex equipments. NMR and chiroptical spectroscopy can instead allow configurational assignments in solution. The NMR approach [4], requiring the derivatization of the molecule with a chiral auxiliary, can, however, be considered an indirect method and essentially empirical. In this framework the chiroptical spectroscopies like electronic circular dichroism (ECD), vibrational circular dichroism (VCD), and optical rotatory dispersion (ORD) can 1

Dedicated to the memory of Professor Carlo Rosini (1948–2010), an unforgettable friend and mentor, in recognition of his outstanding contribution to the development of chiroptical spectroscopy.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

421

422

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

provide a powerful tool for the AC assignment to chiral molecules [5]. A great advantage of these techniques is the possibility to achieve the configurational assignment in solution, then allowing to treat noncrystalline compounds, for which X–ray methods are not applicable. Moreover, the AC assignment is often obtained in a nonempirical fashion, thereby ensuring reliable results. Although ORD and ECD spectroscopies date back to the nineteenth century, in the last decades they had undergone a tremendous development from the experimental, instrumental, and theoretical-computational point of view. In this context the advent of computational approaches to chiroptical spectra interpretation has constituted the real breakthrough during the past few years [6] (see Chapters 22 and 23 in volume 1). This is because the quantum mechanical calculations of ORD and ECD have become much more accessible, and by that they have broadened the prospects for application of these already well-established techniques. Moreover, recent development of commercially available VCD spectrometers, along with implementation of quantum mechanical approaches for VCD calculation in computational packages, has extended the resources for molecular structural analysis with this new powerful technique [7] (see Chapters 23 and 24 in volume 1). The large number of methodologies and techniques now available can offer a wide choice, but, on the other hand, they may confuse nonspecialist in choosing the “right” method for his specific purposes. The aim of this chapter is then to provide a practical guide to nonspecialists in chiroptical spectroscopy, but who are involved in synthetic organic, medicinal, and natural product chemistry, on how to select the most suitable chiroptical techniques for absolute configurational assignment of a particular chiral substrate. For a quick and preliminary configurational assignment, the “first” choice is often the “simplest” and most easily available one, which is not necessarily the most rigorous one. Of course, approaching a structural problem by more than a single technique often allows us to remove a possible ambiguity in the configurational assignment and, by that, to reach a more reliable result [5c, 8]. Accordingly, the scope and limitations of main chiroptical techniques employed for AC assignments will be briefly reviewed. In the discussion we will provide a comparison of the different qualitative versus computational approaches. Finally, some guidelines and representative examples about how to make decision about the most suitable method will be provided, taking into account the chemical structure and other factors, such as substrate stability and available amount.

12.2. THE TECHNIQUES Whereas the previous chapters of this book review in detail the theoretical principles, instrumentation, experimental measurements, and applications of almost all types of chiroptical spectroscopies, here we will focus only on some of them, namely ORD, ECD, and VCD, which are today the most commonly used chiroptical techniques for AC assignments. The Raman optical activity (ROA) [9] (see Chapter 6 in volume 1) and NIR-VCD (see Chapter 10 in Volume 1) will not be discussed because, despite their high potential, these spectroscopies are not widely used yet for absolute configurational assignments. Optical Rotation/Optical Rotatory Dispersion 1. For experimental measurements a simple and less expensive polarimeter is commonly used in organic chemistry laboratories. By applying a set of filters, it can

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

2. 3. 4. 5. 6.

7.

record optical rotation (OR) at a few wavelengths.2 An ORD attachment to a common ECD spectropolarimeter provides a more advanced, continuous recording of OR in broader spectral region. Small quantities (a few milligrams) of sample are usually required. Qualitative methods for assignment of the AC from the sign of some specific ORD bands are available (i.e., octant rule) (vide infra). ORD data can be calculated by several commercially available packages [10]. Medium-size molecules can now be treated by computational approaches.3 The ORD computational simulation provides the sign and order of magnitude of one or a few OR values. Reproduction of the spectral trend (sign/position of some ORs) can be achieved. In this case, more reliable results are obtained when a plane ORD curve (i.e., far from resonant area and corresponding Cotton effects (CEs)) is calculated. According to some authors, a reliable AC assignment at single wavelength can be obtained only for OR values greater than 30–40 units [11], while the calculations of OR values at several frequencies in general is more reliable [12, 13]. It does not require the presence of chromophores, so even UV–vis transparent molecules can be treated.

Electronic Circular Dichroism 1. The instruments for ECD are more expensive than the common polarimeter, yet they are also available at affordable price and quite common in the practice of organic chemistry and biochemistry laboratories. 2. Very small samples are usually required: with substances having very high molar ε values, even micromolar solutions can be used [5]. 3. Qualitative spectrum/structure correlations (i.e., sector and chirality rules) are available, allowing its use also by experimental organic chemists who are not familiar with quantum mechanical calculations and spectroscopy. A drawback for qualitative approaches is that they are sometimes empirical and then not fully reliable. From this point of view, the exciton chirality method [14] (see Chapter 4 in this volume) represents one of the most reliable and versatile tools to be used. 4. ECD data of small and medium-size molecules can be simulated by several commercially available packages [10]. 5. Computational reproduction of (almost) entire spectrum in UV–vis region is possible: sign, position (λ), and intensity of some (1–10) CEs. For reliable results at least 30–50 excited states must be calculated. 6. The presence of UV–vis absorbing moieties (chromophores) is required.

2

For recording a full ORD spectrum it is necessary that an accessory (≥25,000 USD) be implemented on an ECD spectropolarimeter. 3 For instance, the single-wavelength (λ = 589 nm) TDDFT calculation of OR (B3LYP/6-31G*) for a single conformer of the molecule of benzylidene benzotricamphor [C51 H48 O3 (MW 708)] requires (including geometry optimization and frequencies calculations) 229 h with a desktop computer having CPU Intel® Pentium® D 3.20 GHz, 2 GB RAM [G. Mazzeo, E. Giorgio, C. Rosini, F. Fabris, E. Fregonese, U. Toniolo, O. De Lucchi, Chirality, 2009, 21 , E86–E97]. The same calculation with a newer and more powerful PC (Intel® Xeon® Quadcore E5420 at 2.50 GHz, 4 GB RAM) requires 29 h.

423

424

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Vibrational Circular Dichroism 1. The VCD spectra require a more expensive instrument and a good expertise about the factors that may affect the measurements quality. The most common experimental problem is, for example, the baseline stability, an artifact that can be avoided with dual photoelastic modulator (PEM) instruments [15]. 2. Larger amounts of samples are generally required: some authors suggest samples of 50–100 mg [16], but nowadays the minimum quantity is 5–10 mg. Often both enantiomers and the racemic mixture are needed. Alternatively, careful solvent subtraction is required. 3. VCD data can be simulated by several commercially available packages [10]. Medium-size molecules can now be treated computationally. Since the simulation of a VCD spectrum requires only the knowledge of the electronic ground state, these calculations often are simpler than that of ORD/ECD. The simulated VCD spectrum leads to simultaneous reproduction of many (20–50) CEs of different S/N ratios; and in principle, it provides a better opportunity for a safe matching with the experimental data, a necessary step for a safe configurational assignment. 4. It does not require the presence of UV–vis chromophores, therefore even transparent molecules can be treated. It must be recalled that both ORD and ECD rely on electronic properties, namely, on the differential refraction and absorption of circularly polarized UV–vis radiation, respectively. Moreover, both methods are also closely related to each other through Kramers–Kronig (KK) transforms [17], which allow us to convert one into the other.4 It follows that, in principle, both spectroscopies can provide the same type of information. From the experimental point of view, however, ECD is to be preferred over ORD. In fact, ECD bands occur only in correspondence of an optically active absorption band, while ORD spectrum results from the sum of contributions from all the chromophores of the molecule, even those in the far UV. It follows that in ORD it is difficult to separate the contributions from different electronic transitions. Conversely, since in ECD overlaps occur only between very close absorptions, the spectral interpretation is much easier. However, an advantage of ORD over ECD is clearly seen when the compounds are devoid of chromophores absorbing at λ > 180 nm. These compounds are in fact ECD transparent, while still show an ORD curve. In such cases a practical alternative is provided by VCD, which, relying on vibrational transitions, does not require the presence of UV absorbing moieties on the molecule. In general, ECD methods are more sensitive, allowing microscale measurements. This is a relevant feature when dealing with natural products that are usually available in very small amounts. Moreover, these methods often allow us to reach a reliable configurational assignment simply by visual inspection of sign and position of specific diagnostic CEs. As already mentioned, the main limitation of ECD is the need of chromophores in order to get ECD signals intense enough to provide a reliable assignment. On the other hand, VCD allows us to treat also nonchromophoric molecules and gives rise to experimental/predicted spectral comparison on a much larger number of bands. 4

In theory, a complete transformation from ECD to ORD spectrum requires the knowledge of the full (λ from 0 to +∞) ECD spectrum. In practice, KK transform can be applied only to a part of the spectrum given that the accessible spectral window is usually limited to λ > 180 nm. In that case the resulting KK transformation of ECD into ORD takes into account only the contribution given by the section of the spectrum analyzed. The same happens when the reversed ORD to ECD transformation is done.

425

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

VCD

ECD

ORD

(aR) -Vanol B3LYP/6-311G(2d,2p)

(aR) -Vanol B3LYP/6-311G(2d,2p)

x3 (aR) -Vanol B3LYP/6-311G(2d,2p)

(+) -Vanol-expet

(+) -Vanol-experiment

1700

1500 1300 1100 Wavenumber

900 230

260

290 nm

320

(+) -Vanol-expet

350 350

450

550

650

nm

Figure 12.1. Calculated (top line) and experimental (bottom line) VCD, ECD, and ORD spectra of (+)-(aR)-VANOL. (Reprinted with permission from J. Org. Chem. 2009, 74, 5451–5457. Copyright 2009, American Chemical Society.)

Ph

OH

Ph

OH

(aR)-3,3'-diphenyl-[2,2'-binaphthalene]-1,1'-diol (VANOL)

Chart 12.1.

For interpretation of VCD spectra, the use of computational methods is, however, mandatory, making often the overall AC assignment more time-consuming, particularly in the case of conformationally flexible substrates. In Figure 12.1, experimental and calculated VCD, ECD, and ORD spectra of (aR)3,3 -diphenyl-[2,2 -binaphthalene]-1,1 -diol (VANOL) (Chart 12.1) are compared [18]. There we can clearly see the higher complexity of the VCD spectrum, which allows a comparison on a much larger number of bands. An interesting case in which the use of VCD helps to solve a complex structural problem is when the molecule has a chromophoric chiral moiety that gives rise to very intense CEs in both ORD and ECD, and by that causes a significant overlap with bands allied to other chiral moieties in the same molecule. The optical contributions, for example, of molecules containing both axial and central chirality are often difficult to analyze. In such cases, each chirality element can be more conveniently identified and characterized by VCD spectroscopy, so that a safe assignment of AC can be reached [19]. The main operational features of the chiroptical techniques are summarized in Table 12.1.

426

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 12.1. The Main Operational Features of the Chiroptical Techniques Technique ORD ECD VCD

Instrumenta Cost (103 USD)

Chromophores Required

Sample Amount

Qualitative

Computationalb

≥ 25 ≥ 65 ≥ 100

No Yes No

1–10 mg > μg > 10 mg

Yes Yes No

Yes Yes Yes

a

The instruments costs are mean values referring to year 2010. ORD and VCD calculations require quantum mechanical approaches, while calculations of ECD can be performed also by “classical” (coupled oscillators) and semiempirical approaches. b

12.3. THE APPROACHES In general, the assignment of the molecular AC by chiroptical spectroscopy relies on the comparison between the experimental spectrum and predicted spectrum for a given enantiomer. If both spectra agree, then the AC of the molecule corresponds to the one chosen a priori for the spectral prediction. In order to predict the spectrum for a given enantiomer, its preferred conformation(s) must be also known. In fact, AC, conformation, and spectrum are tightly interrelated and knowing two of these information, the third can be obtained (Figure 12.2). For this reason, chiroptical spectroscopies are also employed as probes of the molecular conformation [20]. Therefore the knowledge of both the molecular conformation(s) and the mechanism relating the AC with the chiroptical response are two crucial aspects that have to be addressed. A conformational analysis of the molecule can be performed either experimentally, via NMR spectroscopy, or theoretically, by molecular modeling and calculations (molecular mechanics, semiempirical or quantum mechanical). In the first case, however, only an averaged conformational situation is revealed, while a computational search can allow us to sort out any conformer of the molecule within a given energy window and to determine the relative conformers population. Regarding the spectral prediction, either a qualitative or a quantitative (computational) approach can be pointed out. In the first case, empirical or nonempirical rules have to be considered, both allowing prediction of some diagnostic spectral features for a given enantiomer in a given conformation—that is, for a single geometry. This is the case of the well-known chirality rules that allow us to predict the wavelength position and sign of some diagnostic CEs in the ECD spectrum: octant rule [21], benzene [22] and diene [23] chirality rule, Mislow biphenyl rule [24], exciton chirality method [14]. The advantages of the qualitative approaches rely on their ease and less time required to reach a decision, not requiring a precise spectral prediction. In this case a visual Absolute Configuration

Figure 12.2. Relationship Conformation (NMR, Calculations, Molecular models)

ORD/ECD/VCD Spectrum (Experimental ↔ Predicted)

between molecular structural properties (conformation, AC) and spectrum.

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

inspection of the spectrum allows for the configurational assignment. Moreover, since by qualitative methods only some specific molecular fragments or chromophores of the molecule are taken into account, in such case complex molecules can be investigated as well. However, a major problem by the qualitative approaches arises when more than a single significantly populated conformer is present and the predicted chiroptical response of some conformers is opposite in sign. In this case it is not possible to quantify the contribution of any single conformer to the overall experimental spectrum. A possible solution of this problem is either to (a) transform the compound under investigation into a more rigid derivative which exists as a single conformer or (b) apply a quantitative computational approach for conformational analysis and simulations of chiroptical properties. In fact, the quantitative approach provides a population and numerical values of the rotational strengths for any conformer, thus allowing the calculation of the net contribution by Boltzmann averaging over the populations. In some cases, however, if different conformers have chiroptical contributions of opposite sign, the weighted averaged value may become very small, and hence unreliable for any assignment. The quantitative (computational) approaches allow us to calculate the whole spectrum and therefore provide an estimate of the position, sign, and intensity of all bands allied to a specific structure and geometry. Many types of computational approaches have been described over the years, both quantum mechanical (semiempirical, ab initio) [6, 25]5 (see Chapters 22, 23, and 24 in Volume 1) and classical (coupled oscillators, DeVoe) [26, 27] (see Chapter 20 in Volume 1). The former provides more accurate answers whereas the latter, being computationally much less demanding, can be applied even to large systems. A complete computational treatment requires (a) conformational search and optimization, (b) assessment of the Boltzmann populations, (c) calculation of the chiroptical properties for each conformer, (d) Boltzmann averaging of the chiroptical properties, and (e) comparison with the experimental data. In some cases the solvent effect must be taken into account in the calculations [6e]. Therefore, there are two crucial computational steps for this approach: (a) the exact determination of the structure (geometry) and population of the main conformers and (b) the calculation of their chiroptical properties. In principle, ab initio methods could be applied for any kind of molecule within the reasonable size range. With these approaches, it is unnecessary to know a priori any property of the molecule, its fragments, and its chromophores (e.g., strength and direction of oscillators, group polarizabilities, etc.) or to have a suitable reference compound. However, as mentioned above, a very large size and a very high conformational flexibility can be prohibitive for application of the ab initio methods. In fact, the larger the number of atoms, the higher the number of electronic or vibrational states and transitions to be computed [28]. Since by the computation of electronic transitions the number of electrons is taken into account, the presence of heavy atoms can obviously strongly affect the computation complexity. With flexible molecules the calculation has to be repeated for each conformer and then a Boltzmann average has to be performed. It is therefore clear that the larger the number of accessible conformers, the more complex and time-consuming the spectral prediction. Often a chemical derivatization which provides a molecular “rigidification” is advisable, leading to a decrease in the number of accessible 5

In ab initio electronic structure calculation the Hartree–Fock (HF) method was the first widely used, followed by more accurate methods taking into account electron–electron repulsion correction like Density Functional Theory (DFT) and Coupled Cluster Theory (CC) [S.M. Bachrach, Computational Organic Chemistry, John Wiley & Sons, 2007]. For comparison of Time dependent DFT (TDDFT) and CC in chiroptical calculations see: T. D. Crawford, P. J. Stephens, Phys. Chem. A 2008, 112 , 1339–1345.

427

428

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

conformations that simplify the calculation [29, 30]. Alternatively, the problem of the high molecular mobility can be overcome by performing calculation of ECD spectra in the solid state and then taking into account the single conformation taken by X-ray analysis [31] (see Chapter 6 in this volume). Although computational methods can look as the ultimate solution for the AC assignment, their reliability depends on the right choice of many computational parameters and on the accuracy of the conformational analysis performed. For example, in a recent OR calculation study it has been shown that the use of different levels of theory for geometry optimization may lead not only to different conformers distribution, but also to minor differences in conformers’ geometry [32]. This study revealed in particular the critical effect of even subtle geometrical differences on the overall OR calculation result and hence on the correctness of configurational assignment. It moreover points out some warnings about the use of computational methods. In fact, in many cases, it is impossible to know a priori the right computational protocol and the best method, functional, basis set, and so on [6e]. Therefore, it is desirable that the results provided by one approach be validated by another independent method.

12.4. THE MOLECULAR STRUCTURE Of course, the choice of the suitable chiroptical method to determine the AC of a molecule depends on its structure. It mainly depends on the presence of one (or more) chromophore(s)6 or one (or more) functional group(s), which can allow the introduction of chromophoric moieties, and also on the molecular flexibility (i.e., the number of conformers). As far as the molecular flexibility is concerned, substrates can be roughly divided into rigid and flexible molecules. To the rigid or relatively rigid belong these molecules that adopt only one or a few conformations, but only one has a predominant population. In contrast, the flexible-type molecules exist as a mixture of many conformations, which renders the computational conformational analysis difficult and time-consuming. In order to guide the researcher to the choice of the “right” method, a graphical guide as a flowchart (Scheme 12.1) is provided. This flowchart leads to the “right” choice, intended as the most easily available one or the simplest one. Of course, often the “right” choice cannot be the only one. Other alternative ways perhaps not as simple may also lead to a reliable answer as well. All in this line, more than one approach can often provide more reliable results [5c, 8]; therefore, if possible, the confirmation of the first assignment by an independent method is advisable. In Scheme 12.2 the ECD is given as the “first” choice since it is one of the most available chiroptical approach and it allows also a simpler qualitative spectral analysis. Therefore, the first selection to be made is on the presence of chromophores on the molecule. If chromophores are not present in the molecule and cannot be introduced by chemical reactions, then the molecule is UV–vis transparent and ECD spectroscopy cannot be applied. The absence of chromophores also often leads to low OR values at the D sodium line, therefore sometimes making unreliable even assignment by OR calculation. In this case the “right” choice is then the measurement of the VCD spectrum 6

An organic chromophore can be defined as a part of the molecule containing a functional group or a combination of more functional groups displaying π electrons and responsible for electronic transition giving rise to absorptions in the UV–vis range.

429

Scheme 12.1.

ECD Computational approaches (DeVoe) (Quantum Mechanical)

NO

NO

More than two

Molecular structure

ECD Qualitative exciton coupling approaches (exciton chirality)

YES

Are there couplets in ECD spectrum?

YES

Is it rigid? Can be rigidified?

two

How many?

YES

Are there chromophores?

YES

one

YES

NO

Solid state ECD Quantum Mechanical approach

Is it possible to introduce a second chromophore?

Can they be introduced by chemical derivatization?

NO

ORD/ECD/VCD Quantum Mechanical approaches

NO

Use of Spectroscopic probes

ECD Qualitative approaches (sector/chirality rules)

YES

ORD/VCD Quantum Mechanical approaches

Is it rigid? Can be rigidified?

NO

430

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

and its comparison with a simulated one by quantum mechanical calculations. If a single chromophore with UV–vis absorption is present, the chiral molecule exhibits an ECD spectrum with one or more CEs of same or different sign. In such case the molecule can be treated either qualitatively by applying sector and chirality rules or by quantum mechanical calculation of ECD spectrum, depending on its molecular flexibility. In the case of flexible chromophoric molecules the conformational problem can be overcome by the quantum mechanical simulations of the ECD spectrum taken in the solid state then taking into account the single conformation presents in the crystal structure [31]. Alternatively, when the molecule possesses at least one suitable functional group, an auxiliary that works as a spectroscopic probe, giving rise in the ECD spectrum to CEs related to the AC, can be attached to the molecule. Such probes can be linked to the molecule by either noncovalent [33] or covalent [34] bonds, and the resulting CEs can be qualitatively interpreted by chirality rules. The presence of two chromophores (equal or different) allows the use of nonempirical exciton chirality method. An application of this method requires that the relative orientation of the chromophores and polarization of their interacting electric transition moments be known with certainty, and it also requires that a clear couplet feature (either degenerate or nondegenerate) be visible in the ECD spectrum. On the contrary, if the molecule is too flexible and cannot be chemically rigidified, a computational approach cannot be avoided, being the only way to take into account the individual contribution of the several conformers. Also, molecules with more than two chromophores cannot be easily treated by the qualitative approaches and also by the exciton chirality method, therefore computations will be more desirable. The right choice of the method is better exemplified by looking at some selected examples in which advantages, disadvantages, and limitations of the different approaches are discussed.

12.4.1. Chromophoric Molecules 12.4.1.1. Presence of Only One Chromophore. For compounds having a single chromophore (carbonyl, aryl, diene, enone, etc.), qualitative approaches, such as sector and helicity rules, have been developed. They link the sign of the ECD CEs allied to the main electronic transitions of the chromophore to the AC. These approaches, although widely employed, often present some ambiguity and can lead to incorrect assignments. For this reason, although qualitative methods are often the “first” choice for their simplicity and rapidity, they sometimes need to be supported by more rigorous computational analysis. For several chromophores, belonging to either inherently chiral and achiral type [35], empirical and nonempirical rules relating the AC to the sign of corresponding ECD CEs have been developed. “Sector rules” have been described for treatment of symmetric, dissymmetrically perturbed chromophores, while “chirality rules” allow the treatment of inherently chiral chromophores [21–24] (see Chapters 2–4 in this volume). Some selected recent examples concerning the carbonyl and benzene chromophores are discussed herein, aiming to show some limitations and solution of pertinent problems. For more examples see other relevant chapters in this book. The Carbonyl Chromophore. The oldest and most popular sector rule is the so-called “octant rule,” allowing us to assign the AC of saturated ketones by the sign of the CE allied to the n → π ∗ transition of the carbonyl around 300 nm [20, 21, 36] (see Chapter 2 in this volume). Although it was originally proposed on empirical grounds [37], it was later supported by theoretical studies on the origin of the ketone n → π ∗ CE [21, 38].

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

In order to apply the octant rule for determining the AC of a ketone, the conformation of the latter must be known. A major problem then arises when the molecule shows an equilibrium of several conformations, since the octant rule allows to predict the sign of the n → π ∗ band for each conformer but not its intensity. For this reason the octant rule has usually been applied to cyclic, conformationally defined ketones. Another problem of the octant rule concerns the priority assessment of substituents that fall in different, oppositely signed, octants. Bicyclo[3.3.1]nonanediones provide an interesting example, which highlights the problems and solutions of their AC assignments. By applying the octant rule to these bicyclic diketones, both carbonyl groups in the main conformations must be taken into account. The bicyclo[3.3.1]nonane-2,6-dione (1) (Chart 12.2) was obtained by kinetic resolution of the racemic mixture by baker’s yeast [39], and its AC was established by Berg and Butkus by applying the octant rule [40]. Taking the (1S ,5S ) enantiomer of 1 in its major chair–chair conformation and putting the C2 carbonyl into the octant, the situation depicted in Figure 12.3 is obtained. Here the only atoms not on nodal planes are C4, C9, C8, C7, and the oxygen. C4 and C9 cancel each other, while the oxygen and C7, being far from the carbonyl, give rise to a weak contribution. Therefore, the only effect to be taken into account is given by C8, which falls in a positive octant. Exactly the same situation is obtained when the C6 carbonyl is examined, given the C2 symmetry of the molecule. It follows that for (1S ,5S )-1 a positive CE allied to the n → π ∗ transition is expected in the ECD spectrum. The (+) enantiomer experimentally shows such positive CE, and therefore it has the (1S ,5S ) AC. In this case there is no significant ambiguity in applying the octant rule. The molecule has a largely dominant conformation, and for both carbonyls the same sign of the ECD band is predicted. This assignment was confirmed by chemical correlation [41] and, later, by Stephens et al. [42] by Time-Dependent Density Functional Theory (TDDFT) ECD and OR calculations. These authors performed a DFT conformational analysis of 1 at the B3LYP/6-31G* level, obtaining a 72:27 ratio of the chair–chair:chair–boat conformers. n → π ∗ C=O excitations and OR values were then calculated at the B3LYP/aug-ccpVDZ level for both conformers, and the final values of ECD and OR were obtained by the Boltzmann average of the calculated values. In bicyclo[3.3.1]nonane-2,7-dione (2) (Chart 12.2), MMFF94 calculations give again the chair–chair conformation as the major one [43]. In this case, however, the two carbonyls are not equivalent and the two different situations depicted in Figure 12.4 are obtained. The location of the major conformer into octants, placing each C=O chromophore into the origin of the octants, leads to opposite signs of the n → π ∗ band (Figure 12.4) for the two carbonyls. When the C2 carbonyl group of the (1R,5S ) enantiomer is located into octants, only C4, C9, and C8 do not lie on nodal planes, while the C7 carbonyl lies very close to a nodal plane. Consequently, the major perturber is the axial C8, which falls in a positive

O

O

2 4

1

7 6

3 4

1

8

9 5

2

3

O

8

9

7

5 6

O (1S, 5S)-1

(1R, 5S)-2

Chart 12.2.

431

432

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+

− 4

9

5

1 2

3

6

Figure 12.3. Location of (1S, 5S)-bicyclo[3.3.1]

8

−

+

7

nonane-2,6-dione (1) into octants placing the C2 carbonyl group at the origin of the octants.

octant. Therefore, for this enantiomer a positive CE allied to the carbonyl chromophore is predicted in the ECD spectrum. Interestingly, the placement of the C7 carbonyl group into the origin of the octants led to the prediction of the opposite CE and to the reverse AC for this enantiomer. In fact, on the right projection (Figure 12.4) the C1 and C2 atoms cancels the contribution of C5 and C4. Therefore, oxygen on C2 remains as the sole perturber since all the other atoms are on nodal planes. This projection predicts a negative sign of the CE due to the O perturber on the C2 atom, falling in a negative octant. The authors consider the latter effect weaker than the former and therefore predict a positive CE in the ECD spectrum for (1R,5S )-2. As a result, they assigned this AC to the (+)-enantiomer. Although this assignment was confirmed by chemical correlation [43], obtaining 2 from 1 of known AC, it is clear that in this case the octant rule alone could not provide a reliable result. The AC of 2 was later assigned by means of chiroptical spectroscopy by Stephens et al. [42] through ab initio TDDFT calculations of ECD and OR. DFT B3LYP/6-31G* conformational analysis gave the chair–chair conformer as the only populated one, and TDDFT B3LYP/aug-cc-pVDZ calculations provided both n → π * ECD band and OR values, allowing to reach at the same (1R,5S )/(+) relationship for 2 in a more reliable way. Therefore, in this case, when uncertainty and ambiguity result from application of the octant rule, the use of a computational ab initio approach can provide more confident results. The Benzene Chromophore. The presence of an aromatic ring is one of the most common structural features in organic molecules. Therefore, many efforts have been devoted

+

− 9

5

4

+

−

3 2

4

1 3

2

6 8 7

−

+

−

5

9

1

6

7

8

+

Figure 12.4. Location of (1R, 5S)-bicyclo[3.3.1]nonane-2,7-dione (2) into octants placing the C2 and C7 carbonyl groups (left and right projections, respectively) at the origin of the octants.

433

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

in developing methods for determination of AC of chiral aryl substituted molecules. To this end, various semiempirical benzene sector rules have been proposed relating the sign of the ECD CEs allied to the benzene 1 La [44] and 1 Lb [22] transitions [45] with the AC of a phenyl- or benzyl-substituted stereogenic center. For substituted benzene compounds the benzene chirality rule [46] is instead operative (see Chapter 3 in this volume). The most widely used rule for the unsubstituted benzene chromophore is the “benzene sector rule” [22], which relates the 1 Lb ECD CE with the AC at the benzyl carbon. This rule has been successfully applied to many phenyl- and benzyl-substituted compounds, such as carbinamines, carbinamine salts, and carbinols, as well as for assignment to compounds of unknown AC [47]. It nevertheless failed with simple chiral alkyl benzenes like compounds 3–6 shown in Chart 12.3. For these compounds, all having (R) AC, the benzene sector rule predicts a negative CE allied to the 1 Lb transition. On the contrary, only 6 display a negative CE, while in all other cases positive CEs are observed in the 250- to 280-nm range. For these compounds an ab initio ECD calculation instead provided the right answer [48]. MMFF conformational search for each compound and DFT optimization at the B3LYP/6-31G(d) level was performed. The results showed three populated conformers for 3 and 5, five conformers for 4, and only a single conformation for 6. TDDFT-simulated ECD spectrum for each conformer at the B3LYP/TZVP level, followed by Boltzmann averaging, led to finely reproduced 1 Bb (below 200 nm) and 1 La (205–225 nm) ECD bands, while some disagreements were evident for the 1 Lb bands. This analysis clearly point out the complexity of ECD pertinent to phenyl chromophore. Not surprisingly, more reliable results were obtained after applications of computational approach. Other Chromophores. Several other semiempirical rules have been proposed for absolute configurational assignment such as sector rules for lactones [49], oxiranes [50], thioamides [51], β-lactams [52], and γ -lactams [53] as well as helicity rules for α, βunsaturated ketones [54, 55], dienes [23, 56], disulfides [57], and biphenyls [24, 58] (see Chapter 2 in this volume). In general, AC assignment of these compounds can also be achieved by a theoretical treatment [13c, 59]. 12.4.1.2. Presence of Two Chromophores. Molecules that contain two identical chromophores (i.e., being “dimeric”) with electrically allowed transitions and which do not exchange electrons can be treated by the coupled oscillator or exciton chirality

3

4

5

N

N (–)–(R,R)–7

Chart 12.4.

6

Chart 12.3.

434

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

approach. This treatment finds its origin in the seminal papers of Kuhn [60], Kirkwood [61], and Moffitt [62] on the coupled oscillator model, in both the classical and quantum mechanical formulation. Its first application for determination of the AC of a small molecule, the natural alkaloid calycanthine, was reported by Mason [63]. A major breakthrough in this approach is however due to the introduction by Harada and Nakanishi in early 1970-ies of the so-called exciton chirality method [14] (see Chapter 4 in this volume). Because it is nonempirical approach, the method allows us to establish (safely and qualitatively) a correlation between ECD spectrum and molecular AC, avoiding any kind of calculations. Here, a few examples that point out some limitations of the method and the way to solve them are reported. Two Identical Chromophores. Some failures of the exciton chirality rule have been reported in the literature, but very often these defeats did not depend on the intrinsic nature of the model, but they simply derived from a wrong application of it. In many cases, wrong results are due to wrong placement of the transition dipoles. In fact, to correctly apply the exciton model, it is fundamental to know the right polarization direction of interacting dipoles. This aspect is clearly shown in the case of Tr¨oger’s base (7) (Chart 12.4), for which a (+)/(R, R) correlation was found by Mason et al. [64] by applying the exciton approach. More than 20 years later, Wilen et al. [65] found, by X-ray analysis, that the previous Mason’s assignment for 7 was wrong and that the (−)-enantiomer of 7 has instead (R, R) AC. This problem was later faced by Devlin and Stephens [66] by ab initio DFT calculation of VCD spectra. Equilibrium structures of 7 were calculated at the B3PW91/6-31G* and B3LYP/6-31G* levels finding a single C2 symmetric structure. Minor differences in bond lengths and angles were found using the two functionals. Also, mid-IR absorption and VCD spectra were calculated at the B3PW91/6-31G* level. The comparison of VCD rotational strengths calculated for the (R, R) AC showed excellent agreement with experimental intensities for (−)-7 defining unambiguously the correspondence (R, R)(−)/(S , S )-(+). This assignment is in agreement with the one obtained by X-ray structure, but opposite to the Mason one deduced by the coupled oscillator approach, and it seems to reveal an intrinsic limitation of the exciton model. Actually, even a quantitative coupled oscillator DeVoe calculation [26] in which 7 was treated, similarly to Mason, as a system of two identical aniline chromophores, provided a wrong result [67]. More accurate analysis on the structure of the involved chromophore allowed to determine that in 7 a distorted aniline chromophore is actually present and that this distortion causes a significant change in the polarization directions. Then repeating the DeVoe calculation with the correct oscillators direction and intensity the ECD spectrum of 7 was satisfactorily reproduced [67]. This demonstrates that the DeVoe coupled oscillator approach and the exciton chirality model work well, obviously, only if a correct description of the spectroscopic parameters is used. A main problem in applying the exciton chirality approach is given by the molecular mobility. In fact, when at room temperature the molecule displays more significantly populated conformations, it is difficult to determine the expected couplet feature by summing each conformer contribution. This problem can be overcome by transforming the original flexible bis-chromophoric molecule in a rigid, conformationally defined, derivative. This kind of approach has been followed to determine the AC of anti -1,2-diarylethane-1,2diols [68]. These acyclic diols were transformed into the corresponding 2,2-dimethyl1,3-dioxolanes, where the two hydroxyls form a five-membered ring. Since such ketals are conformationally rigid, the relative position of the two aryl moieties is fixed and can

435

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

be easily determined. Futhermore, the sign of the CD exciton couplet reveals the chiral twist between the two interacting aryl chromophores, and by that the AC of starting diol. For example, in the dioxolanes of (1R, 2R)-1,2-di(4-substituted)phenylethane-1,2diols, both the longitudinally oriented aromatic transitions 1 Ba,b at 180–210 nm and 1 La at 210–240 nm form a positive chiral twist giving rise to a positive exciton couplet (Scheme 12.2), as experimentally observed. Often the positive couplet allied to the 1 La transition appears strongly unsymmetrical or only a single branch of the couplet is visible. This phenomenon is due to an overlap of the negative high-energy branch of such a couplet (centered at ∼220 nm) with the positive low-energy branch of the more intense positive couplet due to the 1 Ba,b transition (at ∼190 nm), giving rise to a reduction or even to a cancellation of the negative component of the 1 La couplet. By this method the absolute configurations of many anti -1,2-diarylethane-1,2-diols symmetrically and nonsymmetrically substituted with phenyl and naphthyl moieties have been assigned in a straightforward manner [68]. Two Different Chromophores. The exciton coupling approach also works when two different chromophores are present in the molecule and the directions of their excitonically coupled dipole transition moments are known. This is the case, for example, of optically active aryl alkyl sulfoxides 8–10 (Figure 12.5) in which the aryl and sulfoxide chromophores interact each other [69, 70]. The compounds 8–10 exemplify a type of structurally close compounds where more caution regarding their similarity will be advisable. In fact, some of them may possess a different degree of complexity if more than a single conformation is present and if more than two electric dipole moments have to be taken into account. In these compounds the absorption allied to the σ → σ ∗ transition of the S=O chromophore occurs at 210 nm, indicating that this chromophore is not conjugated with the aryl one, an effect due to the steric hindrance of the peri hydrogen. Therefore the sulfoxide and the naphthalene chromophore are reciprocally isolated, thus allowing us to apply the exciton coupling method. The transition dipoles to be taken into account are then the allowed σ → σ ∗ transition of the S=O chromophore polarized along the C1 → C2 direction of the C(1)(S=O)C(2) moiety and the long axis polarized 1 B transition of the naphthalene chromophore at 220 nm. Molecular mechanics calculations for (S )-1-naphthyl methyl sulfoxide (8) showed an E /Z equilibrium almost completely shifted toward the E conformer (Figure 12.5). Therefore the exciton chirality rule can be applied to this conformer where in (S )-8 the dipoles of the sulfoxide σ → σ ∗ and naphthalene 1 B transitions define a negative chirality, then giving rise to a negative couplet in the ECD spectrum (Figure 12.6a). The (−) enantiomer of 8 shows indeed a negative ECD coupling, thus leading to the (S )/(−) correlation [69].

HO

O

OH

O X

H

O O

X

X (R,R)

X

X (R,R)

H X POSITIVE chirality Positive Couplet in ECD

Scheme 12.2. Transformation of (R, R)-1,2-di(4-X-phenyl)-1,2-diols in the corresponding 2,2dimethyl-1,3-dioxolanes and application of the exciton chirality rule to the latter.

436

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Me – S O

–O

+

Me

+

S

(S)-8,Z

(S)-8,E Me + – S O

–O

Me

+

S

Me

Me

(S)-9,E

(S)-9,Z

Me – S O

–O

+

Me

+

S

Figure 12.5. Equilibrium between the (E) and (S)-10,E

(S)-10,Z

(a)

(Z) conformations of sulfoxides (S)-8–10.

(b)

Figure 12.6. Exciton chirality defined by the allowed 1 B naphthalene and S=O transitions in (a) E-conformer of (S)-8 and (b) Z-conformer of (S)-9. (See insert for color representation of the figure.)

In the similar compound 1-(2-methylnaphtyl) methyl sulfoxide (9) contrary to what observed in 8 the Z conformer is the most stable one (Figure 12.5), having a 70:30 Z :E ratio [70]. In the Z conformation the sulfoxide σ → σ ∗ and naphthalene 1 B transitions define an opposite chirality in respect to the E one and therefore for (S )-9 a positive chirality is observed and then a positive couplet is expected in the ECD spectrum. This indicated an (S )-configuration for the (−) enantiomer (Figure 12.6b). This qualitative analysis does not take into account the contribution of the minor conformer which may be not negligible, therefore for a reliable assignment a quantitative DeVoe calculation was performed taking both conformers as input geometries and averaging the results over the relative populations. It was found a good agreement between the calculated spectrum

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

for (S )-9 and the experimental one for the (−) enantiomer. Further confirmation for AC was obtained by ab initio calculation of its VCD spectrum [71]. The third compound of this series, the 9-phenanthryl methyl sulfoxide (10), although apparently very similar to the previous ones, was actually more complex and required a different treatment [69]. In this compound the chromophores involved display many complex transition dipoles interactions. Therefore, even if only two chromophores are present, the two dipole treatment of the exciton chirality approach is not suitable to provide reliable results. The ECD spectrum of 10 shows a number of bands, no clear exciton couplets appear, and in correspondence to the strongly allowed band of the phenanthrene chromophore at about 250 nm, only a weak Cotton effect is measurable. All these facts hamper a simple application of the exciton chirality approach and therefore DeVoe calculations were undertaken. Molecular mechanics calculations afforded a 90/10 distribution of the E /Z conformers (Figure 12.5). To describe the phenanthrene chromophore transitions, a series of dipoles were placed at 250 nm (long-axis polarization), 240 nm (short-axis polarization), and 205 nm (short-axis polarization) as suggested by some CNDO/S-CI calculations. For the sulfoxide transition at 210 nm a single oscillator polarized along the CMe –CAr bond was used. A weighted average of the spectra of the two E and Z conformers afforded the calculated ECD spectrum which well reproduced the main features of the experimental one, providing the configurational correlation (−)/(S ). Exciton coupling between different chromophores also arises when, in order to apply the exciton chirality approach, a second chromophore is added by chemical reaction to a monochromophoric substrate. One example of this type is provided by the transformation of 1-arylethane-1,2-diols to the corresponding biphenyl boronates [72]. 1-Arylethane-1,2diols are acyclic diols having only one chromophoric group, the aryl one. Therefore in order to apply the exciton chirality approach, a second chromophore having well-defined electrically allowed transitions and able to couple with the aryl one has to be introduced. Moreover, a great simplification of the analysis is achieved if a rigid derivative is obtained. The diols were then transformed into the corresponding 4-biphenylboronates (Scheme 12.3), thus introducing a second chromophore and, at the same time, transforming acyclic diols into cyclic conformationally defined derivatives. In biphenylboronates, due to sp 2 hybridation of the boron atom, no new stereogenic centers are formed and a strong UV absorption at ∼260 nm, allied to the biphenyl long-axis 1 La transition, is displayed. The aryl and the biphenyl transition dipoles are placed in a fixed and rigid relative disposition whose chiral twist is determined only by the AC of the benzylic stereogenic center of the diol and revealed by the sign of the exciton couplet. In 1-arylethane-1,2diols having (R) AC at the benzylic carbon the aryl and the biphenyl moiety define, in the boronate, a negative chirality and then a negative CE is expected in correspondence to the 1 La biphenyl band at 260 nm (Scheme 12.3). To summarize, boronate approach can allow for the determination of AC at a benzylic stereogenic center of chiral diols, provided that there is no conformational ambiguity in the interpretation of the biphenyl exciton couplet at 260 nm. Many other examples of this type have been described in the literature [14], where the molecule under study contains only one chromophore and a functional group, such as hydroxyl and amino groups, which allows an introduction of a second chromophore. In this case a nondegenerate coupling occurs (see Chapter 4 in this volume). In a special case, if one of the chromophores absorbs below 200 nm (for example, the double bond

437

438

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R OH HO

OH

Ar

R

O

B

O

OH

CHCl3, 4A MS

B

1L (260 a

nm)

1B

negative chirality

negative CE at 260 nm

Scheme 12.3. Transformation of (R)-1-arylethane-1,2-diols in the corresponding biphenylboronates and application of the exciton chirality rule to the latter.

in allylic alcohols or amines), only a single CE as the longer wavelength part of exciton couplet can be observed. Yet, even in such case the sign of this CE revealed the AC of substrate. 12.4.1.3. Presence of More than Two Chromophores. Systems with more than two chromophores cannot be solved by simple two dipole treatments of exciton chirality approach. The (−)-cupressoflavone (11) (Chart 12.5) is an example of such systems. The UV spectrum of (−)-11 exhibits two intense π –π ∗ bands at 324 and 273 nm. In correspondence of the first UV band at 324 nm, positive and negative CEs at 362 and 326 nm, respectively, appear in the ECD spectrum. The second UV band at 273 nm is instead associated with a negative ECD shoulder around 300 nm and a positive CE at 267 nm. At first the authors attempted an interpretation of these spectral features on the basis of the exciton chirality model. These CEs were then initially considered as two oppositely signed couplets deriving from exciton chirality defined by two allowed transitions located on each flavone monomer. The UV bands at 324 and 273 nm were in fact assigned to long-axis polarized transitions of p-methoxycinnamoyl and p-methoxybenzoyl chromophores, respectively. On the basis of such qualitative twodipole treatment, an aS configuration was at first assigned to biflavone (−)-11 [73]. In a subsequent study, Harada et al. [74] determined, by π -electron SCF-CI-DV MO calculations, energy position and rotational strengths of several π –π ∗ transitions of 11, thus obtaining its absorption and ECD spectra. Comparing the calculated and experimental ECD spectrum of (−)-11, they showed that the correct AC of (−)-11 is indeed (R).

OH

H3CO

O

O OCH3 OCH3

H3CO

O

OH

O

(–)–(R)–11

Chart 12.5.

439

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

These authors stated that the first two-dipole treatment gave a wrong result because they did not take into account that in the 400- to 250-nm range there are several allowed electronic transitions with different polarizations [74]. Their interactions give rise to several couplings whose overlap does produce multiple bands of opposite sign; therefore the overall CD has a complicated profile. Successively, the ECD spectrum of (−)-11 was reproduced by coupled oscillators DeVoe calculations [27b] demonstrating that also this simplified approach, which allow us to treat the interaction of several dipoles simultaneously, can afford the correct answer and constitutes a practical alternative to MO calculations.

12.4.2. Transparent (Nonchromophoric) Molecules 12.4.2.1. Presence of No Chromophores but One or Two Functional Groups. Molecules that do not absorb in the accessible UV–vis range are classified as nonchromophoric, transparent compounds and ECD analysis cannot be employed directly to them. However, in such case a suitable substrate with UV–vis absorption can be employed as auxiliary able to provide an ECD spectrum useful for determination of AC. Such an auxiliary can be considered as a chiroptical chromophoric “probe” for determination of molecular chirality. When the molecules are also endowed with a high molecular flexibility, the derivatization with the probe should also reduce the conformational mobility of the molecule. For transparent molecules with two derivatizable functional groups (diols, aminoalcohols, diamines, etc.), the assignment of AC via ECD can be performed by double derivatization with suitable chromophores. A typical application of the exciton chirality rule is in fact the AC assignment of diols by transformation in the corresponding dibenzoates [75], a method extended to a number of transparent bifunctional compounds and to many natural products (see Chapter 4 in this volume). This approach is particularly straightforward when dealing with cyclic, conformationally defined derivatives, but presents severe limitations with flexible, conformationally mobile compounds, in which multiple accessible conformations (and then dipole orientations) must be taken into account. In the second case a practical solution can be provided by the transformation of the diol in a cyclic, conformationally defined derivative. Following this concept, the AC assignment to aliphatic non-chromophoric diols has been approached by introduction of a flexible bridged biphenyl moiety as a probe of the molecular chirality [34a, 76]. Alkyl and aryl substituted diols were transformed in the corresponding biphenyl dioxolanes (Scheme 12.4), thus obtaining a pair

OCH3 OCH3 OH OH

O P

R1 *

n

* R2

n = 0,1,2 R1, R2 = aryl, alkyl, H

O negative CE at 250 nm

R2 * n * R1

O M O

R2 * n * R1

positive CE at 250 nm

Scheme 12.4. Transformation of 1,n-diols in the corresponding biphenyldioxolanes. Equilibrium between P and M twisted dioxolanes.

440

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of diastereoisomers having, respectively, P and M twist of the biphenyl moiety. In these compounds the low rotational barrier (∼14 kcal/mol) of the biphenyl allows, at room temperature, a thermodynamic equilibrium between the diastereoisomers. Therefore the most stable of them is also the major one. The mechanism for chirality transfer from the chiral diol to the biphenyl was clarified, thus establishing a direct relationship between the diol AC and the preferred biphenyl torsion. Moreover, the twist of the biphenyl could be easily determined from the ECD spectrum where the biphenyl P torsion is allied to a negative Cotton effect related to the so-called A band of the biphenyl chromophore at 250 nm, while an M torsion gives rise to a positive A band at the same wavelength [24, 77]. Therefore, by simply looking at the sign of the A band in the CD spectrum, it is possible to identify the biphenyl torsion and then the AC of a particular diol. This method proved to be simple, straightforward, and general with mono- and di-substituted 1,2-, 1,3-, and 1,4-diols, both cyclic and acyclic. The conformational rigidification of acyclic diols through transformation in cyclic chromophoric dioxolanes proved to be a useful tool also for simplifying the computational approaches for AC assignment [30]. By reacting flexible and UV–vis transparent synand anti -1,2- and 1,3-diols with fluorenone dimethyl acetal, the corresponding ketals were obtained (Scheme 12.5). These ketals are conformationally well-defined (only one conformer in most cases). They exhibit medium to high OR values and ECD spectra with several (up to five) CEs in the 350- to 200-nm range, due to valence shell π → π ∗ transitions. These features have allowed simulation of the chiroptical properties of these compounds at the TDDFT/B3LYP/6-31G* level of theory. The simulated ECD spectra provided much better agreement with the corresponding experimental data than with the calculated OR values, where only a satisfactory agreement was found. The transformation of the diols in their fluorenone ketals also turned out to be advantageous in performing VCD analysis for AC assignment, strongly reducing the number of conformers [78]. The assignment of AC to transparent monofunctional compounds by ECD is a really challenging task because in this case there is no possibility to obtain a bis-chromophoric system and then to apply the exciton chirality model. The solution of this problem was found by using “probes” of the molecular chirality—that is, chromophoric moieties which, linked covalently or not covalently to the substrate, give rise to CEs in the ECD spectrum which can be related to the AC of the molecule under investigation. Typical examples of noncovalently bound chirality probes are the “bis-porphyrin tweezers” that are applicable either to bifunctional compounds, such as diamines, amino alcohols, and amino acids, or to monofunctional secondary alcohols, primary and secondary amines, and carboxylic acids, all usually without useful for ECD studies UV–vis MeO

OH R1

*

OMe

OH n

* R2

n = 0,1; R1, R2 = aryl, alkyl, H several conformers LOW ORs; ECD transparent

R1

O

O

*

* R2 n

only one conformer OR values 40–50 degrees intense ECD spectrum

Scheme 12.5. Transformation of chiral 1,n-diols in the corresponding ketals with 9-fluorenone. Conformational and chiroptical properties of the derivatives.

441

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

bands (Chapter 4 in this volume) [33]. In this approach, upon complexation between a chiral substrate as a guest and an achiral bis-metalloporphyrin as a host, a chirality transfer from the guest to the host takes place. The porphyrin moieties of the complex adopt mutually twisted orientations with one more preferred helicity. It was found that the preferred porphyrin helicity of the complex, either clockwise or anticlockwise, depends on the AC of the guest. The porphyrins’ electronic transitions give rise to exciton coupling and then to a typical couplet feature in the ECD spectrum in correspondence of the porphyrin Soret bands (380–420 nm). Therefore, from the sign of the couplet that reflects the porphyrin twist, the AC of the chiral guest can be determined (Figure 12.7). In the past few years, several approaches have been developed for predicting the sign of porphyrin based exciton couplet of the host-guest complex. The earliest one takes into account the relative steric size of the substituents at the stereogenic center estimated on the basis of conformational energy values. Although a qualitative steric-size model provides a simpler and faster assignment of the AC, in some cases it may provide ambiguous results if the steric parameters for some groups are lacking or when factors, other than a steric one, are involved in determining the most stable conformation of the

(R) (R) (R)

O

O

O NH2

O

NH

Zn

Zn +

H Zn

Zn H

O

O

O

N

O

NH2

O

O

O

1-2Zn

conjugate/tweezer complex 15 N

N

10

20

Zn

+ 15'

N

N

N

N

5

10'

20'

Zn

Zn Z n

Zn Zn

N

N

O

−

5'

O

Favored conformation I 1-2Zn O

O

Predicted positive CD exciton couplet

Unfavored conformation II Predicted negative CD exciton couplet

Figure 12.7. Host–guest complexation between I-2Zn tweezers and a conjugate prepared from (+)-isomenthol as a representative of chiral secondary alcohols. The preferred positive helicity between the two porphyrins in the complex, defined as the projection angle between 5–15 and 50–150 directions, is governed by the AC of the isomenthol; in such a case the complex exhibits in solution a positive ECD exciton couplet (not shown). (Reprinted with permission from Chem. Commun. 2009, 5958–5980. Copyright 2009, The Royal Society of Chemistry.)

442

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

host–guest complex. The recently developed approach instead relies on more accurate computational conformational analysis by MC/OPLS-2005 [34, 79] and provides not only reliable results about AC, but also a deeper insight on the mechanism of this host–guest complexation process. Another approach for the assignment of AC to transparent monofunctional molecules is given by the use of covalently bonded biphenyl probes. This approach has been employed specifically for configurational analysis of 2-substituted chiral carboxylic acids [34b]. In this approach the chiral acids are converted in the corresponding biphenyl amides, whose flexible biphenyl moiety acts as a “probe” of the acid chirality giving rise to CEs related to the acid AC. In fact, in these derivatives, the stereogenic center of the acid induces a preferential torsion of the biphenyl moiety, detectable by the sign of the biphenyl A band (at 250 nm) in the ECD spectrum [24, 77] (Scheme 12.6). The mechanism of transfer of chirality from the acid stereogenic center to the biphenyl moiety has been analyzed and understood in detail, defining two different mechanisms operative in amides derived from 2-alkyl and 2-aryl substituted acids, respectively. Therefore, for both classes of compounds, it has been possible to define a model that allows us to predict, for a given acid AC, the preferred twist of the biphenyl moiety and, subsequently, the sign of the A band at 250 nm in the ECD spectrum, related to the biphenyl torsion. Following this protocol, to establish the AC of a 2-substituted chiral acid, it is simply necessary to prepare its biphenyl amide, to record the amide ECD spectrum, and to look at the sign of the A band. From the sign of such a band the torsion of the biphenyl can be deduced and then the acid AC. The main advantages of this approach are the reliability and simplicity of the AC assignment [34b]. In fact, in most of the cases, no conformational analysis is needed and the correlation between the ECD spectrum of the biphenylamide and the acid AC can be established while taking into account only the size of the substituents on the stereogenic center. When the steric size values are unknown or when some ambiguities arise upon the relative size of the substituents, then a simple molecular mechanics conformational analysis can allow a reliable spectrum–structure correlation. A variety of substrates, with different structures and multiple functionalities, has been investigated, in particular this approach displayed its full validity with alkyl and aryl substituted acids, with α-hydroxyacids and α-aminoacids [34b]. 12.4.2.2. Presence of No Chromophores and No Functional Groups. Saturated hydrocarbons are virtually UV-vis transparent, therefore for these compounds ECD and hence qualitative methods based on this technique cannot be applied. These compounds can then be treated either by transforming them into chromophoric derivatives, without affecting the molecular chirality, or by applying OR and VCD calculations [11c].

R2 R3 + R1

COOH

O

O NH

N

P

R1 negative CE at 250 nm

R3 R2

M

N R1

R3 R2

positive CE at 250 nm

Scheme 12.6. Transformation of 2-substituted carboxylic acids in the corresponding biphenylamides. Equilibrium between P and M twisted amides.

443

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

O H H

H

(S)-PHTP (S)-PHTP

(S)-2-keto-PHTP

Scheme 12.7.

(a) stereochemistry of (S)-anti-trans-anti-trans-anti-trans-perhydrophenylene (PHTP) (12). (b) Transformation of (S)-(12) in (S)-13.

An interesting approach to saturated hydrocarbons is provided by the case of perhydrotriphenylene (PHTP) (Scheme 12.7). In 1967 Farina and Audisio described the preparation of optically active anti-trans-anti-trans-anti-trans-perhydrophenylene (12) isomer [80]. The racemic compound 12 was transformed in the corresponding carboxylic acid by radical acylation with oxalyl chloride, followed by hydrolysis. The racemic acid of 12 was then resolved into enantiomers through transformation in its dehydro-abietylamine salt. Recovery of the acid followed by decarboxylation provided optically active 12. Originally, the AC of (−)-12 was assigned as (R) by the Brewster OR sign prediction method [81] Subsequently, the same authors converted an optically active acid of 12 to the 2keto derivative 13 and applied the octant rule to this ketone [82]. The enantiomer (+)-13 exhibited a positive band at 290 nm in ORD, therefore supporting (S ) configuration for (+)-13 and for (+)-12 from which it derived [83]. Yet, the AC assignment of 12 could not be considered as fully reliable because of the empirical nature of the Brewster method and also the uncertain identification of the structure of 13, and hence the octant rule assignment. Therefore it was more recently reinvestigated by Stephens et al. [84] by ab initio DFT calculations of VCD and ORD of 12. Monte Carlo conformational search by MMFF94 force field, followed by optimization using DFT at the B3LYP/6-31G* level, found that the only conformation of 12 significantly populated at room temperature is the one where all four cyclohexane rings adopt a chair conformation. The simulation of mid-IR and VCD spectra of (S )-12 by using DFT at the B3PW91/TZ2P level showed a very good agreement with the experimental spectra for the (+) enantiomer and, by that, confirmed the previous assignment. The TDDFT calculation of OR of 12 at several wavelengths also was performed by TDDFT at B3LYP/aug-cc-pVDZ level. Qualitatively, calculated and experimental rotations were in agreement in sign and in dispersion, supporting once more the (S )/(+) relationship for 12.

12.5. CONCLUSIONS Certainly, selection of most suitable chiroptical method for the purpose of safe absolute configurational assignment depends on many factors which, first of all, include equipment availability and specific chemical structure. The most important structural features of the molecule under investigation that must be taken into account are (a) the presence of chromophores, (b) the presence of functional groups that allow an introduction of chromophores if desirable, and (c) the level of substrate conformational flexibility. For transparent nonfunctionalizable molecules, VCD and ORD computational predictions are the method of choice, while for other compounds also ECD spectroscopy can be used.

444

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The advantages of the latter are its higher sensitivity and the possibility of achieving configurational assignments by qualitative methods, in principle easier and quicker than the computational ones. For the nonspecialist the qualitative methods, when applicable, often still remain the first choice. When the critical factors governing application of nonempirical exciton chirality method are taken into account, and no conformational ambiguity exists, the method can provide a straightforward and fast assignment of AC of a wide variety of chiral substrates even at a microscale level. In general, the computational approaches for AC analysis are becoming much more reliable and affordable, due to the fast methodological and technological advancement, although still some limitations given by the molecular size and conformational mobility exist. There is no doubt that the recent advance in chiroptical methods will open new opportunities for selection of the most suitable method for configurational assignment, or even new ways to apply routinely a few suitable methods instead of one with unparallel efficiency.

REFERENCES 1. (a) T. A. Baillie, K. M. Schultz, in The Impact of Stereochemistry on Drug Development and Use, H. Aboul-Enein, and I. W. Wainer, eds, John Wiley & Sons, New York, 1997, Chapter 2, pp. 21–43. (b) Chirality in Agrochemicals, N. Kurihara and J. Miyamoto, eds., John Wiley & Sons, New York, 1998. (c) Chirality in Industry, A. N. Collins, G. N. Sheldrake, and J. Crosby, eds., John Wiley & Sons, New York, 1997. 2. E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Organic Compounds, John Wiley & Sons, New York, 1994. 3. (a) H. Duddeck, in Houben Weyl Stereoselective Synthesis, G. Helmchen, R. W. Hoffmann, J. Mulzer, and E. Schaumann, eds., Thieme, Stuttgart, 1996, Vol. 1, p. 293. (b) H. D. Flack, G. Bernardinelli, Chirality 2008, 20 , 681–690. 4. (a) M. Seco, E. Qui˜noa, R. Riguera, Chem. Rev . 2004, 104 , 17–117. (b) T. Wenzel, Discrimination of Chiral Compounds Using NMR Spectroscopy, Wiley-Interscience, Hoboken, NJ, 2007. 5. (a) Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000. (b) N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. (c) P. Polavarapu, Chirality, 2008, 20 , 664–672. 6. (a) M. Pecul, K. Ruud, Adv. Quantum. Chem. 2006, 50 , 185–227. (b) T. D. Crawford, Theor. Chem. Acc. 2006, 115 , 227–245. (c) P. L. Polavarapu, Chem. Record 2007, 7 , 125–136. (d) T. D. Crawford, M. C. Tam, M. L. Abrams, J. Phys. Chem. A 2007, 111 , 12057–12068. (e) J. Autschbach, Chirality 2009, 21 , E116–E152. 7. (a) L. A. Nafie, T. B. Freedman, Vibrational optical activity, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000; Chapter 4, pp. 97–132. (b) P. J. Stephens, F. J. Devlin, Chirality 2000, 12 , 172–179. (c) P. L. Polavarapu, Int. J. Quant. Chem. 2006, 106 , 1809–1814. (d) P. J. Stephens, F. J. Devlin, J.-J. Pan, Chirality 2008, 20 , 643–663. 8. P. J. Stephens, D. M. McCann, F. J. Devlin, J. R. Cheeseman, M. J. Frisch, J. Am. Chem. Soc. 2004, 126 , 7514–7521. 9. (a) L. D. Barron, L. Hecht, Vibrational optical activity: From fundamentals to biochemical applications, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 23. (b) L. D. Barron, Molecular Light Scattering and Optical Activity, 2nd ed., Cambridge University Press, Cambridge, 2004. 10. (a) GAUSSIAN09, Revision A.02, Gaussian, Inc., Wallingford CT, 2009, www.gaussian.com. (b) DALTON, Release 2.0 (2005), www.daltonprogram.org. (c) TURBOMOLE, Version

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

11.

12.

13. 14.

15. 16. 17. 18. 19. 20. 21.

22.

23.

24.

25.

26. 27. 28. 29.

6.0, COSMOlogic GmbH & Co. KG, www.turbomole-gmbh.com. (d) ADF2009.01, SCM, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands, www.scm.com. (a) D. M. McCann, P. J. Stephens, J. R. Cheeseman, J. Org. Chem. 2004, 69 , 8709–8717. (b) P. J. Stephens, D. M. McCann, Chirality 2005, 17 (Suppl.) S52–S64. (c) D. M. McCann, P. J. Stephens, J. Org. Chem. 2006, 71 , 6074–6098. (a) E. Giorgio, C. Minichino, R. G. Viglione, R. Zanasi, C. Rosini, J. Org. Chem. 2003, 68 , 5186–5192. (b) E. Giorgio, R. G. Viglione, R. Zanasi, C. Rosini, J. Am. Chem. Soc. 2004, 126 , 12968–12976. J. Autschbach, L. Jensen, G. C. Schatz, Y. C. E. Tse, M. Krykunov, J. Phys. Chem. A 2006, 110 , 2461–2473. (a) N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy: Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, 1983. (b) N. Berova, K. Nakanishi, Exciton chirality method: Principles and applications, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 12, pp. 337–382. L. A. Nafie, H. Bujus, A. Rilling, X. Cao, R. K. Dukor, Appl. Spectr. 2004, 58 , 647–654. T. Furo, T. Mori, T. Wada, Y. Inoue, J. Am. Chem. Soc. 2005, 127 , 8242–8243. (a) H. A. Kramers, Nature 1926, 117 , 775–778. (b) R. D. L. Kronig, J. Opt. Soc. Am. 1926, 12 , 547–557. P. L. Polavarapu, A. G. Petrovic, S. E. Vick, W. D. Wulff, H. Ren, Z. Ding, R. J. Staples, J. Org. Chem. 2009, 74 , 5451–5457. P. L. Polavarapu, N. Jeirath, T. Kurtan, G. Pescitelli, K. Krohn, Chirality 2009, 21 , E202–E207. D. A. Lightner, J. E. Gurst, Organic Conformational Analysis and Stereochemistry from Circular Dichroism Spectroscopy, Wiley-VCH, New York, 2000. D. A. Lightner, The Octant Rule, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 10, pp. 261–303. (a) H. E. Smith, Chem. Rev . 1998, 98 , 1709–1740. (b) H. E. Smith, Applications of the benzene sector and benzene chirality rules, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 14, pp. 397–429. J. Gawronsky, Circular dichroism and chirality of dienes, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 11, pp. 305–335. (a) E. Bunnenberg, C. Djerassi, K. Mislow, A. Moscowitz, J. Am. Chem. Soc. 1962, 84 , 2823–2826. (b) K. Mislow, E. Bunnenberg, R. Records, K. Wellman, C. Djerassi, J. Am. Chem. Soc. 1963, 85 , 1342–1349. J. Sandstr¨om, Determination of conformations and absolute configurations by semiempirical calculation of CD spectra, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 16, pp. 459–490. (a) H. DeVoe, J. Chem. Phys. 1964, 41 , 393–400. (b) H. DeVoe, J. Chem. Phys. 1965, 43 , 3199–3208. (a) C. Rosini, M. Zandomeneghi, P. Salvadori, Tetrahedron: Asymmetry 1993, 4 , 545–554. (b) S. Superchi, E. Giorgio, C. Rosini, Chirality 2004, 16 , 422–451. Today molecules having M.W. up to 1000 u.m.a. can be computationally treated with common desktop PC. F. J. Devlin, P. J. Stephens, P. Besse, J. Org. Chem. 2005, 70 , 2980–2993.

445

446

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30. S. Tartaglia, D. Padula, P. Scafato, L. Chiummiento, C. Rosini, J. Org. Chem. 2008, 73 , 4865–4873. 31. G. Pescitelli, T. Kurt´an, U. Fl¨orke, K. Krohn, Chirality 2009, 21 , E181–E201. 32. G. Mazzeo, E. Giorgio, R. Zanasi, N. Berova, C. Rosini, J. Org. Chem. 2010, 75 , 4600–4603. 33. N. Berova, G. Pescitelli, A. G. Petrovic, G. Proni, Chem. Commun. 2009, 5958–5980. 34. (a) S. Superchi, D. Casarini, A. Laurita, A. Bavoso, C. Rosini, Angew. Chem. Int. Ed. Engl . 2001, 40 , 451–454. (b) S. Superchi, R. Bisaccia, D. Casarini, A. Laurita, C. Rosini, J. Am. Chem. Soc. 2006, 128 , 6893–6902. 35. A. Moskowitz, Tetrahedron 1961, 13 , 48–56. 36. C. Djerassi, Optical Rotatory Dispersion, McGraw-Hill, New York, 1960. 37. (a) W. Moffitt, A. Moscowitz, J. Chem. Phys. 1959, 30 , 648–660. (b) W. Moffitt, R. B. Woodward, A. Moscowitz, W. Klyne, C. Djerassi, J. Am. Chem. Soc. 1961, 83 , 4013–4018. 38. T. D. Bouman, D. A. Lightner, J. Am. Chem. Soc. 1976, 98 , 3145–3154. 39. G. Hoffmann, R. Wiartalla, Tetrahedron Lett. 1982, 23 , 3887–3888. 40. U. Berg, E. J. Butkus, Chem. Res., Synop. 1993, 116–117. 41. H. Gerlach, Helv. Chim. Acta 1978, 61 , 2773–2776. 42. P. J. Stephens, D. M. McCann, E. Butkus, S. Stoncius, J. R. Cheeseman, M. J. Frisch, J. Org. Chem. 2004, 69 , 1948–1958. 43. E. Butkus, S. Stoncius, A. Zilinskas, Chirality 2001, 13 , 694–698. 44. G. Snatzke, P. C. Ho, Tetrahedron 1971, 27 , 3645–3653 45. H. H. Jaff´e, M. Orchin, Theory and Applications of Ultraviolet Spectroscopy, John Wiley & Sons, New York, 1962, p. 242. 46. S. T. Pickard, H. E. Smith, J. Am. Chem. Soc. 1990, 112 , 5741–5747. 47. (a) M. Dawn, H. E. Smith, Chirality, 1993, 5 , 20–23. (b) A. Rumbero, I. Borreguero, J. V. Sinisterra, A. R. Alcantara, Tetrahedron 1999, 55 , 14947–14960. 48. G. Pescitelli, L. Di Bari, A. M. Caporusso, P. Salvadori, Chirality 2008, 20 , 393–399. 49. W. Klyne, P. M. Scopes, The carboxyl and related chromophores, in Fundamental aspects and recent developments in optical rotatory dispersion and circular dichroism, F. Ciardelli, P. Salvadori, eds., Heyden, London, 1973, pp 126–147. 50. (a) A. Gedanken, K. Hintzer, V. J. Schurig, Chem. Soc. Chem. Commun. 1984, 1615–1616; A. Rodger, J. Am. Chem. Soc. 1988, 110 , 5941–5945. (b) A. Gedanken, Chiroptical properties of alcohols, ethers and peroxides, in The Chemistry of Hydroxyl Ether and Peroxide Groups, Vol. 2, S. Patai, ed., John Wiley & Sons, New York, 1993. 51. M. Milewska, M. Gdaniec, H. Maluszynska, T. Polonski, Tetrahedron: Asymmetry 1998, 9 , 3011–3023. 52. H. Rehling, H. Jensen, Tetrahedron Lett. 1972, 27 , 2793–2796. 53. J. Frelek, I. Panfil, Z. Urbanczyk-Lipkowska, M. Chmielewski, J. Org. Chem. 1999, 64 , 6126–6134. 54. J. Gawronski, Tetrahedron 1982, 38 , 3–26. 55. J. Gawronski in The Chemistry of Enones, Eds. S. Patai, Z. Rappoport, John Wiley & Sons, New York, 1989, pp. 55–105. 56. (a) M. Duraisamy, H. M. Walborsky, J. Am. Chem. Soc. 1983, 105 , 3252–3264. (b) M. Duraisamy, H. M. Walborsky, Ibid . 3264–3269. (c) M. Duraisamy, H. M. Walborsky, Ibid . 3270–3273. (d) H. M. Walborsky, S. M. Reddy, J. H. Brewster, J. Org. Chem. 1988, 53 , 4832–4846. 57. (a) L. A. Neubert, M. Carmack, J. Am. Chem. Soc. 1974, 96 , 943–945. (b) R. W. Woody, Tetrahedron 1973, 29 , 1273–1283. 58. J. Gawronski, P. Grycz, M. Kwit, U. Rychlewska, Chem. Eur. J . 2002, 18 , 4210–4215.

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

59. See, for example: P. J. Stephens, D. M. McCann, F. J. Devlin, T. C. Flood, E. Butkus, S. Stoncius, J. R. Cheeseman, J. Org. Chem. 2005, 70 , 3903–3913. 60. W. Kuhn, Trans. Faraday Soc. 1930, 46 , 293–308. 61. J. G. Kirkwood, J. Chem. Phys. 1937, 5 , 479–491. 62. W. Moffitt, J. Chem. Phys. 1956, 25 , 467–478. 63. (a) S. F. Mason, Proc. Chem. Soc. 1962, 362–363. (b) S. F. Mason, G. W. Vane, J. Chem. Soc. (B) 1966, 370–374. 64. S. F. Mason, G. W. Vane, R. J. Wells, J. S. Whitehurst, J. Chem. Soc. (B) 1967, 553–556. 65. S. H. Wilen, J. Z. Qi, P. G. Williard, J. Org. Chem. 1991, 56 , 485–487. 66. F. J. Devlin, P. J. Stephens, Chem. Comm. 1999, 361–362. 67. E. Giorgio, PhD Thesis Universit`a della Basilicata (Italy), 2004. 68. (a) C. Rosini, S. Scamuzzi, G. Uccello-Barretta, P. Salvadori, J. Org. Chem. 1994, 59 , 7395–7400. (b) C. Rosini, S. Scamuzzi, M. Pisani Focati, P. Salvadori, J. Org. Chem. 1995, 60 , 8289–8293. (c) C. Rosini, G. P. Spada, G. Proni, S. Masiero, S. Scamuzzi, J. Am. Chem. Soc. 1997, 109 , 505–512. (d) M. I. Donnoli, P. Scafato, S. Superchi, C. Rosini, Chirality 2001, 13 , 258–265. 69. C. Rosini, M. I. Donnoli, S. Superchi, Chem. Eur. J . 2001, 7 , 72–79. 70. M. I. Donnoli, E. Giorgio, S. Superchi, C. Rosini, Org. Biomol. Chem. 2003, 1 , 3444–3449. 71. P. J. Stephens, A. Aamouche, F. J. Devlin, S. Superchi, M. I. Donnoli, C. Rosini, J. Org. Chem. 2001, 66 , 3671–3677. 72. (a) S. Superchi, M. I. Donnoli, C. Rosini, Org. Lett. 1999, 1 , 2093–2096. (b) S. Superchi, D. Casarini, C. Summa, C. Rosini, J. Org. Chem. 2004, 69 , 1685–1694. 73. M. Parveen, N. U. Khan, B. Achari, P. K. Dutta, Abstracts of Papers, 3rd International Symposium on Flavonoides in Biology and Medicine, Singapore, November 13–17, 1989; Abstract, p. 22. 74. N. Harada, H. Ono, H. Uda, M. Parveen, K. Nizam ud Din, B. Achari, P. K. Dutta, J. Am. Chem. Soc. 1992, 114 , 7687–7692. 75. (a) H. Uzawan, Y. Nishida, H. Ohrui, H. Meguro, J. Org. Chem. 1990, 55 , 116–122. (b) N. Harada, A. Saito, H. Ono, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, T. Kuriki, J. Am. Chem. Soc. 1991, 113 , 3842–3850. (c) N. Harada, A. Saito, H. Ono, S. Murai, H-Y. Li, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, Enantiomer 1996, 1 , 119–138. 76. P. Scafato, S. Superchi, Chirality 2010, 22 (1E), E3–E10. 77. P. Rashidi-Ranjbar, J. Sandstr¨om, J. Mol. Struct. 1991, 246 , 25–32. R. Isaksson, P. RashidiRanjbar, J. Sandstr¨om, J. Chem. Soc. Perkin Trans. 1991, 1147–1152. 78. G. Longhi, S. Abbate, P. Scafato, C. Rosini, Phys. Chem. Chem. Phys. 2010, 12 , 4725–4732. 79. (a) A. G. Petrovic, Y. Chen, G. Pescitelli, N. Berova, G. Proni, Chirality, 2010, 22 , 129–139. (c) Y. Chen, A. G. Petrovic, M. Roje, G. Pescitelli, Y. Yang, M. Kayser, N. Berova, G. Proni, Chirality 2010, 22 , 140–152. 80. M. Farina, G. Audisio, Tetrahedron Lett. 1967, 14 , 1285–1288. 81. J. H. Brewster, J. Am. Chem. Soc. 1959, 81 , 5475–5483. 82. M. Farina, G. Audisio, Tetrahedron 1970, 26 , 1827–1837. 83. M. Farina, G. Audisio, Tetrahedron 1970, 26 , 1839–1844. 84. P. J. J. Stephens, F. J. Devlin, S. Sch¨urch, J. Hulliger, Theor. Chem. Acc. 2008, 119 , 19–28.

447

PART III INORGANIC STEREOCHEMISTRY

13 APPLICATIONS OF ELECTRONIC CIRCULAR DICHROISM TO INORGANIC STEREOCHEMISTRY Sumio Kaizaki

13.1. INTRODUCTION Inorganic stereochemistry has made great progress associated with advances in chiroptical methods. The discovery of chiral structures in coordination compounds has not only played historical roles at the origin of coordination chemistry in the beginning of modern chemistry, but has also provided important clues in expanding it toward asymmetric catalysis, bioinorganic chemistry, and supramolecular chemistry in recent years. The first application of chiroptical spectra to inorganic stereochemistry was made by the measurement of optical rotation (OR) after the syntheses of chiral transition metal complexes by the “father of coordination chemistry” Alfred Werner in the early twentieth century [1], although this technique had been used for organic chemistry. A firm proof of the octahedral coordination theory by synthesis of a chiral “carbon-free” compound is the well-known tetranuclear hexol-type complex [Co{(μ-OH)2 Co(NH3 )4 }]6+ , which was resolved into enantiomers in an epoch-making experiment [2]. The separated complexes were shown to possess equal and oppositely signed optical rotations at the NaD line (589 nm), confirming their mirror-image structure. However, the signs of OR are not suitable for the determination of absolute chiral structures. Though the more reliable wavelength-dependent OR through an absorbing region, called optical rotary dispersion (ORD), was measuring chiral differences associated with the whole absorption region of metal complexes, it is now common and more useful to measure circular dichroism, or absorption differences between right- and left-circularly polarized light, through an absorption band, which is a technique equivalent to ORD [3]. Cotton’s first observation (see Chapter 1 in this volume) of optical rotation measurements through an absorption band, along with interpretation in terms of differential Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

451

452

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

absorption of the circularly polarized light, was performed on solutions of L-tartrate complexes of chromium(III) and copper(II) [4]. This indicates relatively easier application of electronic circular dichroism (ECD) measurements and the merit of most transition metal complexes, which possess moderately intense d –d absorption bands in the more accessible visible and near-infrared region, in contrast to most organic compounds in which CD measurements are limited to the ultraviolet region. The availability of commercial instrumentation in the 1960s led to numerous investigations aimed at the determination of chiral inorganic stereochemistry through the measurement of ECD, as well as a better understanding of UV–vis spectra and electronic structure, leading to advances in the chiroptical theory for metal complexes [5–10]. On the basis of the fundamental information on the structure–spectra relation, recent prime interests are in application-oriented subjects such as asymmetric catalysis and chiral interactions between metal complexes and biomolecules such as proteins and DNA. As described below, however, there are often exceptions and complications in the development of reliable rules relating CD sign patterns to absolute configuration or conformation, with need of confirmation by NMR spectra or X-ray analysis of suitably grown single crystals and/or DFT theoretical calculations. In this chapter, the application of electronic circular dichroism (ECD) to inorganic stereochemistry will be reviewed and is limited to coordination compounds. ECD will mostly originate from the ligand-centered and charge-transfer transitions, as well as the metal-centered d –d or f –f transitions. Special attention will be given to the structure–spectra relation for a variety of chiral structures of coordination compounds, with the emphasis on current experimental and theoretical techniques.

13.2. ECD IN THE D–D TRANSITIONS OF TRIS- OR BIS-BIDENTATE TRANSITION METAL COMPLEXES The relationships between chiral structures and ECD for inert cobalt(III) and chromium(III) complexes with d 6 and d 3 configurations have been the most often studied, since stable chiral complexes can be synthesized, and the d –d bands have been very well characterized on the basis of the ligand-field theory. The standard CD criterion to determine the chiral structure for the tris- or bis-bidentate Co(III) complexes and the related metal complexes is the CD spectrum in the first d –d ligand-field transition of tris(ethylenediamine) cobalt(III) ion, (+)589 -[Co(en)3 ]3+ , with five-membered diamine rings. The reasons to choose this Co(III) complex as the CD criterion are twofold. For one reason, the absolute configuration of this complex was first determined to be the so-called D configuration in (+)589 -[Co(en)3 ]2 Cl6 ·NaCl·6H2 O crystal with anomalous X-ray diffraction techniques in 1955 by Saito et al. [11]. The original stereochemical descriptor D was later renamed to be , according to the helicity of the complex or the IUPAC skew-line convention [12], as shown in Figure 13.1. For the other reason, since the first d –d ligand field 1 T1 ←1 A1 transition of the Co(III) complexes is electric dipole-forbidden, but magnetic dipole-allowed, the CD intensity or dissymmetry factor g is relatively large compared with the second electric dipole-forbidden and magnetic dipole-forbidden d –d ligand field 1 T2 ←1 A1 transition. As shown in Figure 13.2, -(+)-[Co(en)3 ]3+ gives a major positive and then a minor negative CD component from the low-frequency side in solution for the electric dipole-forbidden and magnetic dipole-allowed d –d 1 T1 ←1 A1 transition around 21,000 cm−1 . These correspond to the trigonal split states, 1 E and 1 A2 , which originate from the 1 T1 state in octahedral

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

Figure 13.1. Absolute configuration of (+)589 -[Co(en)3 ]2 Cl6 ·NaCl·6H2 O.

Figure 13.2. Vis–UV and CD spectrum of -(+)-[Co(en)3 ]3+ in water (broken line) and the axial single-crystal CD spectrum (solid line) of -(+)-{[Co(en)3 ]Cl3 }2 ·NaCl·6H2 O. The assignments to the octahedral (1 T1g ) and trigonal (1 E and 1 A2 ) states are represented.

field as shown in Figure 13.2 [13]. Uniaxial single-crystal polarized CD measurements of (+)589 -[Co(en)3 ]2 Cl6 ·NaCl·6H2 O reveal that the CD sign for the 1 E component is positive and hence the major positive band observed in solution is due to the 1 E component (Figure 13.2). The crystal CD intensity is found to be an order of magnitude larger than that of the solution CD [14]. On the other hand, the apparent trigonal splitting estimated from the solution CD is 3000 cm−1 , in contrast to very small splitting of 0–70 cm−1 from the single-crystal measurements. These facts indicate that the observation of weaker CD intensity and the apparently large trigonal splitting in the solution CD results from a mutual cancellation between oppositely signed 1 E and 1 A2 CD components with similar intensities. The net CD sign is that of the 1 E component.

453

454

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The CD component rule may also be applied in situations involving ligand-to-metal charge-transfer (CT) transitions in the UV region. In this application, negative and positive CD peaks in the CT region around 40,000–50,000 cm−1 and 50,000–60,000 cm−1 , respectively, are associated with the  configuration, and positive and negative ones are assigned to a  configuration. The validity of the extension of this rule to the CT region is reasonable if one considers that the major 1 E(1 T1 ) transition “borrows” CD intensity from the lower frequency 1 E(CT) rather than from the higher frequency 1 A2 (CT) on the basis of the assigment to the 1 E(CT) from the single-crystal CD (Figure 13.2) [14]. One exceptional case showing a positive sign results from the large contribution from the conformation of the ptn (2,4-diaminopentane) ligand in the six-membered diamine chelate, -(+)546 -[Co(R,R-ptn)3 ]3+ [7]. An empirical rule for the absolute configurations of tris- and bis-bidentate chelate complexes of the cis-[CoX2 (en)2 ] type was proposed from the experimental results that a positive and negative sign of the major 1 E CD component (of trigonal parentage) indicated a  and  complex, respectively [15]. It is important to note that some exceptions for the diamine complexes were found to be the strained five-membered diamine chelate complex (δδδ)-[Co(R,R-cptn)3 ]3+ (cptn = trans1,2-cyclopentanediamine), the six-membered diamine chelate complex, -[Co(tn)3 ]3+ (tn = trimethylendiamine) and -[Co(R,R-ptn)3 ]3+ (R,R-ptn = 2R, 4R-pentanediamine), and the seven-membered diamine chelate complex -[Co(tmd)3 ]3+ (tmd = 1, 4diaminobutane), which display a major negative CD in the spin-allowed 1 T1 ←1 A1 transition. For the (δδδ)-[Co(R,R-cptn)3 ]3+ complex, the conformational contribution ε(δδδ) from the three RR-cptn ligands exceeds the configurational ε() contribution as found experimentally from the CD for the mixed RR- and SS -cptn complexes, so the apparent solution CD intensity for the lower frequency 1 E component is smaller than that for the 1 A2 one [16]. Thus, the empirical rule still holds with respect to the configurational CD. The results for the -[Co(tn)3 ]3+ complex are due to the oppositely (negative) signed 1 E CD component for the -configuration, probably because of the distortion from a regular octahedron: the chelate N–Co–N bite angle α(N–Co–N) > 90◦ or the radial (azimuthal) expansion (φ > 60◦ ), as compared to the cases for the -five-membered diamine chelate complexes with the smaller bite angle (α(N–Co–N) < 90◦ ) and azimuthal contraction (φ < 60◦ ) giving a positive major 1 E CD as theoretically predicted (vide infra) [17, 18]. However, these exceptions clearly are problematic if one is relying upon ECD measurements to determine absolute configuration. The same empirical rule has been shown to be applicable for tris- or bis-diamine Cr(III) and Ni(II) complexes; that is, the  configuration gives a positive major CD band in the electric dipole-forbidden and magnetic dipole-allowed d –d transition of 4 T2 ← 4A2 for Cr(III) and 3 T2 ← 3A2 for Ni(II) [19]. The solution CD spectra of (+)-[Cr(en)3 ]3+ gives only one positive component (Figure 13.3), which is assigned to one (4 E(4 T2 )) of the trigonal split states by the single-crystal polarized CD measurement [20, 21]. By this CD component rule, using the signs of the 4 E(4 T2 ) component, or the trigonal parentage for the lower symmetry complexes, the absolute configurations of many tris- or bis-bidentate complexes were determined. For Cr(III) complexes, it has been found that characteristic weak but sharp CD patterns in the spin-forbidden transitions within the t2g subshell of Cr(III) complexes yield important information on the CD behavior in the spin-allowed as well as spin-forbidden transitions and are correlated with the absolute configuration. In the

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

455

A(+)589−[Cr(en)3]3+ 2

−0.5 AB log e

log e 2 Δe

−1

1

1

× 100

CD 0 (2A,E) Eb (2A,Ea) 2 2 E A2 2E

−1

(4A1) 4E

Eg 2T1g

15

16

Figure 13.3. CD spectra in the spin-forbidden and

4

2

T2g

20 σ (103 cm−1)

spin-allowed transitions of -(+)-[Cr(en)3 ]3+ in

25

water. The assignments to the doublet and quartet states are given in terms of the single-group and/or double-group representations.

room-temperature solution CD spectra in the spin-forbidden 2 E, 2 Tl ← 4A2 d –d transitions of -(+)-[Cr(en)3 ]3+ , three sharp peaks, (+), (−) and (+), are observed, as shown in Figure 13.3. The CD sign of the lowest-frequency 2 E component is the same as that of the major 4 E CD component [22–24]. This empirical correlation also holds for the circularly polarized luminescence (CPL) of (+)-[Cr(en)3 ]3+ [25] and the chiral ––[EuIII CrIII (L2)3 ]6+ complex (see Figure 13.10) [26] in the 2 E– 4 A2 transitions, as well as the CD of a number of mixed-ligand Cr(III) complexes of tris-chelate type [22–24]. This is elucidated on the basis of the theoretical approaches for the rotational strengths between the spin-forbidden and the spin-allowed transitions in terms of the intensity borrowing mechanism through the spin–orbit coupling between the quartet-doublet states [23, 24]. That is, the rotational strength of the 2 E(2 Eg ) component is predicted to be proportional to the net rotational strength R(4 T2 ) = 2 R(4 E) + R(4 A1 ) : R(2 E(2 Eg )) = 32k [R(4 E) + R(4 A1 )](k = ζ  /18(E (4 T2g ) − E (2 ), where ζ  is the spin–orbit coupling constant and E (4 T2g ) − E (2 ) is the energy interval between 4 T2g and 2 . The CD sign of the 2 E(2 Eg ) component should be the same as that of the 4 E one with |R(4 A1 )| < |R(4 E)|. The other sharp CD peaks were assigned to the 2 A2 (2 T1g ) and 2 E(2 T1g ) from the lower-frequency side (Figure 13.3). The CD signs of the remaining higher-frequency 2 A2 (2 T1g ) and 2 E(2 T1g ) states are negative and positive as predicted in view of R(2 A2 (2 T1g )) = 2k [R(4 E) + 4R(4 A1 )] and R(2 E(2 T1g ))) = 24k [5R(4 E) + 2R(4 A1 )] for the negative 4 A1 and positive 4 E CD components. Some exceptions to the 2 E CD sign rule are observed for (+)546 -[Cr(acac or acaX)2 (en)]+ (acac = acetylacetonate; acaX = 3-halogenoacetylacetonate) and (−)D -[Cr(bgH)3 ]3+ (bgH = biguanide(NH{C(NH2 )NH}2 )) complexes [23]. In these cases, since the 2 E component is strongly split into the 2A(2 Eg ) and E (2 Eg ) levels, where denotes the double-group irreducible representation, owing to the large trigonal

456

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

splitting (K ) of the 4 T2 state, the two components on the lower-frequency side are assigned to the 2A(2 Eg ) and E (2 Eg ), of which the observed signs, (−) and (+), for the -isomer agree with the predicted signs (R(2A(2 Eg )) = 4k [3R(4 E) + 6R(4 A1 )] and (R(E (2 Eg )) = 4k [5R(4 E) + 2R(4 A1 )]) with R(4 E) > 0 and R(4 A1 ) < 0 and the energy ordering between the 2A(2 Eg ) and E (2 Eg ) levels.

13.2.1. Theoretical studies on the structure-spectra relationships There have been a number of qualitative or quantitative attempts to develop a theoretical basis for CD component rules for -[Co(en)3 ]3+ and related complexes by using the crystal-field theory with metal d –p orbital mixing [17] and the molecular orbital models (ligand-field theory) with mixing between the metal d and ligand orbitals [14, 27–31]. These efforts have primarily been concerned with developing possible chiral spectrastructure relationships or so-called “sector rules” with the emphasis on the source of d –d electric dipole-forbidden transition intensity. Mason and Seal [27] have calculated more rigorously the rotational strengths in the d –d transitions for various kinds of diamine Co(III) complexes by using the so-called dynamic-coupling ligand-polarization model where the electric dipole transition moment results from the allowed intraligand transitions, which is complementary to the crystal-field theory. This afforded quantitatively the net rotational strengths, R(1 T1 ), as well as the rotational strengths R(1 E) and R(1 A2 ) for the 1 E and 1 A2 components. These calculations, based on X-ray crystallographic structural data, include the polarizabilities of each XHn (X = C and N) group in the amine ligands together with the allowed higher-order hexadecapole moments in the d –d transitions of the cobalt ion. The calculated signs and magnitudes for the rotational strengths are in fairly good agreement with the observed ones, but still not consistent with the net rotational strength R(1 T1 ) for the -[Co(tmd)3 ]3+ and (λλλ)-[Co(R, R-ptn)3 ]3+ complexes. Later, two ab initio (see Chapter 22 in Volume 1) calculations of the CD of [Co(en)3 ]3+ have been reported to reproduce the CD pattern. Both theoretical studies suggested that the electric dipole transition moments stem from N–H and N–C σ –σ * intra-ligand transitions, as in the dynamic coupling model [8, 32, 33]. This result may help explain the violation of the empirical CD rule for [Co(tn)3 ]3+ . Very recently, time-dependent density functional theory (TD-DFT) calculations have been applied to bis-diamine as well as tris-diamine or tris(π -conjugated bidentate chelate) Co(III) or Rh(III) complexes in the entire experimental spectral region, including the charge-transfer transitions, by Autschbach and Ziegler [34, 35]. This ab initio theoretical method reproduces the experimental CD spectra not only in the d –d transitions, but also in the charge-transfer region. For the Rh(III) complexes, there is a good fit in energy and signed magnitude between the observed and calculated CD, but for the Co(III) complexes, energy shifts in the CT transitions are needed to fit the observed spectra. The agreement between the experimental and theoretical CT energy is improved by a theoretical estimation of the solvent effect for +3 charged Co(III) complexes. The TD-DFT calculations of the hypothetical -[M(NH3 )6 ]3+ show that the CD is dependent on two geometrical deformations from a regular octahedron. One is the azimuthal distortion that allows the even d (eg ) orbital to mix with the odd σ ligand orbitals, resulting in nonzero rotational strengths in the magnetic dipole-allowed d –d transitions. The model calculations predict a positive 1 E CD for the azimuthal contraction with φ < 60◦ and a negative 1 E CD for the azimuthal expansion with φ > 60◦ for the -configuration (vide supra). The other is the polar distortion, giving rise to trigonal splitting or governing the signs of the trigonal splitting parameter K = 2/3{E (E) − E (A)}. It has been demonstrated in these calculations

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

that K > 0 for the polar elongation (s/h < 1.22) and K < 0 for the polar compression (s/h > 1.22). (h denotes the distance between opposite faces, and s is the length of the triangular side in the octahedron.) This simple geometrical consideration accounts for the exceptional case of the major CD component rule for [Co(tn)3 ]3+ . The TD-DFT calculations are also performed for the corresponding paramagnetic Cr(III) complexes with an open-shell ground-state configuration for the tris(en) and the tris(acetylacetonato) or tris(oxalato) complexes with π -conjugated bidentate chelates [36]. It appears that the description of the relative contributions of the d –d and charge-transfer transitions to the CD spectrum is also supported by the DFT calculations. This DFT analysis provides a new reliable tool in developing an understanding of the relationship between CD spectra and absolute configuration of metal complexes. Theoretical simulations of transition metal complexes is discussed in more detail in Chapter 22 in volume 1.

13.3. POLYNUCLEAR COMPLEXES WITH CONFIGURATIONAL CHIRALITY 13.3.1. Dinuclear Complexes

e (mol−1dm3cm−1)

The dihydroxy-bridged dinuclear complex [Cr2 (μ-OH)2 (S , S -chxn)4 ]4+ (S , S chxn = (1S , 2S )-1,2-trans-cyclohexanediamine) is stereospecifically formed to take – absolute configuration [37]. This assignment is in agreement with that based on the major positive CD signs in the first d –d transition. As shown in Figure 13.4, a negative CD peak with a sharp half-band width (ν1/2 = 740 cm−1 ) near 285 nm or 35,100 cm−1 is observed, corresponding to the weak absorption shoulder near 285 nm with ε = 70 M−1 cm−1 . Since the half-band width is small and the transition energy (35,100 cm−1 ) is nearly equal to the sum of 15,000 cm−1 and 20,000 cm−1 , respectively, of the 2 2 E, T1 -4 A2 and 2 T2 −4 A2 transition, this CD peak is due to the double excitation from the lowest singlet (S = 0) level in the ground antiferromagnetic spin-coupled 4 A2 , 4 A2

200

AB

× 15

100 0

×2 N 300

400

500

600

Δe (mol−1dm3cm−1)

N +2

CD

N ×4

+1 300

N H O

700

× 25

400

M

M

O H N Λ

N N

N Λ

0 −1

× 10

500 λ (nm)

600

700

−2

Figure 13.4. UV–vis (upper) and CD spectra (lower) of -[Cr(OH)(S,S-chxn)2 (H2 O)]2+ (broken line) and -[Cr2 (OH)2 (S,S-chxn)4 ]4+ (solid line) in water. The dihydroxo-bridged dinuclear structure is shown on the right-hand side.

457

458

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

state to the excited 2 E, 2 T2 or 2 T1 , 2 T2 singlet level. Thereby, the transitions become formally singlet–singlet spin-allowed, otherwise quartet–doublet spin-forbidden, by the exchange-coupling mechanism for the antiferromagntic interaction (2J = −16 cm−1 ) [37]. This CD peak is a kind of dimer band that intensifies with lowering temperature, giving a probe for dinuclear structures. Another interesting example of binuclear octahedral-based chirality is the chiral polyoxometalate (+)589 -[Co2 Mo10 O34 (OH)4 ]6− , where the CD measurement has confirmed the same D2 symmetry [38] as the binuclear complexes discussed above, revealed by X-ray analysis [39]. (+)589 -[Co2 Mo10 O34 (OH)4 ]6− is used as a dopant to form polypyrrole films in conducting polymers with the magnetotransport properties in terms of a magnetochiral anisotropy effect [40]. Though the temperature-dependent conductivity of the chiral polymer is different from that of the racemic polymer, the CD of the chiral polymer shows that macroasymmetry is not induced in the chiral polyoxometalate dopant.

13.3.2. Tetranuclear Complexes of Hexol Type Shimura et al. [41–44] succeeded in completely resolving the diastereomers (diamine = en, of tetranuclear complexes [Co{(μ-OH)2 Co(diamine)2 }3 ]6+ meso-R, S -butanediamine, R-propylenediamine(R-pn), and (1R, 2R)-1,2-transcyclohexanediamine(R, R-chxn)), which are analogous to the well-known chiral tetranuclear hexol complex [Co{(μ-OH)2 Co(NH3 )4 }3 ]6+ with  and  enantiomers [2]. The structures for these diamine complexes consist of the eight diastereomers ()/(), ()/(), ()/(), and ()/() as shown in Figure 13.5. The absolute configurations of the metal-centered configurations around the Co(O–O)3 moiety were determined on the basis of the CD component rule around 16,500 cm−1 . From the CD signs for the cis-[Co(diamine)2 (H2 O)2 ]3+ generated by acid decomposition of the diastereomers, the absolute configurations of the peripheral Co(O–O)(en)2 moiety could be assigned. The CD spectra of [Co{(μ-OH)2 Co(en)2 }3 ]6+ diastereomers were not found to follow simple additivity for the two main CD contributions due to the central Co(O–O)3 and the peripheral Co(O–O)(N–N)2 moieties. On the other hand, from the CD behavior of the corresponding chiral R-pn or R, R-chxn complexes, it was found that the vicinal contribution of the R-pn and R, R-chxn complexes is due to the asymmetric carbon(s) with R-configuration but not to the λ-ring conformation, in view of the CD intensity being 2ε (R-pn) = ε (R, R-chxn). Such a dominant vicinal contribution of asymmetric carbons is the first case for octahedral six-coordinate complexes, unlike the mononuclear complexes with chiral organic ligands. CD spectra of another type of chiral hexol structures have been measured in Anderson-type heteropoly acids, telluratocobaltate(III) [Co4 Te3 O18 (en)3 ]6− , periodatocobaltate(III) [Co4 I3 O18 (en)3 ]4− , and diastereomers of [Co4 I3 O18 (l-ala)3 ]3− [45]. On the basis of the CD signs in the d –d transitions, the absolute configuration of (+) [Co4 Te3 O18 (en)3 ]6− and (−)-[Co4 I3 O18 (en)3 ]3− are ( )( )3 and ( )( )3 , respectively, where ,  and  concern the absolute configurations around the central Co(MO6 )3 , Co(CoO4 en)3 , and the peripheral Co(en)(MO6 )2 , respectively. The CD intensities of the heteropoly acids are weaker than those of the hexol-type Co(III) complexes. This may be due to the insertion of three MO6 octahedra between the peripheral Co(III) octahedra of the hexol-like Co4 moiety [44]. Recently, Sato and co-workers [46] succeeded in preparing and resolving into eight enantiomers of another hexol type of so-called

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

(a)

459

(b)

Figure 13.5. Four diastereomers of (c)

(d)

[Co{(OH)2 Co(en)2 }3 ]6+ : (a) (), (b) (), (c) (), (d) ().

star-burst tetranuclear tetraacetylethanate(taet)-bridged Ru(III) complexes with acetylacetonate, -[{-Ru(III)(acac)2 (taet)}3 Ru(III)]. Additivity between the CD contributions from the configurational chirality of the central and peripheral moieties is found in contrast to the nonadditivity for [Co{(μ-OH)2 Co(en)2 }3 ]6+ diastereomers mentioned above. This CD behavior is used as a diagnostic tool to discriminate among the diastereomeric ()/(), ()/(), ()/(), ()/() enantiomers.

13.4. ECD IN THE 4f –4f TRANSITIONS Though a number of experimental studies on ECD in the 4f –4f transitions attempted to find the CD-sensitive bands and/or the relationship between the CD signs and the chirality of ligands [47–51], the structure–spectra relationship including the absolute configurations around lanthanide ions could not be examined in detail in contrast to transition-metal complexes. This is because lanthanide complexes in solution are too labile to fix the chiral structures and because the 4f –4f transition intensities are too weak for reliable ECD measurements, which require very high concentrations and long cell pathlengths, and are less understood in theory [52–54]. A nonracemic ground state can be generated by the so-called Pfeiffer effect on addition of a chiral compound [55]. However, this method cannot be used to provide the structure–spectra relationship, since the absolute configuration cannot be definitely identified. On the other hand, there are several examples of configurationally chiral complexes that are stereospecifically formed with chiral ligands, but recently a limited number of CD data for emissive lanthanide(III) complexes, such as Eu(III), Tb(III), or Yb(III) complexes, makes it feasible to provide

460

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the structural probe in connection to circularly polarized luminescence (CPL) [10, 56, 57] (see Chapter 3 in volume 1). Kaizaki and co-workers [58, 59] reported ECD for two series of lanthanide complexes with configurational chirality. Chiral (-)-[(acac)2 Cr(ox)Ln(HBpz3 )2 ] was prepared from a mixture by a chiral interaction between -[Cr(acac)2 (ox)]− and racemic [Ln(HBpz3 )2 ]+ (HBpz− 3 = hydrotris(pyrazol-1-yl)borate), resulting in stereospecific complex formation. The retention of the  absolute configuration around SAPR-8-[Ln(ox)(HBpz3 )2 ]+ (SAPR-8 = square antiprism eight-coordinate) in the Cr(ox)Ln assembly upon combination with inert -[(acac)2 Cr(ox)]− was confirmed by X-ray analysis as well as by the large intensity of 4f –4f CD of the Ln(III) moiety in solution. The CD in the 4f –4f transitions of these structurally well-defined complexes enables us to propose a criterion to determine the absolute configuration, though they are limited to the NIR region in view of overlapping with the d –d bands of the Cr(III) moiety in the UV–vis region. By comparing CD patterns of the (-)-Cr(ox)Ln complexes, an empirical criterion for the relation between the 4f –4f CD signs and the  absolute configurations around the Ln(HBpz3 )2 (ox) moiety was given: a positive sign of the major CD band in the 4 I9/2 → 4F3/2 (Nd), 6 H5/2 → 6F11/2 (Sm), 6 H15/2 → 6F7/2 (Dy), 5 I → 5H (Ho), 3 H →3 H (Tm) transitions for which Richardson’s classifications [54] 8 5 6 4 are type 5 and RII for the rotational strength () and DIII for the dissymmetry factor(g), with S = 0, L ≥ 0, 2 ≤ J ≤ 6(J = 0 = J  ). The 4 I15/2 → 4I11/2 transition of Er is an exception for this relation. On the other hand, much invaluable information on the spectra–structure relation in the CD of the --Cr(ox)Ln complexes could be provided by a series of ECD spectra in the vis–NIR region for the Cs[Ln((+)-hfbc)4 ] (Cs-Ln) complexes, where (+)-hfbc is heptafluorobutyryl-(+)-camphorate, taking a -SAPR-8-(llll ) configuration (l between sites in different squares) with four helically bladed propellers, as shown on the right side of Figure 13.6 [60, 61]. The configurational chirality of Cs[Ln((+)-hfbc)4 ] is retained by the intramolecular interaction between the fluorocarbon and cesium ion, as revealed by the exciton CD spectra (vide infra). The Cs–Yb complex gives a strong positive CD peak in the 2 F7/2 → 2F5/2 transition with g = +0.318, much larger than that (+0.06) of the Cr(ox)Yb complex and the reported values (−0.14) for -[Yb(DOTMA)][56] (see Chapter 11 in volume 1). The CD peak in the 2 F7/2 → 2F5/2 transition gives a suitable criterion for the absolute configuration of the Yb complexes: a positive sign for the  configuration, because this transition is CD-sensitive in view of the transition type 1 with S = 0, L = 0, J = 0, 1(J = 0 = J  ) and the transition properties belong to the RI and DII class for which the CD and g values are larger, respectively, than those of RII and DIII [54]. For the 6H 6 5/2 → H7/2 transition for the Cs–Sm complex with the same type 1 and the same DII class as that for the Yb complexes, however, this criterion could not be adopted. That is, two large negative CD peaks are observed around 1100 cm−1 with g = −0.01 to −0.005 for the  absolute configuration of the Cs–Sm complex, not in accordance with that for the Yb complexes [62]. The aforementioned criterion for the Cr(ox)Ln complexes can be applied to the other Cs–Ln complexes. There are two exceptions, for the Cs–Sm complex (6 H5/2 → 6 F11/2 ) and the Cs–Er complex(4 I15/2 → 4I11/2 ), which give a negative sign for the  configuration. Therefore, this criterion is not suitable for the Cs–Ln complexes. So far, Richardson’s classification or selection rules could not provide a criterion of the absolute configuration of the lanthanide(III) complexes.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

0.05 0

Δε

−0.05 Nd

−0.15

ν (10 −0.25 14.00

3 cm−1)

16.00

18.00

20.00

0.3 MI

Δε

0 −0.3 −0.6 −0.9 20.00

Ln Ho Δ-SAPR-8-C4(llll)-M[Ln(+)-(hfbc)4] with an encapsulated alkali metal ion

ν (103 cm−1) 21.00

22.00

23.00

24.00

25.00

0.04

Δε

0 −0.04 −0.08 −0.12

ν (103 cm−1)

Er

−0.16 12.00 14.00 16.00 18.00 20.00 22.00 24.00

Figure 13.6. CD spectra in the hypersensitive 4f –4f transitions of -Cs[Ln((+)-hfbc)4 ] in CHCl3 (left) and the proposed structure in solution (right). (See insert for color representation of the figure.)

Close comparison among the ECD in the UV–vis to NIR region of the -Cs–Ln complexes reveals that the CD signs in the hypersensitive transitions with environmental sensitivity are correlated to the absolute configurations of the Cs–Ln complexes [62], though theoretically the hypersensitive transitions belonging to RII or DIII class of type 5 for the S = 0 or type 11 for the S = 0 with L ≥ 0, 2 ≤ J ≤ 6(J = 0 = J  ) may not be the most favorable chiroptical probe as claimed by Richardson [54]. That is, as shown in Figure 13.6, the CD components are negative in the 4 I9/2 → 4G5/2 (Cs–Nd), the 5 I8 → 5G6 (Cs–Ho), and the 4 I15/2 → 2H11/2 (Cs–Er) transitions. For other hypersensitive transitions, the -Cs–Eu complex gives a negative 5 D0 – 7 F2 CPL peak at 16,300 cm−1 [63] as well as a negative 5 D2 – 7 F0 CD shoulder near 21,700 cm−1 on the intense negative peak at 24,600 cm−1 which may be assigned to the singlet–triplet π –π * transition of (+)-hfbc in view of the intensity and position as well as the band width [62]. For -Na3 [Eu(ODA)3 ]·2NaClO4 ·6H2 O(ODA = oxydiacetate) with trigonal D3 symmetry, a negative 5 D2 – 7 F0 CD peak was clearly observed at 21,500 cm−1 [64], and the CPL (see Chapter 3 in Volume 1) in the 5 D0 – 7 F2 emission showed a negative major component at 16,200 cm−1 [57, 65]. --[EuIII CrIII (L2)3 ]6+ with distorted trigonal D3 symmetry around Eu(III) gave a negative 5 D0 – 7 F2 CPL peak at 21,500 cm−1 [26].

461

462

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

This relation of the CD and CPL signs with the absolute configuration is in contrast to the case of the CD-sensitive 5 D0 – 7 F1 magnetic dipole-allowed transition of which the CPL and/or CD sign for the -Cs–Eu and --[EuIII CrIII (L2)3 ] is opposite to that for -Na3 [Eu(ODA)3 ]·2NaClO4 ·6H2 O. It is noted that these CD or CPL peaks in the hypersensitive transitions are composed of dominant negative component(s) and are not disturbed by neighboring positive peak(s) and are thus without mutual cancellation. Therefore, these hypersensitive CD or CPL components could be an appropriate criterion of the chiral structure–spectra relation for SAPR-8(llll ) or trigonal D3 lanthanide(III) complexes, even though there are some differences in ligand electronic structure and coordination polyhedra among these Eu(III) complexes [57].

13.5. EXCITON ECD IN THE INTRALIGAND TRANSITIONS 13.5.1. Tris- and Bis-Chelate Complexes For tris- and bis-chelated octahedral six-coordinate complexes having conjugated aromatic bidentate ligands, a large CD couplet can be observed in the UV region, as predicted (Figure 13.7) [66] in contrast to the corresponding mono-chelated complexes that give a single CD component with a positive sign for the -configuration and vice versa. These CD bands are considered to be due to exciton splitting (see Chapter 4 in this volume) and the zero-order rotational strengths are induced from the helical charge displacement by the coulombic coupling of the two or three long-axis π –π * transitions λ phen

bpy

[M(phen)3]

[M(phen)2X2]

[M(bpy)(phen)X2]

[M(phen)2(bpy)]

[M(bpy) 2(phen)]

Figure 13.7. Predicted exciton CD patterns of bpy and phen complexes.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

of the identical ligands. For tris-chelate complexes with D3 symmetry, there are E and A2 couplings for which the rotational strengths are positive and negative, respectively, going from lower to higher frequency for the  configuration, on the basis of the simple exciton approximation. This pattern of oppositely signed CD bands is called a couplet and is designated as a positive couplet if the low-frequency band is positive, as in this case, and as a negative couplet if the low-frequency band is negative. The situation is the same for bis-chelated complexes as for the tris complexes. On the basis of this exciton model, the absolute configurations can be determined for tris- and bis-(chelated) complexes with 2,2 -bipyridine (bpy), 1,10-phenanthroline (phen), acetylacetonate (acac), 1,2-benzenediolate or catecholate (cat), and related aromatic ligands. The exciton CD for an octahedral six-coordinate metal complex with a closed-shell configuration was observed for -(+)-[Si(acac)3 ]+ by Larsen et al. [67], -(+)-[As(cat)3 ]+ by Mason and Mason [68], and -(+)-[Si(bpy or phen)3 ]+ by Yoshikawa et al. [69] and were shown to give a positive couplet in the pure π –π * transition of the acac, cat, or phen ligand. This assignment was confirmed by the X-ray analysis of -(+)-[As(cat)3 ]− by Ito et al. [70]. Even for the transition-metal complexes with an open-shell configuration, such as [M(bpy)3 ]n+ and [M(phen)3 ]n+ (M = Fe2+,3+ ; Ru2+,3+ ; Os2+,3+ ), where there is mixing with the d –d and CT transitions, exciton couplets have been observed for the intraligand π –π * transitions. In particular, most Co(III) and Cr(III) complexes exhibit agreement between the absolute configurations based on the d –d CD and the exciton CD and/or the X-ray structure determinations [71]. For -[Co(bpy)3 ]3+ , -[Ni(bpy)3 ]2+ , and [Ni(phen)3 ]2+ , however, the absolute configuration based on the exciton CD analysis is in agreement with the X-ray crystal structure, but the empirical d –d CD gives the incorrect result [71, 72]. The reverse situation is encountered for the tris-biguanide Co(III) complex. -(−)D -[CoIII (bgH)3 ]3+ (bgH = NH{C(NH2 )NH}2 ), for which the major CD rule gives the same absolute configuration as the X-ray analysis, exhibits a UV CD couplet opposite to the exciton CD prediction, unlike the corresponding -(−)D -[CrIII (bgH)3 ]3+ . This may be due to uncertain assignment of the intraligand transition arising from large mixing of the d –π orbital with the bgH orbital in the Co(III) complex [73, 74]. Recent time-dependent density functional theory (TD DFT) studies showed that the exciton CD for [M(phen)3 ]2+ (M = Fe, Ru, Os) complexes [75, 76] and related complexes [77] are reproduced in good agreement with the experimental pattern. For nonidentical mixed-ligand complexes, such as -(−)589 -[Cr(ox)(bpy)(phen)]+ , the CD pattern would be two positive peaks if there were no coupling between the nonidentical π –π * transitions of bpy and phen, since a positive single CD component in the intraligand π –π* transition is observed for mono-phen or bpy complexes such as -[Co(en)2 (phen)]3+ [78] and --[Cr2 (L-tart)2 H(bpy or phen)2 ]− [79]. However, this mixed-ligand complex gives a positive couplet in the region of the bpy and phen transitions [80] as predicted by the exciton theory as in Figure 13.7. This type of nondegenerate exciton CD may also be applied with success to more complicated nonidentical couplings in [M(bpy)2 (phen)]n+ or [M(bpy)(phen)2 ]n+ . These complexes exhibit the expected characteristic exciton CD patterns with three components: (+),(+),(−) from lower to higher frequency for the -isomers, as theoretically predicted (Figure 13.7) [71, 81]. The chiral tetrahedral (T-4) and distorted square-planar four-coordinate (SP-4) metal complexes which are stereospecifically formed with bis-(−)-α-pinene-bpy-type ligands give exciton CD spectra [82]. The absolute configurations of the Ag(I), Pd(II), and Zn(II) complexes predicted by the exciton CD model nicely agree with those determined by the X-ray analysis.

463

464

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

13.5.2. (β-Diketonato) Lanthanide(III) Complex Several achiral tris-β-diketonato lanthanide(III) complexes have been known as a chiroptical probe for specific recognition of chiral amino acids, amino alcohols, and diols since the pioneering work of Nakanishi and Dillon [83]. The CD method is based on the CD couplet near 300 nm which is useful to determine the chiral configuration of organic compounds functioning as chiral bidentate ligands [84, 85]. The signs of the CD couplets depend on the conformation of the bidentate chelates, with the bulky groups in the equatorial positions. That is, R-bidentate ligands with the anticlockwise (λ) conformation give a positive couplet and vice versa. Assuming the helical bladed arrangement around the central tetrakis-chelated Ln ion with the R-bidentate ligand, it is supposed to take a  absolute configuration on the basis of the exciton theory. However, it is difficult to assume the exciton CD bands in the β-diketonate intraligand transition region, since SAPR-8(ssss) or -(ssll ) configurations are more stable than the SAPR-8-(llll ) (Figure 13.8), in view of more favorable chelation at sites within the same square planes due to steric hindrance, as found for most tetrakis(β-diketonato) lanthanide(III) complexes. The exciton CD spectra of the β-diketonate Ln complexes are actually observed for tetrakis-3-heptafluorobutyryl-(+)-camphorate Ln(III) complexes with an encapsulated alkali metal ion, M[Ln((+)-hfbc)4 ]is (Figure 13.6) [60, 61]. Since complexes in this group are stereospecifically formed with chiral -SAPR-(C4 (llll )) configurations with the aid of CF· · ·M+ intramolecular interaction, as shown in Figure 13.9, a negative CD couplet in the π –π * transition of the β-diketonate chromophore is safely assigned to exciton CD, leading to the -SAPR-8 absolute configuration (Figure 13.6) [60, 61]. It is noted that the exciton CD intensities (ε) depend on the alkali metal ion size, Cs+ > Rb+ > K+ > Na+ , reflecting the variation of twist angles around the C4 axis in the helically four-bladed C4 (llll ) chiral configuration. The absolute configuration for the Cs–Ln(III) complexes is supported by other chiroptical methods: CPL shows the largest g value for the Cs–Eu [63], VCD [86] and 4f –4f CD (vide supra).

Figure 13.8. Possible configurations of a square antiprism (SAPR)-8 complex. (a) llll and (b) ssss means chelation in the sites within different and the same

(a)

square, respectively.

(b)

200

Δε

100 0 −100 −200 250

Figure 13.9. Exciton CD spectra of 300 λ (nm)

350

M[La((+)-hfbc)4 ] in CHCl3 . M: Cs (red), Rb (green), K (blue), Na (black). (See insert for color representation of the figure.)

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

internuclear (net effect) intranuclear (Ln end) intranuclear (Cr end)

Δε (M−1cm−1)

50

CrIII

LnIII 0

N

N

−50

N

N

N

N

L2 275

300

375 325 350 Wavelength/nm

400

425

O N

Figure 13.10. Right: Structure of the ligand L2(below) and -[LnIII CrIII (L2)3 ]6+ (above). Left: Schematic vertical lines summering the dominant coupling effects in the CD spectra of -[LnIII CrIII (L2)3 ]6+ . The black line corresponds to the CD spectrum of -[GdCr(III)(L2)3 ]6+ in CH3 CN. (See insert for color representation of the figure.)

In some polynuclear complexes, including helicates, exciton CD is observed in the intraligand π –π * transition for each metal center, but for many of these species the CD depends on the bridging units connecting the individual chromophores, leading to anomalous CD patterns [87]. Recently, the CD spectra observed in the π –π * transitions of the dichlorido-bridged dinuclear complexes, -[(L1)2 Co(μ-Cl)2 Co(L1)2 ]2+ (L1 = chiral tetradentate Schiff base chelate with diimine chromophores) [87] and -[LnCr(L2)3 ]6+ (Figure 13.10) [26], show an exciton CD pattern opposite to that expected for the absolute configuration from the X-ray analysis. This has been elucidated on the basis of semiempirical (ZINDO) calculations by considering the internuclear exciton coupling [87, 88]. The other dinuclear complexes -[(bpy)2 Ru(bpm)Ru(bpy)2 ]2+ (bpm = 2, 2 bipyrimidine) and the related trinuclear complexes show weak exciton CD bands with the correct signs. Such anomalous CD intensity as compared with that of the corresponding analogous mononuclear complexes can be reproduced with the oppositely signed internuclear exciton CD contributions by the ZINDO calculations, as schematically depicted in Figure 13.10. This theoretical approach provides a potential tool to determine the absolute configuration on the basis of the exciton CD bands for polynuclear complexes, especially playing a crucial role for systems exhibiting the incorrect CD signs, opposite to those of the mononuclear complexes, leading to the correct determination of the absolute configurations. In summary, even though the implicit approximations used in the application of exciton CD to determine absolute configuration ignore important complications such as the exciton splitting order, d -orbital mixing, and vibronic coupling, it has been a remarkably successful method for determination of the absolute configuration with careful consideration of internuclear exciton contributions for polynuclear complexes.

13.6. ATROPISOMERISM AND ECD Atropisomers with rotamer conformations of coordinated monodentate ligands were found for the diastereomeric trans-[CoCl2 py4 ](hydrogen L-dibenzoyltartrate) compound in a

465

466

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

solid powder by CD [89] and later confirmed by X-ray analysis [90]. Other examples are found for the conglomerates in space group P 21 21 21 of isomorphous nitrito-κ-N K[Co(NO2 )4 (NH3 )2 ] and NH4 [Co(NO2 )4 (NH3 )2 ] [91], as well as mer-[Co(NO2 )3 (NH3 )3 ] [92] by neutron diffraction studies, and the CD spectra were measured later [93]. For these pyridine and nitrito-κ-N complexes, the racemization of the atropisomers of monodentate lignds are too fast in solution to permit measurement of the solution CD spectra.

13.6.1. Nitrito-κ-O Complexes Such lability in the pyridine and nitrito-κ-N complexes could be reduced by chiral peripheral ligands near the monodentate ligands. This was observed for a dianionotetramine Co(III) complex with a chiral tetradentate ligand. The CD spectra in the first ligand field d –d band region of trans-R, R-[Co(N3 )2 (3,2,3-tet)]+ (3, 2, 3-tet(NH2 (CH2 )3 NH(CH2 )2 NH(CH2 )3 NH2 ) = chiral tetradentate linear tetramine ligand with δ conformation of the central ethylene backbone) gave enantiomeric patterns in H2 O and DMSO [94]. This suggests chirality due to rotational conformations of axial monodentate aniono ligands. However, such rotational or atropisomeric chirality cannot be evaluated by ECD because the d –d CD is susceptibile to chiral contributions from the ring conformation or geometrical structures of complexes. On the other hand, CD in the intraligand transitions of monodentate ligands could provide chiral rotational information on the monodentate ligand itself. Monodentate nitrito-κ-O Cr(III) complexes provide examples. Kaizaki [95] observed unique vibronic CD spectra with about 1000 cm−1 progression due to the NO stretching vibrations in the intraligand n –π * transition of the nitrito-κ-O ligands from 24,000 to 30,000 cm−1 for trans-[Cr(ONO)2 (R-pn)2 ]+ and trans-[Cr(ONO)2 (R,R-ptn)2 ]+ , as shown in Figure 13.11. The considerable solvent-dependent vibronic CD intensity enhancement reveals stereoselective solvation to the amino equatorial NH protons in increasing order of the bulkiness as well as the donor number of the solvents. The intraand intermolecular interactions of nitrito ligands with diamine chelates and/or the solvent molecules determine the most probable location of the nitrito ligands at equilibrium near amino nitrogen atoms in their predominant chiral rotamer conformations, which are sensitive to the N–Cr–N diamine chelate angles and/or the crowding around the chelate or the amino protons, causing the CD sign inversion between the R-pn and R, R-ptn complexes and very weak vibronic CD for -(+)450 ε -cis-[Cr(ONO)2 (en)2 ]+ as shown in Figure 13.11.

13.6.2. Guanine Derivative–Platinum(II) Complexes Analogous rotational isomers or atropisomers to the above nitrito complexes are found in square-planar four-coordinate SP-4 platinum(II) complexes. CD spectra of a number of model guanine platinum(II) complexes have been studied in order to reveal the selective binding to target DNA of platinum anticancer drugs that form a critical lesion by interor intrastrand cross-linking with two adjacent guanine bases or an adenine and a guanine base, respectively. A simple model complex, [Pt(R, R-dach)(9-EtG)2 ] (R, R-dach or R, R-chxn = R, R-cyclohexanediamine; 9-EtG = 9-ethylguanine), showed two CD couplets centered at 280 nm and 230 nm in the intraligand π –π * transitions of the 9-EtG, and the corresponding S , S -dach complex gives the enantiomeric CD pattern [96]. This clearly indicates a transmission of chirality of the dach ligand to the coordinated cis-guanine

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

467

2 AB

0.8 1.5 0.6 log ε

Δε (mol−1 dm3 cm−1)

1

0.4 (+) 0.2

1

0 CD

0.2 (−) 0.4

Figure 13.11. UV–vis(AB) and CD spectra 18

20

22

24

26

28

30

σ (103 cm−1)

of (+)-cis-[Cr(ONO)2 (en)2 ]+ (— —), trans-[Cr(ONO)2 (R-pn)2 ]+ (— — —) and trans-[Cr(ONO)2 (R, R-ptn)2 ]+ (· · ·).

bases through the diamine protons. Though the CD inversion was assumed to result from a change of the tilting direction of the guanine bases relative to the coordination plane [96], Natile et al. [97] elucidated more convincingly the CD results by recent studies on more rigid diamine complexes with [Pt{(S , R, R, S or R, S , S , R)-Me2 dab}G2 ] (Me2 dab = N , N  -dimethyl-2,3-diaminobutane with four asymmetric centers at the N, C chelate ring atoms). A correlation is demonstrated between the CD signs and the two chiral conformers among three possible ones depicted in Figure 13.12: the major chiral head-to-tail (HT) forms, in which the guanine bases have their six-membered rings located on the opposite sides, and a head-to-head (HH) conformer where the guanine bases have their six-membered rings located on the same side [98]. The CD couplets are interpreted on the basis of the exciton coupling theory, which predicts the sign inversion from HT to HT or vice versa. As shown in Figure 13.13, the guanine derivative complexes [Pt(diamine)(3 -GMP)2 ] (diamine = (NH3 )2 , en or tn: 3 -GMP = guanosine3 -monophosphate) give -type CD with the HT atropisomer, whereas the CD patterns of the corresponding 5 -GMP (guanosine-5 -monophosphate) complexes are the reverse of those of the 3 -GMP complexes. The same CD behavior is observed for the R, R- and S , S -dach complexes. This indicates that the chirality of the D-ribose in the GPM plays a more important role than that of the diamine in forming the HT or HT conformation,

N

N

N

N

N

N

Pt

Pt

Pt

ΔHT

HH

ΔHT

Figure 13.12. Schematic representation of HH(head-to-head), HT (head-tail), and HT (headto-tail) atropisomers for cis-[PtA2 X2 ] complexes.

468

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

a

8

a

b c

Δε (M−1cm−1)

Δε (M−1cm−1)

5

0 −5

b

4

c 0 −4 −8

250

300

250

nm

300 nm

Figure 13.13. CD spectra of cis-[Pt(amine)2 (n -GMP)2 ] at pH 7. (Left) 3 -GMP. (Right) 5 -GMP. Amine = (a) (NH3 )2 , (b) tn(trimethylediamine, (c) en(ethylenediamine).

through an internucleotide hydrogen bond between a phosphate group and the N1H of the neighboring GMP. These intramolecular interactions are sensitive to the bite angles of amine ligands, in view of the weaker CD intensity for both the (en) complexes than that for the (NH3 )2 and (tn) complexes (Figure 13.13) [98]. Other than monodentate complexes, there have been a number of atropisomers of metal complexes with bidentate chelates. Examples are the optical resolution and CD spectra of dinuclear atropisomeric (3,4-diacetyl-2,5-hexanedionato)bis[(2,2 , 2 triaminotriethylamine)cobalt(III), [{Co(tren)}2 tae], and the related mononuclear 3-aryl2,4-pentanedionato complexes by Nakano et al. [99]. It is noted that the CD in the intraligand or charge-transfer transition region is comparable in intensity to the d –d CD for the dinuclear complex, whereas much larger CD intensities are observed in the d –d CD for the mononuclear complexes than for [Co(acac)(en)2 ]+ . This may depend on whether the major chirality originates from the relative configuration of the tren moiety, but not directly from the β-diketonate rings. Recently, McCormick and Wang [100] reported the atropisomers of a polypyridyl N , N -chelate induced by zinc(II) and a chiral carboxylate, Zn(R- or S -O2 CCH(Br)CHMe2 )2 , where the zinc(II) ion functions as a mediator to facilitate the recognition event between the atropisomeric ligand and the chiral Zn(II) carboxylate, according to CD measurements.

13.7. SUMMARY Application of ECD to inorganic stereochemistry is going to become even more essential as chiral-selective structural probes, as asymmetric catalysts, and as key components of biomolecular processes. It will become increasingly important for researchers to explore the chiral structure–spectra relations that will make it possible to understand the various mechanisms and to design new chiral substances. Since there are still some problems in establishing reliable rules relating CD signs to absolute configuration or conformation, the development of synthetic methods for chiral complexes, the progress of ECD techniques, and the advances in theoretical computation of chiroptical properties in recent years have certainly resulted in newer, more consistent ways to understand the inorganic stereochemistry in both flexible and rigid systems.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

REFERENCES 1. A. Werner, V. L. King, E. Scholze, B er. Dtsch. Chem. Ges. 1911, 44 , 1887–1898. 2. A. Werner, Ber. Dtsch. Chem. Ges. 1914, 47 , 3087–3094. 3. S. F. Mason, Molecular Optical Activity and the Chiral Discriminations, Cambridge University Press, Cambridge, 1982. 4. A. Cotton, Ann. Chim. Phys. 1896, 8 , 347–432. 5. J. Fujita, Y. Shimura, Optical rotary dispersion and circular dichroism, in Spectroscopy and Structures of Metal Chelate Compounds, K. Nakamoto and P. McCarthy, eds., John Wiley & Sons, New York, 1968, pp. 156–215. 6. R. D. Peacock, B. Stewart, Coord. Chem. Rev . 1982, 46 , 129–157. 7. R. Kuroda, Y. Saito, Circular dichroism of inorganic complexes: Interpretation and applications, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., John Wiley & Sons, New York, 2000, pp. 563–599. 8. R. D. Peacock, B. Stewart, Chiral molecules spectroscopy, in Comprehensive Coordination Chemistry II: From Biology to Nanotechnology, Vol. 2. Fundamentals: Physical Methods, Theoretical Analysis and Case Studies, A. B. P. Lever, ed., Elsevier Pergamon, Amsterdam, 2004, pp. 65–81. 9. H. Amouri, M. Gruselle, Chirality in Transition Metal Chemistry: Molecules, Supramolecular Assemblies and Materials. John Wiley & Sons, Chichester, UK, 2008. 10. J. P. Riehl, S. Kaizaki, Absolute chiral structures of inorganic compounds, in Physical Inorganic Chemistry, Vol. 1. Principles, Methods, and Models, A. Bakac, ed., John Wiley & Sons. Hoboken, NJ, 2010, pp. 143–197. 11. Y. Saito, K. Nakatsu, M. Shiro, H. Kuroya, Acta Crystallogr. 1955. 8 , 729–730. 12. N. G. Connelly, T. Damhus, R. M. Hartshorn, A. T. Hutton, eds., The Red Book— Nomenclature of Inorganic Chemistry: IUPAC Recommendations, 2005, The Royal Society of Chemistry, Cambridge, 2005. 13. A. J. McCaffery, S. F. Mason, Mol. Phys. 1963, 6 , 359–371. 14. S. F. Mason, R. H. Seal, Chem. Commun. 1975, 331–333. 15. A. J. McCaffery, S. F. Mason, and B. J. Norman, J. Chem. Soc. 1965, 5094–5107. 16. M. Kunimatsu, H. Kanno, M. Kojima, K. Kashiwabara, J. Fujita, Bull. Chem. Soc. Jpn. 1980. 53 , 1571–1576. 17. T. S. Piper, A. G. Karipides, Mol. Phys. 1962. 5 , 475–483. 18. R. R. Judkins, D. J. Royer, Inorg. Chem. 1974, 13 , 945–950. 19. D. F. Harding, S. F. Mason, B. J. Peart, J. Chem. Soc. Dalton Trans. 1973, 955–960. 20. S. F. Mason, Fundamental Aspects and Recent Developments in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli and P. Salvadori, eds., Heyden & Son, 1973, 21. U. Geiser, H. U. G¨udel, Inorg. Chem. 1981. 20 , 3013–3019. 22. S. Kaizaki, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1970. 43 , 1100–1109. 23. S. Kaizaki, J. Hidaka, Y. Shimura, Inorg. Chem. 1973, 12 , 135–142. 24. S. Kaizaki, M. Ito, N. Nishimura, Y. Matsushita, Inorg Chem. 1985. 24 , 2080–2088. 25. G. L. Hilmes, H. G. Brittain, F. S. Richardson, Inorg. Chem. 1977. 16 , 528–533. 26. M. Cantuel, G. Bernardinelli, G. Muller, J. P. Riehl, C. Piguet, Inorg. Chem. 2004, 43 , 1840–1849. 27. S. F. Mason, R. H. Seal, Mol. Phys. 1976, 31 , 755–775. 28. S. F. Mason, Mol. Phys. 1979, 37 , 843–853. 29. F. S. Richardson, Chem. Rev . 1979, 79 , 17–36. 30. R. S. Schreiner, P. J. Hauser, Inorg. Chem. 1974. 13 , 2185–2192.

469

470

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

31. R. W. Strickland, F. S. Richardson, Inorg. Chem. 1973. 12 , 1025–1036. 32. M. C. Ernst, D. J. Royer, Inorg. Chem. 1993, 32 , 1226–1232. 33. B. Stewart, R. D. Peakock, L. Alagna, T. Prosperi, S. Turchini, J. Goulon, A. Rogalev, C. Goulon-Ginet, J. Am. Chem. Soc. 1999, 121 , 10233–10234. 34. J. Autschbach, F. E. Jorge, T. Ziegler, Inorg. Chem. 2003, 42 , 2867–2877. 35. F. E. Jorge, J. Autschbach, T. Ziegler, Inorg. Chem. 2003, 42 , 8902–8910. 36. J. Fan, M. Seth, J. Autschbach, T. Ziegler, Inorg. Chem. 2008, 47 , 11656–11668. 37. S. Kaizaki, N. Azuma, A. Fuyuhiro, Inorg. Chim. Acta, 1998 274 , 210–214. 38. T. Ama, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1970, 43 , 2654. 39. H. T. Evans, Jr., J. S. Showell, J. Am. Chem. Soc. 1969. 91 , 6881–6882. 40. E. Coronado, S. Curreli, C. Gimenez-Saiz, C. J. Gomez-Garcia, J. Roth, Synthetic Metals, 2005, 154 , 241–244. 41. T. Kudo, Y. Shimura, Bull. Chem. Soc. Jpn. 1979, 52, 1648–1652. 42. T. Kudo, Y. Shimura, Bull. Chem. Soc. Jpn. 1979, 52, 3553–3558. 43. T. Kudo, Y. Shimura, Bull. Chem. Soc,. Jpn. 1980, 53 , 1588–1594. 44. Y. Shimura, Rev. Inorg. Chem. 1984. 6 , 149–193. 45. Y. Hosokawa, Y. Shimura, Bull. Chem. Soc. Jpn. 1986, 59 , 2913–2914. 46. N. Fujimoto, Y. Mori, A. Yamagishi, H. Sato, Chem. Commun. 2010, 46 , 5473–5475. 47. S. Misumi, S. Kida, T. Isobe, Spectrochim. Acta A, 1968, 24 , 271–275. 48. S. Misumi, S. Kida, T. Isobe, Y. Nishida, H. Furuta, Bull. Chem. Soc., Jpn. 1969, 42 . 3433–3438. 49. L. I. Katzin, Inorg. Chem. 1968, 7 , 1183–1191. 50. L. I. Katzin, Inorg. Chem. 1969, 8 , 1649–1654. 51. B. Norden, I. Grenthe, Acta Chem. Scand . 1972, 26 , 407–409. 52. J. D. Saxe, T. R. Faulkner, F. S. Richardson, J. Chem. Phys. 1982, 76 , 1607–1623. 53. F. S. Richardson, T. R. Faulkner, J. Chem. Phys. 1982, 76 , 1595–1606. 54. F. S. Richardson, Inorg. Chem. 1980, 19 , 2806–2812. 55. N. C¸oruh, G. L. Hilmes, J. P. Riehl, Inorg. Chem. 1988, 27 , 3647–3651. 56. L. Di Bari, P. Salvadori, Coord. Chem. Rev . 2005, 249 , 2854–2879. 57. P. Gawryszewska, J. Legendziewicz, Z. Ciunik, N. Esfandiari, G. Muller, C. Piguet, M. Canteul, J. P. Riehl, Chirality 2006, 18 , 406–412. 58. M. A. Subhan, T. Suzuki, S. Kaizaki, J. Chem. Soc., Dalton Trans. 2002, 1416–1422. 59. M. A. Subhan, T. Suzuki, S. Kaizaki, J. Chem. Soc., Dalton Trans. 2001 , 492–497. 60. D. Shirotani, T. Suzuki, S. Kaizaki, Inorg. Chem. 2006, 45 , 6111–6113. 61. D. Shirotani, T. Suzuki, Y. Yamanari, S. Kaizaki, J. Alloys Compd . 2008, 451 , 325–328. 62. D. Shirotani, H. Sato, K. Yamanari, S. Kaizaki, Unpublished results, 2010. 63. J. L. Lunkely, D. Shirotani, K. Yamanari, S. Kaizaki, G. Muller, J. Am. Chem. Soc. 2008, 130 , 13814–13815. 64. A. C. Sen, M. Chowdhury, J. Chem. Soc., Faraday Trans., 2 , 1981, 77 , 1393–1300. 65. J. P. Morley, J. D. Saxe, F. S. Richardson, Mol. Phys. 1982, 47 , 379–406. 66. B. Bosnich, Acc. Chem. Res. 1969, 2 , 266–273. 67. E. Larsen, S. F. Mason, G. H. Searle, Acta Chim. Scand . 1966, 20 , 191–196. 68. J. Mason, S. F. Mason, Tetrahed . 1967, 23 , 1919–1927. 69. Y. Ohmori, M. Namba, Y. Kuroda, M. Kojima, Y. Yoshikawa, Inorg. Chem. 1992, 31 , 2299–2300. 70. T. Ito, A. Kobayashi, F. Marumo, Y. Saito, Inorg. Nucl. Chem. Lett. 1971, 7 , 1097–1102. 71. S. F. Mason, Inorg. Chim. Acta Rev . 1968, 89–109.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

72. Y. Ohashi, K. Yanagi, Y. Mitsuhashi, K, Nagata, Y. Kaizu, Y. Sasada, H. Kobayashi, J. Am. Chem. Soc. 1979, 101 , 4739–4740. 73. G. R. Brubaker, L. E. Webb, J. Am. Chem. Soc. 1969, 91 , 7199–7200. 74. K. Michelsen, Acta Chem. Scand . 1965, 19 , 1175–1179. 75. A. Le Guennic, W. Hieringer, A. Autschbach, J. Phys. Chem. A, 2005, 109 , 4836–4846. 76. J. Fan, J. Autschbach, T. Ziegler, Inorg. Chem. 2010, 49 , 1355–1362. 77. F. J. Coughlin, K. D. Oyler, R. A. Pascal, S. Bernhard, Inorg. Chem. 2008, 47 , 974–979. 78. J. Hidaka, B. Douglas, Inorg. Chem. 1964, 3 , 1180–1184. 79. S. Kaizaki, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1969, 42 , 988–994. 80. S. Kaizaki, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1970, 43 , 3024. 81. B. Bosnich, Inorg. Chem. 1968, 7 , 2379–2386. 82. O. Mamula, Alex von Zelwsky, T. Bark, H. Stoeckli-Evans, A. Neels, G. Bernardinelli, Chem. Eur. J . 2000, 6 , 3575–3585. 83. K. Nakanishi, J. Dillon, J. Am. Chem. Soc. 1971, 93 , 4058–4060. 84. L. Di Bari, M. Lelli, G. Pintacuda, P. Salvadori, Chirality 2002, 14 , 262–273. 85. H. Tsukube, S. Shinoda, H. Tamaki, Coord. Chem. Rev . 2002, 226 , 227–234. 86. H. Sato, D. Shirotani, K. Yamanari, S. Kaizaki, Inorg. Chem. 2010. 49 , 356–358. 87. S. G. Telfer,, N. Tajima, R. Kuroda, J. Am. Chem. Soc. 2004, 126 , 1408–1418. 88. S. G. Telfer, N. Tajima, R. Kuroda, M. Cantuel, C. Piguet, Inorg. Chem. 2004, 43 , 5302–53210. 89. S. Utsuno, J. Am. Chem. Soc. 1982, 104 , 5846–5847. 90. N. C. Payne, M. E. Thachuk, Chirality 1989, 1 , 284–292. 91. I. Bernal, Inorg. Chim. Acta 1985, 96 , 99–110. 92. I. Bernal, Inorg. Chim. Acta 1986, 115 , 25–29. 93. S. Utsuno, A. Suganuma, Y. Yoshikawa, Inorg. Chim. Acta 1996, 244 , 1–2. 94. B. Bosnich, J. M. Harrowfield, Inorg. Chem. 1975, 14 , 828–836. 95. S. Kaizaki, Bull. Chem. Soc. Jpn. 1983, 56 , 3625–3631. 96. M. Gullotti, G. Pacchioni, A. Pasini, R. Ugo, Inorg. Chem. 1982, 21 , 2006–2014. 97. M. Benedetti, L. G. Marzilli, G. Natile, Chem. Eur. J . 2005, 11 , 5302–5310. 98. G. Natile, L. G. Marzilli, Coord. Chem. Rev . 2006, 250 , 1315–1331. 99. Y. Nakano, Y. Yoshikawa, H. Kondo, J. Am. Chem. Soc. 1986, 108 , 7630–7632. 100. T. M. McCormick, S. Wang, Inorg. Chem. 2008, 47 , 10017–10024.

471

PART IV BIOMOLECULES

14 ELECTRONIC CIRCULAR DICHROISM OF PROTEINS Robert W. Woody

14.1. INTRODUCTION Circular dichroism is extensively used in the characterization of protein structure and folding because of its high conformational sensitivity. In this chapter, we will discuss electronic CD almost solely, so we will use the common abbreviation CD to refer to electronic CD and use VCD when we refer to vibrational CD. CD provides a sensitive and convenient method for determining the secondary structure of proteins. As such, it is one of the principal methods for the initial characterization of proteins and is an invaluable tool for the investigation of protein folding. CD is widely used to validate the conformational integrity of mutant proteins and to assess their conformational stability relative to the wild-type protein. CD also has important applications in detecting and quantitating ligand binding to proteins. Additional chiroptical methods for studying protein conformation are discussed in other chapters of this volume: VCD (Chapter 22) and Raman optical activity (Chapter 23). Related topics of interest to readers of this chapter include: the CD of peptides (Chapter 15), peptidomimetics (Chapter 16), nucleic acids (Chapter 17), peptide nucleic acids (Chapter 18), protein–nucleic acid interactions (Chapter 19), drug and natural products binding to nucleic acids (Chapter 20), ligand binding to serum proteins (Chapter 21), and glycoconjugates (Chapter 24). Theoretical methods for predicting protein CD are discussed in Chapter 20 of volume 1. The CD spectrum of a protein is conveniently divided into three regions, each of which is dominated by different types of chromophores and provides different kinds of information. The far UV ranges from 250 nm to the lower limit of measurements, ∼170 nm in water, and ∼120 nm in films. The dominant chromophore is the amide group, and the CD is largely determined by the secondary structure of the protein. In Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

475

476

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the near UV from 250 to 300 nm, protein CD is dominated by aromatic side chains and provides information about tertiary structure. In the proximal near UV, visible, and near IR, from 300 nm to ∼1000 nm, CD is observed only in proteins that have chromophoric prosthetic groups or other bound chromophores with absorption bands in this region. Thus this region is useful for monitoring ligand binding and ligand conformation. Recent reviews of protein CD include those of Sreerama and Woody [1, 2], Kelly and Price [3, 4], Martin and Schilstra [5], and Wallace [6]. The present chapter will focus on developments of the past decade.

14.2. THE PEPTIDE BACKBONE The far-UV CD of proteins is largely determined by the peptide backbone contributions, which arise from the peptide chromophore. Various types of secondary structure have characteristic CD patterns, which makes the far-UV CD of proteins useful for secondary structure analysis.

14.2.1. The Amide Chromophore The absorption spectra of amides have been studied extensively in the gas phase [7, 8], in solution [9], and in crystals [10]. In condensed phases, where Rydberg transitions are suppressed, only three excited states are observed down to ∼130 nm in simple amides: the n –π ∗ transition at ∼220 nm and the first (NV1 ) and second (NV2 ) π –π ∗ transitions at ∼190 and 140 nm, respectively. The n –π ∗ transition is analogous to that in other carbonyl chromophores. It is symmetry-allowed in the Cs symmetry of amides, but it is weak in absorption (εmax ∼100 M−1 cm−1 ). A large magnetic dipole transition moment (∼1BM) is directed along the carbonyl bond. The energy of the n –π ∗ transition depends on the extent and strength of hydrogen bonding. In nonpolar solvents, the amide n –π ∗ transition is at ∼230 nm, whereas in strongly hydrogen-bonding solvents, λmax ∼210 nm. The n –π ∗ transitions in tertiary amides are red-shifted by 5–10 nm relative to those in secondary amides [9]. The 190-nm (NV1 ) π –π ∗ transition has a moderate intensity (εmax ∼9000 M−1 cm−1 with a transition moment magnitude |μ|∼3 D). It occurs at 185–190 nm in secondary amides and near 200 nm in tertiary amides. The transition is polarized approximately along the C–N bond direction [10]. The NV2 transition has been observed in crystals near 140 nm, is less intense than the NV1 , and its polarization is approximately orthogonal to that of the NV1 . Ab initio calculations for amides as isolated molecules [11–13] and in cyclohexane and water [14] have been reported. These ab initio results are in good agreement with experiment.

14.2.2. General Aspects of Protein CD Isolated amide chromophores are achiral and therefore have no CD. Interactions of the n –π ∗ and π –π ∗ transitions in polypeptides are largely responsible for the far-UV CD of proteins. These interactions are of three types: (1) coupled electric dipole transition moments on the amide groups, μ–μ coupling, of which the exciton coupling between NV1 transition moments is the most important [15] (Chapters 20 in volume 1 and 4 in this volume); (2) coupling of the electric dipole transition moment on one peptide and

477

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

the magnetic dipole transition moment on another (μ–m coupling) [16]; (3) mixing of NV1 (NV2 ) and n –π ∗ excited states within a peptide group induced by the electric field of the polypeptide (one-electron or static-field coupling) [17]. These effects are all taken into account in Tinoco’s first-order perturbation theory [18] and the matrix method of Schellman and co-workers [19]. The first mechanism (μ–μ coupling) is considered in DeVoe theory [20] and in Applequist’s atom dipole interaction model [21]. These models are described in Chapter 20 in volume 1.

14.2.3. CD of Secondary Structural Elements We briefly describe the CD spectra of the major types of secondary structure that appear in proteins. More extensive information may be found in Chapter 15 in this volume. 14.2.3.1. α-Helix. Pauling and Corey’s [22] α-helix is the major element of secondary structure in many proteins and accounts for about one-third of the residues in globular proteins [23]. The α-helix was the first secondary structure to be characterized by CD [24, 25], with measurements on poly(Glu), poly(Lys), and poly(GluOMe). The α-helix CD spectrum was found to be largely insensitive to the side chains and solvent [26], so long as the α-helix was maintained. The spectrum of α-helical poly(Glu), shown in Figure 14.1, has three bands above 180 nm. The negative band at 222 nm is assigned to the n –π ∗ transition [27, 28]. The negative band at ∼207 nm and the positive band at ∼190 nm are both attributed to the π –π ∗ (NV1 ) transition, resulting from exciton coupling [15] among the π –π ∗ transition moments. The 207-nm band is polarized along the helix axis, and the 190 nm band is polarized in the plane perpendicular to the axis. In addition to these well-characterized bands, the α-helix CD spectrum has a positive band near 175 nm, which appears as a shoulder on the 190-nm band; a negative band

ππ* (perp)

30

Δε (M−1 cm−1)

20

10

0

–10 ππ* (par) –20

180

200

nπ* 220

240

λ (nm)

Figure 14.1. CD spectra of model polypeptide secondary structures. α-helix (

), poly(Glu) in water, pH. 4.5 [29]. The band assignments are indicated by labels adjacent to the extrema. β-sheet (– – –), poly(Lys-Leu), 0.5 M NaF, pH 7 [62]. Unordered conformation (· · · · · ·), poly(Glu) in water,

pH 8 [29].

478

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

near 160 nm; and a positive band below 140 nm [29]. The latter band is attributable to the NV2 transition, but the other two bands have no counterpart in simple amides. They probably result from charge-transfer transitions between neighboring amides. This assignment is consistent with ab initio calculations on small peptides [30–33]. 14.2.3.2. 310 -Helix. The 310 -helix occurs less frequently in proteins (∼3% of residues) [23]. Its CD spectrum has been characterized experimentally by Toniolo et al. [34, 35]. The long-wavelength negative bands are dramatically weaker than those of the α-helix and especially that of the n –π ∗ transition, which is only a shoulder in the 310 helix spectrum. The ratio ε222 /ε204 is ∼0.2–0.3, compared to ∼1.0 for the α-helix. Based upon theoretical calculations, a value for this ratio significantly below unity was proposed [36] as a criterion for 310 -helix formation. Although this criterion has been strongly questioned [37], the work of Toniolo et al. supports it. The 310 -helix appears to have a positive band near 190 nm, but its magnitude is uncertain. 14.2.3.3. β-Sheet. The CD spectrum of a model β-sheet is shown in Figure 14.1. A negative band near 217 nm, a positive band near 195 nm, and a negative band near 175 nm are characteristic of the β-sheet conformation [38, 39]. The 217-nm band is assigned to the n –π ∗ transition, and the 195- and 175-nm bands are attributable to exciton splitting of the π –π ∗ transition. A positive band at 168 nm is observed [40] in the spectrum of concanavalin A, a β-rich protein, and is attributed to a charge-transfer transition. In contrast to the α-helix, β-sheet CD spectra show much more variation in absolute magnitude and in relative magnitudes of the bands. This is probably a reflection of the inherent variability of β-sheets. β-sheets may be parallel, antiparallel, or mixed; the extent of twisting varies from weak to strong; they vary in the number and length of strands; and they may contain defects such as β-bulges [41]. Calculations [42] suggest that the CD of parallel and antiparallel β-sheets are similar, and this is supported by the observation [43] that the CD spectrum of pectate lyase C, a protein rich in parallel β-sheet, is similar to that of poly(Lys), which is an antiparallel β-sheet. 14.2.3.4. β-Turns. β-Turns are an important element of secondary structure in proteins. Woody [44] predicted the CD spectrum for various types of β-turns described by Venkatachalam [45]. He found several CD patterns, but the two most common types of β-turns, types I and II, were predicted to have a β-sheet-like CD spectrum, with bands red-shifted by 5–10 nm (a class B spectrum). For one type of β-turn, expected for the sequence D-X-L-Pro that adopts a type II β-turn (the prime indicates a mirror image form), an α-helix-like CD pattern (class C) was predicted. Experimental data on cyclic peptides [46–49] have supported these predictions for the type II and II turns, but indicate an α-helix-like spectrum for type I turns, in contrast to the predictions. The observed spectra for type I and II β-turns in cyclic peptide models are shown in Figure 14.2. 14.2.3.5. Unordered Polypeptides. Unordered polypeptides have been modeled by the charged homopolypeptides poly(Glu) and poly(Lys) at neutral pH. These polypeptides have a relatively weak positive CD band at ∼217 nm and a strong negative band at ∼197 nm, as shown in Figure 14.1. Tiffany and Krimm [50, 51] pointed out that this spectrum bears a strong resemblance to that of poly(Pro)II, the left-handed threefold helical form adopted by poly(Pro) in water. On this basis, they suggested that poly(Glu) and

479

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

15

Δε (M−1cm−1)

10

5

0

–5

–10

200

220

240

λ (nm)

Figure 14.2. CD spectra of model β-turns [49]. Type I turn (• • •), cyclo(L-Ala-Aha) (Aha = ε-aminohexanoic acid). Type II turn (

), cyclo(L-Ala-D-Ala-Aha).

poly(Lys) are not truly unordered but have a significant amount of the poly(Pro)II (PII ) conformation. This proposal was not widely accepted at the time, but is now supported by a large body of experimental and theoretical evidence [52, 53]. Although the positive 217-nm band of the PII or unordered conformation is assigned to the n –π ∗ transition, the strong negative band at 200 nm and the absence of any positive counterpart are not consistent with a dominant exciton effect in the π π ∗ region. Woody [54] has shown that mixing of the NV1 π –π ∗ transition with transitions in the deep UV occurs in the PII conformation and is responsible for suppressing the exciton contributions and generating the strong negative CD band near 200 nm.

14.2.4. Secondary Structural Analysis of Proteins One of the most important and widely used applications of CD spectroscopy is the analysis of the secondary structure of proteins. The CD spectra of proteins frequently show the characteristics of their dominant secondary structure, as illustrated in Figure 14.3, which shows the CD spectra of hemoglobin, an α-rich protein; concanavalin A, a βrich protein; and α-synuclein, a predominantly disordered protein. However, because the α-helix spectrum is much more intense than the β-sheet spectrum, proteins with significant amounts of both secondary structures generally show the qualitative features of the α-helix. Figure 14.4 shows the CD spectra of ribonuclease A (21% α, 33% β) and subtilisin (30% α and 18% β), which both have α-like features. Also shown in Figure 14.4 is the CD spectrum of elastase (11% α, 34% β, and 34% unordered), the spectrum of which resembles that of unordered polypeptides. Clearly, protein CD spectra contain important information about secondary structure, but extracting this information is not a simple task. Methods of secondary structure analysis have been reviewed extensively [2, 55–61].

480

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

14

Δε (M−1 cm−1)

9

4

–1

–6

170

180

190

200

210 220 λ (nm)

230

240

250

), an α-rich protein; concanavalin A (• • •), a β-rich protein; α-synuclein (ooo), an intrinsically disordered protein. The myoglobin and

Figure 14.3. CD spectra of three proteins: myoglobin (

concanavalin A spectra are from the CDPro database [83] and were a private communication from W. C. Johnson. The α-synuclein spectrum is from reference 210.

6

Δε (M−1 cm−1)

4

2

0

–2

–4 170

180

190

200

210 220 λ (nm)

230

240

250

), subtilisin (• • •), and elastase (ooo). The spectra are from the CDPro database [83] and were a private communication from W. C. Johnson.

Figure 14.4. CD spectra of ribonuclease A (

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

A basic assumption in secondary structure analysis by CD is that the spectrum is a linear combination of component secondary structure spectra:  fi Bi λ , (14.1) Cλ = λ

where Cλ is the observed CD at wavelength λ, fi is the fraction of secondary structure i , and Bi λ is the CD of secondary structure i at wavelength λ. Additional implicit assumptions are that the contributions are additive and that the effects of tertiary structure, aromatic side chains, and variations in secondary structure geometry and segment length are negligible. Although these assumptions are not always valid, current methods minimize their effects by using flexible methods in the analysis. Initially, spectra of the secondary structure types, Bi λ , were taken from model systems as in Figure 14.1 [38, 62]. However, the model systems have long and regular helices and strands and thus differ significantly from the relatively short and irregular secondary structure elements that occur in globular proteins. Consequently, later methods have derived, explicitly or implicitly, the secondary structure CD spectra from the CD spectra of proteins with structures known from X-ray diffraction (reference proteins). The secondary structure spectra and the secondary structure fractions have been derived by many methods, using simple least squares [63–66], ridge regression [67, 68], singular value decomposition [69–72], neural networks [73–77], principal component factor analysis [78], convex constraint analysis [79], partial least squares [80, 81], and support vector machines [80]. The importance of flexibility in secondary structure analysis by CD was recognized by Manavalan and Johnson [70] and implemented in their Variable Selection Method. Flexibility permits the program to adapt the set of reference proteins to best describe the CD spectrum of the protein under analysis and thus helps overcome limitations imposed by non-peptide CD contributions and variations in helix and strand geometry and length. Flexibility has also been introduced by ridge regression (CONTIN) [67], the locally linearized model [71], the self-consistent method (SELCON) [72], CDsstr [82], and neural network methods [73–77]. The databases of CD spectra for structurally characterized proteins have undergone significant development in recent years. The CDPro website [83] (a list of website addresses and contacts for software useful in the analysis and prediction of protein CD spectra is provided in Section 14.6) offers 10 sets of CD spectra for various wavelength ranges (from 178–260 nm to 190–240 nm) and permits analysis for various combinations of secondary structure. Some sets include denatured proteins, and these are well-suited for analyzing intrinsically disordered proteins and protein folding/unfolding transitions [84]. Other sets contain integral membrane proteins and are useful for the analysis of such proteins [85]. A basis set containing 50 proteins (RaSP50) has been constructed by Oberg et al. [86]. The proteins were selected to cover all known protein folding types and the full range of α-helix and β-sheet contents. The authors acquired the CD spectra from 185–260 nm for each protein and analyzed the reported crystal structure data for secondary structure by DSSP [87]. Increasing availability of synchrotron light sources for CD spectroscopy [6] has stimulated the development of protein CD basis sets that extend into the vacuum UV. Matsuo et al. [88, 89] reported a database for 31 proteins that extends to 160 nm. Lees et al. [90] described two databases, SP175 and SP170, with short-wavelength cutoffs of 175 nm (72 proteins) and 170 nm (39 proteins), respectively. Like the RaSP50 database

481

482

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

[86], the proteins were selected to provide a broad representation of protein folds and secondary structure contents. The SP175 data set is available on the DICHROWEB server [91, 92] and in the Protein Circular Dichroism Data Bank [93, 94]. Although publications describing methods of secondary structure analysis report performance statistics, most have reported comparisons with only a few other methods. There have been few broader comparisons of methods. Greenfield [58] compared CONTIN [67], VARSLC [70], SELCON [72], and K2D [74], as well as several older methods that lacked flexibility. Greenfield recommended SELCON, CONTIN, and K2D for analysis of globular protein structure. Lees et al. [80] compared SELMAT3 (their version [90] of SELCON3 [83]); PLS, partial least squares; SIMPLS, simultaneous partial least squares; PCR, principal component regression; NN, neural network; SIMPL-NN, simultaneous partial least squares neural network; and SVM, support vector machine methods. Cross-validation was performed with the SP175 database [90]; that is, each of the 72 proteins was analyzed using the remaining proteins as the data base. Two performance measures were used: the Pearson product-moment correlation coefficient [95], r, which is 1 for perfect agreement between calculated (CD) and observed (X-ray) structure fractions and is 0 for random agreement; and δ, the root-mean-square deviation between calculated and observed fractions. In all tests, the performance of the various methods was quite similar for all methods. For example, in a three-state model (α-helix, β-sheet, other), r for the α-helix content varied from 0.957 to 0.971, and δ ranged from 0.052 to 0.063. No one method was best for the all three structural types, but the SIMPL-NN model performed best for the β-sheet and other, whereas the PLS method did best for the α-helix. Tests were also performed for the six-state model used by Sreerama et al. [96] to calculate the numbers of α-helix and β-strand segments. Again, the range of performance parameters was narrow for each structural type. In this test, only two methods (SELMAT3 and PLS) were compared. SELMAT3 performed best for two structural types, and PLS worked best for four types. Secondary structure analysis by the methods described above requires an accurate protein concentration because the per-residue ε or [θ ] is used. The effects of errors in the protein concentration were discussed by Hennessey and Johnson [69], who showed that a 5–10% error in concentration propagated as proportional changes in secondary structure contents. Therefore, it is important to use spectrophotometrically determined concentrations and not ones based upon less accurate methods such as dye binding. For some samples (e.g., with protein films), it is impossible to obtain accurate protein concentrations. McPhie [97–99] and Goormaghtigh and co-workers [100, 101] have developed methods that are independent of protein concentration. While not as accurate as the methods previously described, they are sufficiently accurate to be useful in cases in which accurate protein concentrations are unavailable. McPhie’s method [97–99] utilizes the Kuhn anisotropy ratio, g, which is the ratio ε/ε. He measured g-factor spectra from 190nm to 240 nm for a basis set of 32 proteins and used these in a constrained least squares or ridge regression analysis to derive secondary structures. The analyses for α-helix and β-sheet were satisfactory with r ∼0.85 and 0.75, respectively, and rms deviations of ∼0.1. Results for turns and unordered structure were inferior, with r∼0.4. In the method of Raussens et al. [100] (note also the erratum [101]), the observed spectrum is first normalized to an ellipticity of 1 at 207 nm, a wavelength selected to minimize the deviations of the α-helix contents of the set of 50 reference proteins [86]

483

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

from X-ray values. Then, for each type of secondary structure, a model of the form fi = a1 + a2 Eλ1 + a3 Eλ1 2 ,

(14.2)

where Eλ1 is the normalized ellipticity of the test protein at wavelength λ1 , is used. The wavelength λ1 and the coefficients a1 , a2 , and a3 were optimized to give the best fit to the secondary structure contents of type i in the set of reference proteins. For the α-helix and β-sheet, this model was augmented by adding linear and quadratic terms for a second wavelength. For the turns and unordered conformation, the additional terms were unnecessary. This method gives standard deviations of ∼10–12% for the α-helix and β-sheet, comparable to those from the method of McPhie [97–99]. It is possible to extract information from far-UV CD spectra beyond the content of secondary structures. The number of α-helices and β-strands in a protein can be estimated [96, 102]. The basis for this lies in the fact that amide groups at the ends of α-helical or β-strand segments differ in their environment and hence their CD from those in the middle of the segments. Sreerama and Woody assumed that nα and nβ residues per helix and strand, respectively, are affected by being terminal residues. For each helix in a reference protein, nα residues were assigned as “distorted,” hD , and the remainder were considered as “regular,” hR , and correspondingly for each beta strand, giving βD and βR . Analysis of the reference proteins showed that the values of nα and nβ that gave the best values for helix and strand numbers are nα = 4 and nβ = 2. With these choices, the numbers of helices and strands can be estimated with an uncertainty of ∼ ±3. Protein secondary structure can also be derived from IR absorption [103–105], Raman [106–108], VCD (Chapter 21 in this volume) and Raman optical activity (Chapter 22 in this volume). Methods for combined analysis of CD and IR absorption spectra have been developed [109–113]. VCD and CD spectra have also been combined [111, 114]. In general, the combined analysis of CD and vibrational spectroscopy, IR absorption, or VCD has been found to improve the results over the individual methods. CD gives superior results for the α-helix whereas vibrational spectroscopy provides better results for β-sheets. Thus, the two types of spectroscopy complement each other.

14.3. PROTEIN SIDE CHAINS Side-chain chromophores in proteins include the three aromatic side chains (Phe, Tyr, and Trp) and the disulfide group of cystine. These chromophores have absorption bands in the near UV and are responsible for the near-UV CD of proteins. They also have higher-energy transitions, in the far UV, which are much more intense than the near-UV bands. In the far-UV, side-chain CD is usually obscured by the much more numerous amide groups but, in some cases, especially strong aromatic CD combined with weak backbone CD permits the detection and characterization of side-chain CD bands in the far UV. Side-chain CD has been reviewed [115–118].

14.3.1. Near-UV CD The near-UV CD of proteins is dominated by aromatic side-chain contributions. The CD of these chromophores is determined by coupling with the peptide backbone and with other aromatic groups. Globular proteins with a well-defined tertiary structure generally exhibit a near-UV CD spectrum, often with some distinct fine structure. Thus, the nearUV CD spectrum is sensitive to the protein’s tertiary structure. Conformational changes

484

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

or ligand binding that affect the geometry or environment of one or more aromatic side chains can therefore perturb the near-UV CD spectrum. Because of the relatively small numbers of aromatic chromophores, such perturbations may be more easily detectable than those of the far-UV CD, which require a significant fraction of the residues in the protein to be perturbed. Each of the three aromatic amino acids gives rise to absorption and CD bands with characteristic wavelengths and fine structure. These features frequently permit assignment of specific features in the near-UV CD spectrum to one type of aromatic side chain. Site-directed mutagenesis can provide definitive assignment to a specific residue in the sequence. A chromophoric residue can be mutated to a weaker chromophore or a nonchromophoric residue. The difference CD spectrum of (wild-type − mutant) provides the CD contribution of the mutated residue to the wild-type spectrum, if the mutation has not significantly altered the conformation. Such studies have been performed for a substantial number of aromatic residues in several proteins: barnase [119], carbonic anhydrase II [120], dihydrofolate reductase [121], gene protein 5 from bacteriophage fd [122], pancreatic trypsin inhibitor [123], human tissue factor [124], and ribonuclease A [125]. For a number of these systems, theoretical calculations have given good agreement with the sign and approximate magnitude of the individual aromatic CD contributions [118, 123, 125–127]. Disulfides also may contribute significantly to protein CD. Their near-UV CD generally consists of a relatively weak and broad band with a maximum near 260 nm. The disulfide peak is generally obscured by aromatic CD bands, but a long-wavelength tail may extend to 300 nm or above.

14.3.2. Far-UV CD Aromatic CD bands in the far-UV are most readily observed in β-rich proteins with relatively weak backbone CD—for example, immunoglobulin folds, lectins, and snake toxins. Positive bands near 230 nm are generally due to aromatic side chains because, except for the PII conformation, none of the common secondary structures have positive CD in this region, and the long-wavelength positive band of the PII conformation is relatively weak. Proteins that exhibit positive CD bands near 230 nm include avidin [128], gene 5 protein from filamentous phages [122, 129], and cobrotoxin [130]. For the most favorable backbone and side-chain conformations, coupling between the La transition of Tyr or Phe and the amide transitions of nearest-neighbor peptide groups favors positive rotational strength in the La band [131]. However, there appears to be no strong bias toward positive rotational strengths for the 225-nm Bb band of Trp [132]. It is interesting to note that all of the aromatic amino acids in the form of their N -acetyl, N  -methylamides exhibit positive La (Tyr, Phe) and Bb (Trp) CD bands [133]. Aromatic side chains are frequently found in clusters in proteins [134], so exciton couplets involving aromatic side chains might be expected. The dependence of the exciton couplet strength on the square of εmax (Chapter 4) excludes observable exciton couplets in the near-UV. However, the intense Bb band of Trp (εmax ∼35, 000 M−1 cm−1 ) can give rise to strong couplets in Trp–Trp pairs and, through coupling with the nearly degenerate La band of Tyr (εmax ∼10, 000), in Trp–Tyr pairs. Trp–Trp exciton couplets have been reported in chymotrypsin and chymotrypsinogen [126, 135], dihydrofolate reductase [121, 136], and hemoglobin variant III from Chironomus thummi thummi [137]. A Trp–Tyr exciton couplet has been established by site-directed mutagenesis in PagP, a bacterial outer membrane enzyme that transfers fatty acyl groups [138]. The coat protein

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

of bacteriophage fd has an exciton couplet-like feature superimposed on an α-helix CD pattern [139], with a negative lobe at 223 nm and a positive lobe at 210 nm. This may represent a Trp (Bb )-Phe (La ) exciton couplet.

14.4. EXTRINSIC CHROMOPHORES For proteins that lack prosthetic groups, the electronic CD spectrum effectively ends at 300 nm. Therefore, the CD bands of chromophoric prosthetic groups, metal ions, substrates and substrate analogues, and inhibitors can be studied at wavelengths above 300 nm without interference from intrinsic protein groups. CD bands from such chromophores are called extrinsic CD bands. Thus, CD in the very near UV and visible regions is useful for studying protein–ligand interactions. In most cases, the free ligands are achiral or, if chiral, are flexible and therefore have only weak CD when free. Thus, the extrinsic CD is characteristic of the bound ligand and its interaction with the protein. The CD of protein–ligand complexes has been reviewed [140–143], and Chapter 21 also provides many examples. Extrinsic CD bands can arise from three mechanisms: (1) inherent chirality of the bound chromophore, (2) coupling of transitions on the chromophore with transitions in the peptide and aromatic groups of the protein, and (3) mixing of transitions within the chromophore caused by the static field of the protein. Of these mechanisms, the most readily analyzed case is (1) because it depends only on the geometry of the ligand, whereas the other two mechanisms involve interactions with the protein. Many extrinsic chromophores that are achiral when free are actually unresolvable mixtures of enantiomeric or nearly enantiomeric conformers. Binding to a protein selects one of these conformers. Examples include 11-cis retinal [144], bilirubin [145], and di- and triphenyl methyl dyes [146–148].

14.5. APPLICATIONS 14.5.1. Protein Folding CD, NMR, and fluorescence are the major biophysical tools in the study of protein folding. The use of CD in monitoring the kinetics of protein folding has been reviewed [149, 150]. CD secondary structural analysis provides information about the native and the unfolded states of the protein, the two end points of the folding process. CD has also yielded important information about a key type of intermediate in the folding process, the molten globule (MG) in which the protein largely retains its native secondary structure but lacks a defined tertiary structure. MGs were first recognized as partially unfolded proteins observed at low pH (for reviews, see references 151 and 152). The protein αlactalbumin is a widely studied example. Far- and near-UV CD spectra of α-lactalbumin are shown in Figure 14.5 [153]. In the far-UV CD, spectra 1 and 2 are for the native protein at neutral pH in the presence and absence of Ca2+ , respectively; spectrum 3 is taken at pH 2 and is that of the MG; spectra 4 and 5 are taken at higher temperatures of 41◦ C and 78◦ C, respectively; and spectrum 6 is that of the fully unfolded protein in 6 M guanidinium chloride (GuCl). The squares and circles will be discussed in the next paragraph. The MG spectrum differs from those of the native protein and the protein that is unfolded by heat or GuCl, but it resembles the native spectrum more than those of the unfolded forms. In the near UV, spectra 1–3 are as for the far UV; spectrum 4 is that

485

486

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3

0

4 5

[θ] (deg.cm2.dmol–1)

[θ] x10–3 (deg.cm2.dmol–1)

0

–5 5 1

–10

5

4

–15

200

2

–100

–200

–300

1 2

3

220 240 Wavelength (nm)

260

(a)

240

280 300 260 Wavelength (nm)

320

(b)

Figure 14.5. CD spectra of the native, unfolded, and molten globule (A-state) of α-lactalbumin in the far-UV (a) and near-UV regions (b). The open circles and squares show the CD values obtained by extrapolating to zero time of the refolding curves. In both panels, spectra 1 and 2 are for the holo and the apo forms, respectively, in the native state; spectrum 3 is for the A state. In (a), spectra 4 and 5 are of the thermally unfolded protein at 41◦ C and 78◦ C, respectively; spectrum 6 is that of the unfolded state in GuCl. In (b), spectrum 4 is that of the thermally unfolded protein at 62.5◦ C; spectrum 5 is that of the GuCl-unfolded state. (Reprinted with permission from Kuwajima et al. [153], © 1985 American Chemical Society.)

of thermally unfolded protein at 62.5◦ C; and spectrum 5 is that of the GuCl-unfolded protein. We see that the MG has a near-UV CD spectrum comparable to those of the unfolded forms and qualitatively different from that of the native form. These data lead to the conclusion that the MG of α-lactalbumin has secondary structure that resembles the native protein but lacks well-defined tertiary structure. The open symbols in the CD spectra of α-lactalbumin (Figure 14.5) are derived from kinetic folding experiments [153]. Stopped-flow experiments monitored by CD were used to study the refolding of the GuCl-unfolded protein on dilution in neutral buffer. Extrapolation of the signal observed at each wavelength to zero time generated the points shown in Figure 14.5. That these points nearly coincide with the equilibrium MG spectrum (curve 3) implies that during the dead time of the experiment (3 s in the data shown, but subsequent measurements with a 15-ms dead time [154] demonstrate that the intermediate forms on the millisecond timescale) the protein folds to a conformation with substantial secondary structure but no well-defined tertiary structure. In addition to the similarity of the CD spectrum of this rapidly formed intermediate to that of the equilibrium MG, the unfolding of the kinetic intermediate by GuCl, monitored by CD, coincides with that of the equilibrium MG [154]. Kinetic bursts corresponding to the formation of MG-like intermediates have been observed for many proteins (reviewed in reference 155), and such species are now generally accepted as key intermediates in the folding of most proteins [152]. Another aspect of protein folding in which CD has had a major impact is the recognition and characterization of intrinsically disordered proteins (IDPs) [156, 157]. Recently,

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

it has been established that many proteins, especially from eukaryotes, have extensive regions that are disordered in the native state, in some cases extending throughout the protein [156]. Many lines of evidence, including far-UV CD, now support this onceheretical view. IDPs have far-UV CD spectra characteristic of unordered polypeptide chains (Section 14.2.3.5). These proteins usually fold into well-defined structures upon binding to their targets—other proteins, nucleic acids, or other ligands. These coupled folding/binding reactions are conveniently followed by CD. The CD of IDPs generally becomes more negative near 220 nm and less negative near 200 nm with increasing temperature. These changes are nearly linear in temperature. This behavior has sometimes been interpreted as indicating “temperature-induced formation of secondary structure” [158]. Although the direction of the observed changes is consistent with α-helix and/or β-sheet formation, it is also consistent with a melting of PII -helix [157]. Recently, a combined CD, NMR, and small-angle X-ray scattering study of several IDPs has provided strong evidence for this latter interpretation [159].

14.5.2. Membrane Proteins Membrane proteins constitute about one-third of the proteins coded for in the human genome. CD spectroscopy should be especially useful in characterizing the secondary structure of these proteins because both X-ray diffraction and NMR are difficult to apply to these systems—they are difficult to crystallize and their size poses problems for NMR methods. Despite this, relatively few CD studies of membrane proteins have been reported. Membrane protein CD is known to be subject to artifacts produced by inhomogeneous samples and by differential scattering of left- and right-circularly polarized light [160]. Both of these problems can be avoided by solubilizing the protein in nonionic detergents, thereby generating a molecularly dispersed solution. It has been claimed that the lipid environment of membrane proteins leads to wavelength shifts in their CD spectra that make soluble proteins unsuitable for analyzing the CD spectra of membrane proteins. Wallace et al. [161] analyzed the CD spectra of eight membrane proteins and reported that the use of soluble proteins in the analysis gave inaccurate results. They attributed this to the problem of lipid-induced wavelength shifts and called for the development of a basis set of membrane proteins. Sreerama and Woody [85] developed a basis set of membrane proteins, based upon data of Park et al. [162], who reported CD spectra for 30 membrane proteins. At the time of the work of Park et al., structures were available for only four of the proteins, and one of these was of low resolution. Twelve years later, high-resolution structures were available for nearly half (13) of the proteins, providing an adequate set of reference proteins. Sreerama and Woody found that secondary structure analysis of the membrane proteins using a soluble protein basis set gave acceptable results, only slightly inferior to those obtained using the membrane protein basis set. The best results were obtained with a combined basis set. Sreerama and Woody pointed out that the data of Park et al. do not support the idea [161] of significant wavelength shifts in membrane protein CD. The wavelength maximum of the positive π –π ∗ CD band for α-rich membrane proteins ranges from 192 to 196 nm, whereas that for α-rich soluble proteins ranges from 192 to 195 nm. Thus, a special set of reference proteins is not required for analyzing the secondary structure of membrane proteins. Turk et al. [163] used the combined soluble + membrane protein basis set [85] to analyze the secondary structure of a bacterial Na+ /galactose co-transporter. They found 85% α-helix, in good agreement with sequence analysis and topological analysis that

487

488

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

indicated 14 trans-membrane helices. Turk et al. also analyzed the CD of a fusion protein between the co-transporter and green fluorescent protein (GFP), a β-rich protein with only one turn of α-helix. They obtained 60% α-helix which, corrected for the presence of the β-rich GFP component, is equivalent to the helix content of the co-transporter. Matsuo et al. [164] have used vacuum-UV CD to study the conformation of α1 glycoprotein (AGP) in solution and bound to liposomes. Spectra of the protein component were obtained by subtraction of the glycan spectrum, which contributes 10–20% of the CD at 193 nm. Matsuo et al. [89] used their vacuum-UV CD data for 31 reference proteins to determine the helix content and the number of helices and strands. Native AGP was found to contain 14% α-helix in three segments and 38% β-sheet in 10–11 strands. On binding to liposomes, the α-helix content increased to 50% in 7–8 segments and the β-sheet content decreased to 3–6% in two strands. Matsuo et al. [165] combined this information with a sequence-based secondary structure prediction method using a neural network and thus predicted the locations of the α-helical and β-strand segments in the amino acid sequence for the native and liposome-bound protein. The bacterial membrane protein phospholipid, lipid A palmitoyltransferase (PagP), is a β-barrel protein that transfers acyl chains from phospholipids in the outer leaflet of the outer membrane to a lipid A precursor to form lipid A. It has a strong preference for C16 , palmitoyl, chains. The NMR [166] and X-ray [167] structures of PagP reveal that the central cavity of the β-barrel is closed at the bottom by a glycyl (Gly88) residue, thus defining the preferred length of the acyl chain that can be transferred. Bishop and coworkers [138] probed the nature of this “molecular ruler” by mutating the Gly at the bottom of the cavity to Ala, Cys, S-methyl Cys, and Met. As the length of the replacement side chain increases, one would expect the length of the preferred acyl chain to decrease by one methylene group for each additional atom in the side chain. This was found to be the case for all mutants except Gly88 → Cys, in which case C14 and C15 acyl chains were transferred with equal facility; that is, the acyl chain specificity is lost. Concomitant studies by CD showed that the wild-type protein and all mutants except Gly88 → Cys show a positive CD band at about 232 nm that was suspected to arise from an aromatic side chain. Theoretical calculations [138] based upon the crystal structure [167] predicted a strong positive couplet centered at about 228 nm associated with Tyr26 and Trp66, the side chains of which are near each other and near Gly88 at the bottom of the pocket. This prediction was verified by site-directed mutagenesis, which demonstrated that the positive 232-nm band disappears on mutation of either Tyr26 (Tyr26 → Phe) or Trp66 (Trp66 → Phe), whereas the negative CD band at 218 nm decreases. Thus, both theory and experiment show the existence of a Trp–Tyr exciton couplet (see Chapter 4 in this volume) in PagP, and the presence of this couplet correlates exactly with the specificity of acyl-chain transfer. The CD at 232 nm provides an excellent probe for the intact tertiary structure of PagP and is useful for monitoring protein stability in the wild-type and mutant proteins. The cause of the breakdown of the couplet and of acyl chain-length specificity was elucidated later [168], again with help from the exciton CD. In the original study [138], wild-type and mutant proteins were refolded at pH 8, the pH optimum for enzyme activity. When the Gly88 → Cys mutant was refolded at lower pH, down to pH 5, the exciton couplet was observed and strong acyl chain-length specificity for C14 was observed at pH7. Using the exciton intensity at 232 nm, PagP refolding at various pHs was shown to follow the titration of Cys88, and a pKa of 7.5 was obtained, which is about 2 pH units below the normal pKa of a Cys–SH group (∼9.5). Above the pKa , the charged thiolate group at the bottom of the cavity perturbs both the Tyr26–Trp66 exciton coupling and

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

the operation of the molecular ruler. In this system, exciton CD permits the determination of the pKa of a specific ionizable group in PagP.

14.5.3. Heme Proteins Heme proteins have been extensively studied by CD, and there has been great interest in the extrinsic bands of the heme [169–171]. The origin of the heme CD bands was elucidated by Hsu and Woody [172], who calculated the CD of the Soret and other heme bands for myoglobin and hemoglobin, using the available crystal structures. They were able to account for the sign and approximate magnitude of the Soret and visible bands and of the UV bands near 350 and 260 nm by the coupled oscillator interactions of the heme transitions with the π –π ∗ transitions of the aromatic side chains. The contributions of coupling with peptide transitions were found to be small. This interpretation was strengthened by calculations on hemoglobin from an insect [173, 174] and a lamprey [175], which showed that the same mechanism accounts for the CD bands in these proteins, with Soret bands opposite in sign to those of mammalian hemoglobins. A recent study by Dartigalongue and Hache [176] utilized Applequist’s method [177] to calculate the Soret CD of MbCO and Mb in the native form and in postulated intermediates following photolysis of the CO from MbCO. Their results for the native form are consistent with those of Hsu and Woody [172]. They found that although the net contribution of the proximal His to the Soret CD is small, in accord with Hsu and Woody, the orientation of the proximal His has a dramatic effect on the Soret CD spectrum. Rotation of the proximal His by 30◦ is predicted to lead to nearly an order of magnitude decrease in the Soret rotational strength. A potential role for inherent chirality in heme protein CD has been suggested by the observation that heme isomers in myoglobin (Mb) have very different Soret CD bands. The isomers differ by a 180◦ rotation about an axis through the α- and γ -methine carbons. NMR studies of Lamar et al. [178] showed that native MbCN consists of a 9:1 mixture of isomers A and B. MbCO freshly reconstituted from heme and globin consists of a 50:50 A:B mixture and its Soret CD has only about 50% of the magnitude of the equilibrium species [179]. Extrapolation of the linear relationship between the Soret CD intensity and the fraction of isomer B to zero and to one showed that the Soret band of the dominant isomer, A, is strong and positive, whereas that of the minor isomer, B, is weakly negative [180]. The model of Hsu and Woody [172] does not predict such a dramatic change upon 180◦ rotation of a planar heme about an in-plane axis. A nonplanar heme, however, might explain the results because the 180◦ rotation could reverse the sign of the intrinsic heme contributions while the aromatic coupling contributions undergo only minor changes. MD simulations for the major isomer, combined with CD calculations, indicate that the heme–aromatic coupling is the largest single contributor (∼40 %) and that intrinsic contributions of heme nonplanarity and heme–peptide contributions each contribute ∼30% of the Soret CD intensity [181]. The heme in cytochrome c is substantially distorted from planarity by its covalent attachment to the protein [182], and this distortion is likely to be preserved in the heme undecapeptide produced by proteolytic cleavage. A combination of molecular dynamics and CD calculations on the heme undecapeptide provided a good description of the Soret CD, which showed a significant contribution from heme nonplanarity [183]. Near-UV CD has been useful for spectroscopically monitoring the allosteric transition in hemoglobin (Hb). Perutz et al. [184] reported that deoxyHb has a distinct negative band at 287 nm, whereas the CD of HbO2 is weak from 280 to 300 nm. They proposed

489

490

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the negative band as a marker for the R → T transition, where R is the relaxed quaternary structure characteristic of liganded Hb, such as HbO2 , and T is the taut conformer found in deoxyHb. Perutz et al. proposed that the 287-nm band in deoxyHb arises from two aromatic side chains in the α1 β2 interface that is critical in the R → T transition: Tyr42 in the α-chain (α42 Tyr) and Trp37 in the β-chain (β37 Trp). These aromatic side chains are H-bonded in T but not in R, and this might account for the difference in CD. There are two additional aromatic residues in the α1 β2 interface, α140 Tyr and α145 Tyr. Nagai and co-workers [185] have explored the contributions of the four aromatic side chains to the difference CD of the R → T transition, using mutants in which each of the residues is replaced by a nonchromophoric residue. The residues identified by Perutz et al. [184] contributed 4% (α42 Tyr) and 18% (β37 Trp), whereas α140 Tyr and α145 Tyr contributed 32% and 27%, respectively. Jin et al. [185] suggested that α140 Tyr and α145 Tyr may be more important in the difference CD spectrum because they shift from solvent-exposed positions in the R conformer to hydrophobic pockets in the T conformer. By contrast, although α42 Tyr and β37 Trp change their H-bonding states, their hydrophobic environments do not undergo a change in the R → T transition. In another study, Li et al. [186] compared the near-UV CD of isolated α- and βsubunits of Hb to demonstrate that the change in CD at 287 nm is not solely due to quaternary structural differences. About 50% of the difference CD is associated with changes in tertiary structure of the individual subunits induced by oxygen binding. Applications of CD spectroscopy to the study of cytochrome c have been reviewed [187]. Far-UV CD spectra of cytochrome c down to 175 nm have been analyzed [187] by DichroWeb [92, 188] to characterize the secondary structure of the protein in four states (II to V, in order of increasing pH). In state III, which prevails at physiological pH, the content of α-helix, β-sheet, and β-turns was found to agree well with the X-ray structure [189]. State II, at pH 3, appears to be a molten globule state, with both α-helix and β-turn contents showing small decreases relative to the native form. The two high-pH forms, IV and V, showed significant decreases in secondary structure relative to state III, but even state V had 31% α-helix and 20% β-turn, suggesting that the pH-induced conformational changes are predominantly at the tertiary structure level. Schweitzer-Stenner and co-workers [190, 191] have used the visible CD spectrum of cytochrome c, associated with the Soret and visible bands, to measure the electrostatic field created by the protein at the heme iron. The Soret and visible bands are each associated with two nearly degenerate transitions that are split by the electrostatic field (Stark effect) arising from the protein. This splitting manifests itself in the CD spectrum as couplets. Simulation of the CD spectra permitted determination of the field strength for both oxidized and reduced cytochrome c [191]. Oxidized cytochrome c has absorption and CD bands at 695 nm, attributed to charge transfer from heme → Fe3+ [192] or Met80 S → Fe3+ [193]. This band has several subbands, probably arising from different protein conformers. The 695-nm band disappears upon opening of the heme crevice during unfolding. The temperature dependence of the sub-bands was analyzed, providing thermodynamic data for the unfolding of the various conformers [194].

14.5.4. Retinal Proteins The CD of rhodopsin has long been of interest because of its potential for providing information about the chirality of the retinal chain. A major concern is the question of

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

whether the extrinsic CD of rhodopsin is dominated by inherent chirality [195–200] or by coupling with protein transitions [201–203]. Recently, Pescitelli et al. [204] have reported theoretical calculations that strongly support inherent chirality as the dominant mechanism. They used time-dependent density functional theory (see Chapter 22 in volume 1) to calculate the CD of the bound chromophore, using geometries from several crystal structures, and to generate the transition parameters necessary to calculate the contributions of coupling (see Chapter 20 in volume 1) with aromatic and peptide groups in the protein. The two contributions were of opposite signs for both the α (500 nm) and β (340 nm) bands. For the α band, the inherent chirality contribution was twice as large as that from coupling, and for the β band the inherent contribution was much more dominant. Thus, in this case, inherent chirality is dominant, but this may not be true in other retinal proteins (e.g., bacteriorhodopsin).

14.6. COMPUTER RESOURCES Computer programs for performing secondary structure analysis and for simulating protein CD spectra are available at the websites or through the contacts listed below. CDPro [83] SELCON3 [83], CDsstr [82], CONTINLL [67, 83] Contact: [email protected] or [email protected] Internet: http://lamar.colostate.edu/∼sreeram/CDPro CDsstr [82] CD analysis and XTLsstr [205] crystal structure analysis http://biochem.science.oregonstate.edu/dr-johnsons-softwar-downloadinstructions CONTIN [67] ridge regression Internet: http://s-provencher.com/pages/contin-cd.shtml K2D2 [77] neural network Internet : http://www.ogic.ca/projects/k2d2 CCA[79] convex constraint analysis Internet : http://www.chem.elte.hu/departments/jimre/ DICHROWEB [91, 92] web-based secondary structure analysis Internet: http://www.cryst.bbk.ac.uk/cdweb/html/home.html DICHROPROT [206] linear regression [38], SELCON2 [72], SELCON3 [83], CONTIN [67], VARSLC [70] Internet: http://dicroprot-pbil.ibcp.fr PROTEIN CD DATA BANK [93, 94] deposition of and access to protein CD data Internet: http://pcddb.cryst.bbk.ac.uk/home.php DICHROCALC [207] web-based protein CD calculation Internet: http://comp.chem.nottingham.ac.uk/dichrocalc MATMAC [208] Tinoco [18] and matrix method [19] calculations of protein CD Contact: joerg.fl[email protected] PROTEIN [209] matrix method [19] calculations of protein CD Contact: [email protected]

491

492

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

REFERENCES 1. N. Sreerama, R. W. Woody, Circular dichroism of peptides and proteins, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., John Wiley & Sons, New York, 2000, pp. 601–620. 2. N. Sreerama, R. W. Woody, Meth. Enzymol . 2004, 383 , 318–351. 3. S. M. Kelly, N. C. Price, Biochim. Biophys. Acta Protein Struct. Mol. Enzymol . 1997, 1338 , 161–185. 4. S. M. Kelly, T. J. Jess, N. C. Price, Biochim. Biophys. Acta Proteins Proteom. 2005, 1751 , 119–139. 5. S. R. Martin, M. J. Schilstra, Meth. Cell Biol . 2008, 84 , 263–293. 6. B. A. Wallace, Q. Rev. Biophys. 2009, 42 , 317–370. 7. M. B. Robin, Amides, in Higher Excited States of Polyatomic Molecules, Vol. 2, Academic Press, New York, 1974, pp. 122–146. 8. M. B. Robin, Amides, in Higher Excited States of Polyatomic Molecules, Vol. 3, Academic Press, Orlando, FL, 1985, pp. pp. 303–308. 9. E. B. Nielsen, J. A. Schellman, J. Phys. Chem. 1967, 71 , 2297–2304. 10. L. B. Clark, J. Am. Chem. Soc., 1995 117 , 7974–7986. 11. L. Serrano-Andres, M. P. F¨ulscher, J. Am. Chem. Soc. 1996, 118 , 12190–12199. 12. L. Serrano-Andres, M. P. F¨ulscher, J. Am. Chem. Soc. 1996, 118 , 12200–12206. 13. J. D. Hirst, D. M. Hirst, C. L. Brooks, J. Phys. Chem. A 1997, 101 , 4821–4827. 14. N. A. Besley, J. D. Hirst, J. Phys. Chem. A 1998, 102 , 10791–10797. 15. W. Moffitt, J. Chem. Phys. 1956, 25 , 467–478. 16. J. A. Schellman, Acc. Chem. Res. 1968, 1 , 144–151. 17. E. U. Condon, W. Altar, H. Eyring, J. Chem. Phys. 1937, 5 , 753–775. 18. I. Tinoco, Jr., Adv. Chem. Phys. 1962, 4 , 113–160. 19. P. M. Bayley, E. B. Nielsen, J. A. Schellman, J. Phys. Chem. 1969, 73 , 228–243. 20. H. DeVoe, J. Chem. Phys. 1965, 43 , 3199–3208. 21. J. Applequist, J. Chem. Phys. 1979, 71 , 4332–4338. 22. L. Pauling, R. B. Corey, H. R. Branson, Proc. Natl. Acad. Sci. USA 1951, 37 , 205–211. 23. D. J. Barlow, J. M. Thornton, J. Mol. Biol . 1988, 201 , 601–619. 24. G. Holzwarth, P. Doty, W. B. Gratzer, J. Am. Chem. Soc. 1962, 84 , 3194–3196. 25. G. Holzwarth, P. Doty, J. Am. Chem. Soc. 1965, 87 , 218–228. 26. R. W. Woody, J. Polym. Sci., Part D: Macromol. Rev . 1977, 12 , 181–320. 27. J. A. Schellman, P. Oriel, J. Chem. Phys. 1962, 37 , 2114–2124. 28. R. W. Woody, I. Tinoco, Jr., J. Chem. Phys. 1967, 46 , 4927–4945. 29. W. C. Johnson, Jr., I. Tinoco, Jr., J. Am. Chem. Soc. 1972, 94 , 4389–4390. 30. L. Serrano-Andres, M. P. F¨ulscher, J. Am. Chem. Soc. 1998, 120 , 10912–10920. 31. L. Serrano-Andres, M. P. F¨ulscher, J. Phys. Chem. B 2001, 105 , 9323–9330. 32. M. T. Oakley, J. D. Hirst, J. Am. Chem. Soc. 2006, 128 , 12414–12415. 33. B. M. Bulheller, A. J. Miles, B. A. Wallace, J. D. Hirst, J. Phys. Chem. B 2008, 112 , 1866–1874. 34. C. Toniolo, A. Polese, F. Formaggio, M. Crisma, J. Kamphuis, J. Am. Chem. Soc. 1996, 118 , 2744–2745. 35. C. Toniolo, F. Formaggio, S. Tognon, Q. B. Broxterman, B. Kaptein, R. Huang, V. Setnicka, T. A. Keiderling, I. H. McColl, L. Hecht, L. D. Barron, Biopolymers 2004, 75 , 32–45. 36. M. C. Manning, R. W. Woody, Biopolymers 1991, 31 , 569–586.

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

37. 38. 39. 40. 41. 42. 43. 44.

45. 46. 47. 48. 49.

50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61.

62. 63. 64. 65. 66. 67. 68. 69. 70. 71.

T. S. Sudha, E. K. S. Vijayakumar, P. Balaram, Int. J. Pept. Protein Res. 1983, 22 , 464–468. N. J. Greenfield, G. D. Fasman, Biochemistry, 1969, 8 , 4108–4116. S. Brahms, J. Brahms, G. Spach, A. Brack, Proc. Natl. Acad. Sci. USA 1977, 74 , 3208–3212. J. G. Lees, B. A. Wallace, Spectroscopy Int. J ., 2002, 16 , 121–125. J. S. Richardson, E. D. Getzoff, D. C. Richardson, Proc. Natl. Acad. Sci. USA 1978, 75 , 2574–2578. R. W. Woody, Biopolymers 1969, 8 , 669–683. D. E. Kamen, Y. Griko, R. W. Woody, Biochemistry 2000, 39 , 15932–15943. R. W. Woody, Studies of theoretical circular dichroism of polypeptides: Contributions of β turns, in Peptides, Polypeptides, and Proteins, E. R. Blout, F. A. Bovey, M. Goodman, and N. Lotan, eds., New York, John Wiley & Sons, 1974, pp. 338–350. C. M. Venkatachalam, Biopolymers 1968, 6 , 1425–1436. C. A. Bush, S. K. Sarkar, K. D. Kopple, Biochemistry 1978, 17 , 4951–4954. V. S. Ananthanarayanan, N. Shyamasundar, Biochem. Biophys. Res. Commun. 1981, 102 , 295–301. L. M. Gierasch, C. M. Deber, V. Madison, C. H. Niu, E. R. Blout, Biochemistry 1981, 20 , 4730–4738. J. Bandekar, D. J. Evans, S. Krimm, S. J. Leach, S. Lee, J. R. McQuie, E. Minasian, G. Nemethy, M. S. Pottle, H. A. Scheraga, E. R. Stimson, R. W. Woody, Int. J. Pept. Protein Res. 1982, 19 , 187–205. M. L. Tiffany, S. Krimm, Biopolymers 1968, 6 , 1379–1382. M. L. Tiffany, S. Krimm, Biopolymers 1969, 8 , 347–359. R. W. Woody, Adv. Biophys. Chem. 1992, 2 , 37–79. Z. S. Shi, R. W. Woody, N. R. Kallenbach, Adv. Prot. Chem. 2002, 62 , 163–240. R. W. Woody, J. Am. Chem. Soc. 2009, 131 , 8234–8245. J. T. Yang, C. S. C. Wu, H. M. Martinez, Meth. Enzymol . 1986, 130 , 208–269. W. C. Johnson, Annu. Rev. Biophys. Biophys. Chem. 1988, 17 , 145–166. W. C. Johnson, Proteins Struct. Funct. Genet. 1990, 7 , 205–214. N. J. Greenfield, Anal. Biochem. 1996, 235 , 1–10. N. J. Greenfield, Meth. Enzymol . 2004, 383 , 282–317. N. J. Greenfield, Nat. Protoc. 2006, 1 , 2876–2890. S. Y. Venyaminov, Determination of protein secondary structure, in Circular Dichroism and the Conformation of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 69–107. S. Brahms, J. Brahms, J. Mol. Biol . 1980, 138 , 149–178. V. P. Saxena, D. B. Wetlaufer, Proc. Natl. Acad. Sci. USA 1971, 68 , 969–972. Y. H. Chen, J. T. Yang, Biochem. Biophys. Res. Commun, 1971, 44 , 1285–1291. Y. H. Chen, J. T. Yang, H. M. Martinez, Biochemistry 1972, 11 , 4120–4131. I. A. Bolotina, V. O. Chekhov, V. Y. Lugauskas, A. V. Finkelshtein, O. B. Ptitsyn, Mol. Biol . (English translation of Molekul. Biol .) 1980, 14 , 701–709. S. W. Provencher, J. Gl¨ockner, Biochemistry 1981, 20 , 33–37. V. V. Shubin, M. L. Khazin, T. B. Efimovskaya, Mol. Biol . (English translation of Molekul. Biol .) 1990, 24 , 165–176. J. P. Hennessey, W. C. Johnson, Jr., Biochemistry 1981, 20 , 1085–1094. P. Manavalan, W. C. Johnson, Jr., Anal. Biochem. 1987, 167 , 76–85. I. H. M. van Stokkum, H. J. W. Spoelder, M. Bloemendal, R. van Grondelle, F. C. A. Groen, Anal. Biochem. 1990, 191 , 110–118.

493

494

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113.

N. Sreerama, R. W. Woody, Anal. Biochem. 1993, 209 , 32–44. G. Bohm, R. Muhr, R. Jaenicke, Prot. Eng. 1992, 5 , 191–195. M. A. Andrade, P. Chacon, J. J. Merelo, F. Moran, Protein Eng. 1993, 6 , 383–390. N. Sreerama, R. W. Woody, J. Mol. Biol . 1994, 242 , 497–507. P. Unneberg, J. J. Merelo, P. Chacon, F. Moran, Proteins Struct. Funct. Genet. 2001, 42 , 460–470. C. Perez-Iratxeta, M. A. Andrade-Navarro, BMC Struct. Biol . 2008, 8 , 25–29. P. Pancoska, T. A. Keiderling, Biochemistry 1991, 30 , 6885–6895. A. Perczel, M. Hollosi, G. Tusnady, G. D. Fasman, Protein Eng. 1991, 4 , 669–679. J. G. Lees, A. J. Miles, R. W. Janes, B. A. Wallace, BMC Bioinform. 2006, 7 , 507–516. S. Navea, R. Tauler, E. Goormaghtigh, A. de Juan, Proteins Struct. Funct. Bioinform. 2006, 63 , 527–541. W. C. Johnson, Proteins Struct. Funct. Genet. 1999, 35 , 307–312. N. Sreerama, R. W. Woody, Anal. Biochem. 2000, 287 , 252–260. N. Sreerama, S. Y. Venyaminov, R. W. Woody, Anal. Biochem. 2000, 287 , 243–251. N. Sreerama, R. W. Woody, Protein Sci . 2004, 13 , 100–112. K. A. Oberg, J. M. Ruysschaert, E. Goormaghtigh, Protein Sci . 2003, 12 , 2015–2031. W. Kabsch, C. Sander, Biopolymers 1983, 22 , 2577–2637. K. Matsuo, R. Yonehara, K. Gekko, J. Biochem. 2004, 135 , 405–411. K. Matsuo, R. Yonehara, K. Gekko, J. Biochem. 2005, 138 , 79–88. J. G. Lees, A. J. Miles, F. Wien, B. A. Wallace, Bioinformatics 2006, 22 , 1955–1962. A. Lobley, L. Whitmore, B. A. Wallace, Bioinformatics 2002, 18 , 211–212. L. Whitmore, B. A. Wallace, Nucl. Acids Res. 2004, 32 , W668–W673. B. A. Wallace, L. Whitmore, R. W. Janes, Proteins Struct. Funct. Bioinform. 2006, 62 , 1–3. L. Whitmore, R. W. Janes, B. A. Wallace, Chirality 2006, 18 , 426–429. W. H. Press, B. P. Flannery, S. A. Teukolsky, W. T. Vetterling, Numerical Recipes: The Art of Scientific Computing, Cambridge University Press, Cambridge, England, 1989. N. Sreerama, S. Y. Venyaminov, R. W. Woody, Protein Sci . 1999, 8 , 370–380. P. McPhie, Anal. Biochem. 2001, 293 , 109–119. P. McPhie, Biopolymers 2004, 75 , 140–147. P. McPhie, Anal. Biochem. 2008, 375 , 379–381. V. Raussens, J. M. Ruysschaert, E. Goormaghtigh, Anal. Biochem. 2003, 319 , 114–121. V. Raussens, J. M. Ruysschaert, E. Goormaghtigh, Anal. Biochem. 2006, 359 , 150–150. P. Pancoska, V. Janota, T. A. Keiderling, Anal. Biochem. 1999, 267 , 72–83. N. N. Kalnin, I. A. Baikalov, S. Y. Venyaminov, Biopolymers 1990, 30 , 1273–1280. W. K. Surewicz, H. H. Mantsch, D. Chapman, Biochemistry 1993, 32 , 389–394. A. Barth, C. Zscherp, Q. Rev. Biophys. 2002, 35 , 369–430. R. W. Williams, J. Mol. Biol . 1983, 166 , 581–603. R. W. Williams, Meth. Enzymol . 1986, 130 , 311–331. B. M. Bussian, C. Sander, Biochemistry 1989, 28 , 4271–4277. R. W. Sarver, W. C. Krueger, Anal. Biochem. 1991, 199 , 61–67. R. Pribic, Anal. Biochem. 1994, 223 , 26–34. V. Baumruk, P. Pancoska, T. A. Keiderling, J. Mol. Biol . 1996, 259 , 774–791. K. A. Oberg, J. M. Ruysschaert, E. Goormaghtigh, Eur. J. Biochem. 2004, 271 , 2937–2948. E. Goormaghtigh, R. Gasper, A. Benard, A. Goldsztein, V. Raussens, Biochim. Biophys. Acta Proteins Proteom. 2009, 1794 , 1332–1343.

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

114. P. Pancoska, E. Bitto, V. Janota, M. Urbanova, V. P. Gupta, T. A. Keiderling, Protein Sci . 1995, 4 , 1384–1401. 115. E. H. Strickland, CRC Crit. Rev. Biochem. 1974, 2 , 113–175. 116. P. C. Kahn, Meth. Enzymol . 1979, 61 , 339–378. 117. R. W. Woody, A. K. Dunker, Aromatic and cystine side-chain circular dichroism in proteins, in Circular Dichroism and Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 109–157. 118. R. W. Woody, Aromatic side-chain contributions to protein circular dichroism, in Methods in Protein Structure and Stability Analysis, V. N. Uversky and E. Permyakov, eds., Nova Science Publishers, Hauppauge, NY, 2007, pp. 291–344. 119. S. Vuilleumier, J. Sancho, R. Loewenthal, A. R. Fersht, Biochemistry 1993, 32 , 10303–10313. 120. P. O. Freskg˚ard, L. G. M˚artensson, P. Jonasson, B. H. Jonsson, U. Carlsson, Biochemistry 1994, 33 , 14281–14288. 121. E. Ohmae, Y. Sasaki, K. Gekko, J. Biochem. 2001, 130 , 439–447. 122. T. M. Thompson, B. L. Mark, C. W. Gray, T. C. Terwilliger, N. Sreerama, R. W. Woody, D. M. Gray, Biochemistry 1998, 37 , 7463–7477. 123. N. Sreerama, M. C. Manning, M. E. Powers, J. X. Zhang, D. P. Goldenberg, R. W. Woody, Biochemistry 1999, 38 , 10814–10822. 124. D. Andersson, U. Carlsson, P. O. Freskg˚ard, Eur. J. Biochem. 2001, 268 , 1118–1128. 125. A. Y. M. Woody, R. W. Woody, Biopolymers 2003, 72 , 500–513. 126. I. B. Grishina, R. W. Woody, Faraday Disc. 1994, 99 , 245–262. 127. D. M. Rogers, J. D. Hirst, Biochemistry 2004, 43 , 11092–11102. 128. N. M. Green, M. D. Melamed, Biochem. J . 1966, 100 , 614–621. 129. L. A. Day, Biochemistry 1973, 12 , 5329–5339. 130. C. C. Yang, C. C. Chang, K. Hayashi, T. Suzuki, K. Ikeda, K. Hamaguchi, Biochim. Biophys. Acta 1968, 168 , 373–376. 131. R. W. Woody, Biopolymers 1978, 17 , 1451–1467. 132. R. W. Woody, Eur. Biophys. J . 1994, 23 , 253–262. 133. M. Shiraki, Sci. Pap. Coll. Gen. Educ. Univ. Tokyo 1969, 19 , 151–173. 134. S. K. Burley, G. A. Petsko, Science 1985, 229 , 23–28. 135. C. R. Cantor, S. N. Timasheff, Optical spectroscopy of proteins, in The Proteins, 3rd ed., H. Neurath and R. L. Hill, eds., Academic Press, New York, 1982, pp. 145–306. 136. K. Kuwajima, E. P. Garvey, B. E. Finn, C. R. Matthews, S. Sugai, Biochemistry 1991, 30 , 7693–7703. 137. A. Wollmer, Konformationsanalyse von Proteinen mit Hilfe des Circulardichroismus und der optischen Rotationsdispersion. Habilitation thesis, RWTH Aachen, 1972. 138. M. A. Khan, C. Neale, C. Michaux, R. Pomes, G. G. Prive, R. W. Woody, R. E. Bishop, Biochemistry 2007, 46 , 4565–4579. 139. G. E. Arnold, L. A. Day, A. K. Dunker, Biochemistry 1992, 31 , 7948–7956. 140. G. Blauer, Struct. Bonding (Berlin) 1974, 18 , 69–129. 141. N. J. Greenfield, CRC Crit. Rev. Biochem. 1975, 3 , 71–110. 142. F. Zsila, Z. Bikadi, I. Fitos, M. Simonyi, Curr. Drug Discov. Technol . 2004, 1 , 133–153. 143. A. Rodger, R. Marrington, D. Roper, S. Windsor, Meth. Mol. Biol . 2005, 305 , 343–363. 144. B. Honig, T. G. Ebrey, Annu. Rev. Biophys. Bioeng. 1974, 3 , 151–177. 145. G. Blauer, G. Wagniere, J. Am. Chem. Soc. 1975, 97 , 1949–1954. 146. J. S. Wicken, R. W. Woody, Biochemistry 1973, 12 , 3459–3466. 147. J. S. Wicken, R. W. Woody, Arch. Biochem. Biophys. 1974, 162 , 12–16.

495

496

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

148. 149. 150. 151. 152. 153. 154. 155. 156. 157.

158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171.

172. 173. 174. 175.

176. 177. 178. 179. 180.

J. F. Towell, R. W. Woody, Biochemistry 1980, 19 , 4231–4237. K. Kuwajima, Meth. Mol. Biol . 1995, 40 , 115–135. N. J. Greenfield, Nat. Protoc. 2006, 1 , 2891–2899. O. B. Ptitsyn, Adv. Protein Chem. 1995, 47 , 83–229. M. Arai, K. Kuwajima, Adv. Protein Chem. 2000, 53 , 209–282. K. Kuwajima, Y. Hiraoka, M. Ikeguchi, S. Sugai, Biochemistry 1985, 24 , 874–881. M. Arai, K. Kuwajima, Fold. Design 1996, 1 , 275–287. R. W. Woody, Meth. Enzymol . 2004, 380 , 242–285. P. Tompa, Structure and Function of Intrinsically Disordered Proteins, CRC Press, Boca Raton, FL, 2010. R. W. Woody, Circular dichroism of intrinsically disordered proteins, in Instrumental Analysis of Intrinsically Disordered Proteins: Assessing Structure and Conformation, V. N. Uversky and S. Longhi, ed., John Wiley & Sons, Hoboken, NJ, 2009, pp. 303–321. V. N. Uversky, J. R. Gillespie, A. L. Fink, Proteins Struct. Funct. Genet. 2000, 41 , 415–427. M. Kjaergaard, A.-B. Nørholm, R. Hendus-Altenburger, S. F. Pedersen, F. M. Poulsen, B. B. Kragelund, Protein Sci . 2010, 19 , 1555–1564. D. W. Urry, T. H. Ji, Arch. Biochem. Biophys. 1968, 128 , 802–807. B. A. Wallace, J. G. Lees, A. J. W. Orry, A. Lobley, R. W. Janes, Protein Sci . 2003, 12 , 875–884. K. Park, A. Perczel, G. D. Fasman, Prot. Sci . 1992, 1 , 1032–1049. E. Turk, O. K. Gasymov, S. Lanza, J. Horwitz, E. M. Wright, Biochemistry 2006, 45 , 1470–1479. K. Matsuo, H. Namatame, M. Taniguchi, K. Gekko, Biochemistry 2009, 48 , 9103–9111. K. Matsuo, H. Watanabe, K. Gekko, Proteins-Struct. Funct. Bioinform., 2008, 73 , 104–112. P. M. Hwang, W. Y. Choy, E. I. Lo, L. Chen, J. D. Forman-Kay, C. R. H. Raetz, G. G. Prive, R. E. Bishop, L. E. Kay, Proc. Natl. Acad. Sci. USA 2002, 99 , 13560–13565. V. E. Ahn, E. I. Lo, C. K. Engel, L. Chen, P. M. Hwang, L. E. Kay, R. E. Bishop, G. G. Prive, EMBO J . 2004, 23 , 2931–2941. M. A. Khan, J. Moktar, P. J. Mott, R. E. Bishop, Biochemistry 2010, 49 , 2368–2379. Y. P. Myer, Meth. Enzymol . 1978, 54 , 249–284. Y. P. Myer, A. Pande, Circular dichroism studies of hemoproteins and heme models, in The Porphyrins, D. Dolphin, ed., Academic press, New York, 1978, pp. 271–322. R. Woody, Circular dichroism probes of hemoglobin structure, in Biochemical and Clinical Aspects of Hemoglobin Abnormalities, W. S. Caughey, ed., Academic press, New York, 1978, pp. 279–298. M. C. Hsu, R. W. Woody, J. Am. Chem. Soc. 1971, 93 , 3515–3525. J. Fleischhauer, A. Wollmer, Z. Naturforsch. 1972, B 27 , 530–532. W. Strassburger, H. Thiele, W. Steigemann, E. Weber, J. Fleischhauer, A. Wollmer, HoppeSeylers Z. Physiolog. Chem. 1978, 359 , 1155–1156. R. W. Woody, Contributions of aromatic amino acid residues to the optical activity of protein–ligand complexes, in Protein–Ligand Interactions, H. Sund and G. Blauer, eds., Walter de Gruyter, Berlin, 1975, pp. 60–77. T. Dartigalongue, F. Hache, J. Chem. Phys. 2005, 123 , 184901-1–184901-9 J. Applequist, K. R. Sundberg, M. L. Olson, L. C. Weiss, J. Chem. Phys. 1979, 70 , 1240–1246. G. N. Lamar, H. Toi, R. Krishnamoorthi, J. Am. Chem. Soc. 1984, 106 , 6395–6401. H. S. Aojula, M. T. Wilson, A. Drake, Biochem. J . 1986, 237 , 613–616. H. S. Aojula, M. T. Wilson, G. R. Moore, D. J. Williamson, Biochem. J . 1988, 250 , 853–858.

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

181. C. Kiefl, N. Sreerama, R. Haddad, L. S. Sun, W. Jentzen, Y. Lu, Y. Qiu, J. A. Shelnutt, R. W. Woody, J. Am. Chem. Soc. 2002, 124 , 3385–3394. 182. J. D. Hobbs, J. A. Shelnutt, J. Protein Chem. 1995, 14 , 19–25. 183. G. Blauer, N. Sreerama, R. W. Woody, Biochemistry 1993, 32 , 6674–6679. 184. M. F. Perutz, J. E. Ladner, S. R. Simon, C. Ho, Biochemistry 1974, 13 , 2163–2173. 185. Y. Jin, H. Sakurai, Y. Nagai, M. Nagai, Biopolymers 2004, 74 , 60–63. 186. R. Li, Y. Nagai, M. Nagai, J. Inorg. Biochem. 2000, 82 , 93–101. 187. R. Schweitzer-Stenner, A. Hagarman, D. Verbaro, J. B. Soffer, Meth. Enzymol . 2009, 466 , 109–153. 188. L. Whitmore, B. A. Wallace, Biopolymers 2008, 89 , 392–400. 189. G. V. Louie, G. D. Brayer, J. Mol. Biol . 1990, 214 , 527–555. 190. I. Dragomir, A. Hagarman, C. Wallace, R. Schweitzer-Stenner, Biophys. J . 2007, 92 , 989–998. 191. R. Schweitzer-Stenner, J. Phys. Chem. B 2008, 112 , 10358–10366. 192. W. A. Eaton, R. M. Hochstrasser, J. Chem. Phys. 1967, 46 , 2533–2539. 193. J. McKnight, M. R. Cheesman, A. J. Thomson, J. S. Miles, A. W. Munro, Eur. J. Biochem. 1993, 213 , 683–687. 194. R. Schweitzer-Stenner, R. Shah, A. Hagarman, I. Dragomir, J. Phys. Chem. B 2007, 111 , 9603–9607. 195. M. J. Burke, D. C. Pratt, T. R. Faulkner, A. Moscowitz, Exp. Eye Res. 1973, 17 , 557–572. 196. B. Honig, P. Kahn, T. G. Ebrey, Biochemistry 1973, 12 , 1637–1643. 197. H. Kakitani, T. Kakitani, S. Yomosa, J. Phys. Soc. Japan 1977, 42 , 996–1004. 198. V. Buss, K. Kolster, F. Terstegen, R. Vahrenhorst, Angew. Chem. Intl. Ed ., 1998, 37 , 1893–1895. 199. Q. Tan, J. Lou, B. Borhan, E. Karnaukhova, N. Berova, K. Nakanishi, Angew. Chem. Intl. Ed . 1997, 36 , 2089–2093. 200. Y. Fujimoto, N. Fishkin, G. Pescitelli, J. Decatur, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 2002, 124 , 7294–7302. 201. A. S. Waggoner, L. Stryer, Biochemistry 1971, 10 , 3250–3254. 202. A. Kropf, B. P. Whittenberger, S. P. Goff, A. S. Waggoner, Exp. Eye Res. 1973, 17 , 591–606. 203. E. M. Johnston, R. Zand, Biochemistry 1973, 12 , 4637–4643. 204. G. Pescitelli, N. Sreerama, P. Salvadori, K. Nakanishi, N. Berova, R. W. Woody, J. Am. Chem. Soc. 2008, 130 , 6170–6181. 205. S. M. King, W. C. Johnson, Proteins Struct. Funct. Genet. 1999, 35 , 313–320. 206. G. Deleage, C. Geourjon, Comput. Appl. Biosci . 1993, 9 , 197–199. 207. B. M. Bulheller, J. D. Hirst, Bioinformatics 2009, 25 , 539–540. 208. G. Kurapkat, P. Kruger, A. Wollmer, J. Fleischhauer, B. Kramer, E. Zobel, A. Koslowski, H. Botterweck, R. W. Woody, Biopolymers 1997, 41 , 267–287. 209. R. W. Woody, N. Sreerama, J. Chem. Phys. 1999, 111 , 2844–2845. 210. V. N. Uversky, J. Li, A. L. Fink, J. Biol. Chem. 2001, 276 , 10737–10744.

497

15 ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES Claudio Toniolo, Fernando Formaggio, and Robert W. Woody

15.1. INTRODUCTION Peptides are of great importance in biology for their many activities: hormones, intracellular signals, neurotransmitters, immunogens, antibiotics, toxins, and so on. The biological significance of peptides makes them important for the pharmaceutical industry and biotechnology. Isolation and synthesis of natural peptides and generation of a wide range of analogues are increasingly relevant activities in these fields. Knowledge of the conformation of a peptide is essential for understanding its mechanism of action. Electronic circular dichroism (hereafter the common abbreviation CD will be used for this spectroscopy) has been an invaluable tool for the conformational analysis of peptides since the introduction of commercial instrumentation in the early 1960s, replacing its dispersive counterpart, optical rotatory dispersion. The unique stereochemical sensitivity of a chirospectroscopic method such as CD guarantees it an important role in solution studies, along with NMR, IR absorption, and Raman. Over the past two decades, CD has been joined by two methods that combine the advantages of chiral sensitivity with the richness of vibrational spectroscopy: vibrational circular dichroism and Raman optical activity. The applications of these methods to peptide systems are described in Chapters 22 and 23 of this volume, respectively. Readers of this chapter may also find the chapters on the CD of peptidomimetics (Chapter 16, this volume), the CD of proteins (Chapter 14, this volume), and the independent systems theory of CD (Chapter 20, Volume 1) of interest. Of the many general reviews of the applications of CD to protein and peptide systems, those of Sears and Beychok [1], Johnson [2, 3], Woody [4] and Kelly et al. [5] may be found especially useful. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

499

500

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

15.2. α-HELIX The α-helix is well defined, since it corresponds to a narrow region of the Ramachandran (φ, ψ) map [6], and it is the most thoroughly investigated type of polypeptide secondary structure. It was first proposed by Pauling et al. [7] in 1951 and accounts for approximately one-third of the residues in globular proteins [8]. For a detailed discussion of the 3D-structural parameters of this helix, the reader should refer to Section 15.3 dealing with the closely related 310 -helix. The α-helix has an intense and characteristic CD spectrum [9] (Figure 15.1). The negative band at ∼222 nm is attributable to the n → π ∗ transition of the peptide group [10, 11]. The peptide π → π ∗ transition gives rise to the other two bands, through exciton splitting (see Chapter 4, this volume) [11–13]: the negative band at ∼208 nm (parallel exciton band) and the positive band at ∼190 nm (perpendicular exciton band). The α-helix CD spectrum shows only minor solvent effects. The case of poly(lAla) in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) is an apparent exception [14]. In HFIP, poly(l-Ala) has a CD spectrum that is reduced about twofold in amplitude, maxima that are blue-shifted by several nanometers, and an n → π ∗ band that is diminished to a shoulder, relative to the normal α-helix spectrum. Other helix-forming polypeptides in HFIP exhibit normal CD spectra. Substantial differences are also seen in the UV

80

[θ] x 10–3 (deg x cm2 x dmol–1)

40

0

–40

Figure 15.1. CD spectrum of poly(L-Glu) in 180

200

220 Wavelength (nm)

240

260

aqueous solution (pH 4.3). (Redrawn from reference 9.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

and IR absorption spectra of poly(l-Ala) in HFIP, compared to those of other α-helical polypeptides. Parrish and Blout [14] proposed that poly(l-Ala) in HFIP adopts a modified α-helix conformation in which the peptide planes, nearly parallel to the helix axis in the Pauling and Corey structure [7], are tilted so that the carbonyl oxygens can H-bond to the solvent, while retaining the intramolecular 1 ← 5 H-bonding. The small size of l-Ala makes this conformation, known as the αII conformation [15], feasible, in contrast to other amino acids. Monitoring helix-coil equilibria [16–19] is one of the most common applications of CD, which is the method of choice because of the large difference between the spectrum of the α-helix and that of the unordered conformation (see Section 15.7). For this and other applications, it is important to quantitate α-helix content. Figure 15.2 shows a family of CD spectra generated by melting of α-helix. A sharp isodichroic point near 203 nm is characteristic of helix-coil equilibria and implies a two-state equilibrium. Helix-coil equilibria are commonly monitored by measuring the CD at 222 nm, where the difference between helix and unordered CD is at a maximum and the signal-to-noise ratio is favorable. If only the α-helix and unordered forms are present, as is implied by an isodichroic point, the fraction of helix, fα can be calculated: fα = ([θ ]222,obs − [θ ]222,u )/([θ ]222,α − [θ ]222,u ) where [θ ]222,obs is the observed 222-nm ellipticity, and [θ ]222,α and [θ ]222,u are the ellipticities of the α-helix and unordered forms, respectively. Chen and Yang [20] proposed values of −32, 640 and −2340 deg cm2 dmol−1 for the α-helix and unordered forms, respectively, so fα = −([θ ]222,obs + 2340)/(30,300) This equation is useful for obtaining approximate values of fα , but precise work requires more refined parameters. Section 15.7 should be consulted for a discussion of the work of Park et al. [21] on this topic. For peptides, the dependence of the CD upon helix length becomes significant. Amide groups at the ends of a helix have a different environment and contribute differently to the CD than those in the middle of the helix, giving rise to end effects. Yang and co-workers [22, 23] proposed an empirical equation for the helix-length dependence of CD: [θn ]λ = [θ∞ ]λ (n − kn )/n where [θn ]λ is the mean residue ellipticity at wavelength λ of a helix with n residues; [θ∞ ]λ is the mean residue ellipticity of an infinite helix at the same wavelength; and kn is a constant that corrects for end effects. The constant kn can be thought of as the number of residues effectively missing from the helix. Its value will depend on the CD band studied. Chen et al. [23] showed that this equation fits the theoretical helix-length dependence predicted in two different studies [11, 24]. The helix-length dependence has also been formulated in terms of the number of amides, r, which has a sounder theoretical basis than the number of residues: [θr ]λ = [θ∞ ]λ (r − kr )/r where [θr ]λ is the ellipticity per amide at wavelength λ and kr is the correction for the number of missing amides. For an unblocked peptide, r = n − 1, so kr = kn − 1; for

501

502

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60

[θ] x 10–3 (deg x cm2 x dmol–1)

40

20

0 70°C 30°C 15°C

–20

1°C

Figure 15.2. CD spectra of an Nα -acetylated (Ac) 17-mer peptide amide based on L-Ala, L-Glu,

–40 200

220 Wavelength (nm)

240

and L-Lys in aqueous solution (pH 7) as a function of heating from 1◦ C to 70◦ C. (Redrawn from reference 18.)

an N α -acyl peptide or a C α -amidated peptide, r = n and kr = kn ; for an N α -acyl, C α amidated peptide, r = n + 1, and kr = kn + 1. Gans et al. [25] used the latter formulation to fit the theoretical CD results of Manning and Woody [26] and experimental data for peptides with r = 14–22 under conditions where they were expected to be nearly 100% helical. The experimental data gave kr = 4.7 and [θ∞ ]222 = −40,000 deg cm2 dmol−1 . This value of [θ∞ ]222 is consistent with the data for synthetic polypeptides in a variety of solvents [27]. The CD of the α-helix also depends on temperature, although not so strongly as that of the unordered conformation. Baldwin and co-workers [18, 28] reported a linear temperature dependence of [θ∞ ]222 = −42,500 + 125t deg cm2 dmol−1 (t = ◦ C) for lAla-rich peptides, assuming kr = 3. The values of the slope and intercept vary somewhat with the peptide sequence. CD, along with 220 MHz (in 1969!) 1 H NMR, was instrumental in detecting the onset of the helix conformation at the heptamer level in terminally-protected l-Glu(OEt) (where OEt is ethoxy) homo-oligopeptides in structure-supporting organic solvents [29]. Remarkably, in this work TFE (2,2,2-trifluoroethanol), a well-known UV-transparent and helicogenic solvent, was used in CD for the first time. Following the same approach (CD technique, TFE as the solvent), high amounts of α-helical conformation were reported

503

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

11

12–15

60 10 9

[θ] x 10–3 (deg x cm2 x dmol–1)

40

8 20 7 6 1 0 5 6 –20

15

–40 180

200

220 Wavelength (nm)

240

Figure 15.3. CD spectra of Boc-(L-Met)n -NHPEG (n = 1–15 ; Boc, tert-butyloxycarbonyl) in TFE solution. (Redrawn from reference 30.)

for monodisperse (l-Met)n (Figure 15.3) and [l- Lys(Z)]n , where Z is benzyloxycarbonyl, homo-oligopeptides to the 15-mer level [30, 31]. These two peptide series were solubilized by covalently linking a polyethylene glycol (PEG) chain to their C-terminus. The threshold for helix stability results from the entropic barrier for helix nucleation [32], which can be overcome by providing a nucleus to initiate the helix. Kemp and co-workers [33–35] designed a series of nuclei, RHelOH, where HelOH = (1S , 4S , 7S , 10S )-2-oxo-9-thia-3,12-diazatricyclo[8.2.1.03.7 ]tridecane-4carboxylic acid and R = Ac, Acβ-l-Asp. These conformationally restricted templates can adopt a conformation in which three carbonyl groups are oriented like the three carbonyls at the N-terminus of an α-helix and thus provide a nucleus for helix formation. A CD study [36] of l-Ala–rich peptides of sequence AcHel–(l-Ala–l-Ala–l-Ala–lAla–l-Lys)n –l-Ala–l-Ala–NH2 (n = 1–5) gave unusually large [θ ]222 values: −50,00 deg cm2 dmol−1 at 0◦ C in 16 mol% ethylene glycol/water, pH 1, 0.2 M NaClO4 , and −60,000 cm2 dmol−1 at −20◦ C. The 208-nm band is not enhanced under these conditions (data are not available for the 190-nm band). A different kind of helix nucleus was introduced by Bierzynski and co-workers [37], who found that when the Ca2+ -binding loop of calmodulin binds La3+ ions, the four peptide groups at the C-terminus adopt an α-helical conformation, and additional residues extend this helix. This finding permitted the measurement of the CD of α-helices

504

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

with 4, 8, and 11 peptides [38]. The important result is that the α-helix CD pattern persists even when the helix contains only four amides, a single turn, although the amplitude is diminished and the shape is altered. Earlier theoretical studies [11, 13, 24, 26] predicted that the 208-nm negative band would be detectable only for helices of at least 10 residues. However, calculations [38, 39] using the experimental π → π ∗ transition dipole moment direction [40] predict a 208-nm band discernable even at the dipeptide level. The induction or augmentation of α-helix can also be accomplished by a strongly helicogenic Cα -tetrasubstituted α-amino acid placed at the N-terminus [41]. According to a CD analysis, Ac-(S )-Iva (Iva, isovaline) is able to increase the α-helix content of an already partially folded 13-mer peptide by about 7% in aqueous solution. Significantly, here the intensity of the 222-nm band is not unusual; that is, it is lower than that at 206 nm. Kemp and co-workers [42–44] discovered that some l-Ala-rich peptides show anomalous temperature dependence of the n → π ∗ CD at low temperatures. In such peptides, [θ ]222 values as large as −65,000 deg cm2 dmol−1 have been observed at −50◦ C [42]. Peptides exhibiting this behavior can be recognized by a dual-wavelength parametric test in which [θ ]222 is plotted against [θ ]208 for a family of spectra generated, for example, by varying temperature. If the shape of the spectrum of the helix is independent of temperature, this plot will be a straight line with a slope of 1.1–1.3. For such peptides, helix content can be analyzed using a limiting ellipticity at 222 nm of −40,000 deg cm2 dmol−1 as described above. However, for a number of l-Ala-rich peptides studied by Kemp’s group, the dual wavelength plot curves upward, corresponding to enhanced [θ ]222 relative to [θ ]208 as temperature decreases. The temperature and length dependence of the CD of such systems has been characterized [44]. What is the origin of the remarkable enhancement of the n → π ∗ band in these peptides? This anomalously strong CD reflects a conformational modification of the αhelix, perhaps a conformation more like the Pauling and co-workers [7] helix, which is predicted [9–11] to have a more intense n → π ∗ band than the Barlow and Thornton [8] helix characteristic of proteins. Lending some credence to this is the observation [43] of an Hα –HN coupling constant that indicates φ = −50◦ , close to that for the Pauling and co-workers helix (−47◦ ) and differing from that for the Barlow and Thornton helix (−62◦ ). Woody [45] suggested that the high amplitude might arise from a strong electrostatic mixing of the n → π ∗ and π → π ∗ transitions in peptide groups perturbed by the proximity of a charged side-chain l-Lys ammonium group [46]. However, the observation of an intensified n → π ∗ band in pure (l-Ala)n sequences [43], as well as a natural protein [47] and other peptides [48], excludes this explanation. Dang and Hirst [49] have suggested a shortening of the H-bonds in the helix caused by the low temperatures and the co-solvents. They calculated a [θ ]max of ∼ − 40, 000 deg cm2 ˚ increasing in magnitude to −50, 000 deg dmol−1 for the normal H-bond length of 3.0 A, ˚ cm2 dmol−1 for an H-bond length of 2.7 A.

15.3. 310 -HELIX The 310 -helix (or, more properly, the 3.010 -helix) (Figure 15.4b) is an alternative helical conformation that has been observed in many peptide and protein structures [50–56]. It is more tightly wound than the α-helix (or 3.613 -helix) (Figure 15.4a). The backbone torsion angles of the right-handed 310 -helix (ideally φ = −60◦ , ψ = −30◦ ) are within

505

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

(a)

(b)

1 ← 5 trans, trans

1← 4 trans (III)

3.613-helix (α-helix)

3.010-helix

Figure 15.4. (a) A 10-residue, right-handed 3.613 -helix (α-helix) and its building block, one of the 1 ← 5 trans, trans intramolecularly H-bonded peptide conformations (also termed α-bend or C13 conformation). (b) A 10-residue, right-handed 3.010 -helix and its building block, one of the 1 ← 4 trans intramolecularly H-bonded peptide conformations (also termed type-III β-turn or C10 conformation). Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white.

the same region of the conformational map as those of the α-helix (ideally φ = −55◦ , ψ = −45◦ ). However, the intramolecular C=O · · · H–N H-bonding schemes are significantly different in the two helices, being of the 1 ← 4 type (trans C10 -form or type-III β-turn) in the 310 -helix, while of the 1 ← 5 type (trans –trans C13 -form or α-bend) in the α-helix. A long polypeptide 310 -helix formed by Cα -monosubstituted α-amino acid residues is less stable than the α-helix, its van der Waals energy is less favorable (it has several close, although not forbidden, short contacts), and the H-bond geometry is not optimal. Thus, for many years it was considered unlikely that long stretches of 310 -helix would be observed. However, there is no disallowed region of the conformational (φ, ψ) space separating these two regularly folded secondary structures. Therefore, the α-helix may be gradually transformed into a 310 -helix (and vice versa) maintaining a near-helical conformation of the chain throughout. Furthermore, if one of the conformations should turn out to be impossible (say, as a result of side-chain interactions), the main chain may slip into the other conformation. In fact, the 310 -helix appears to derive its relevance mainly from its proximity in the conformational energy map to the more stable α-helix. Thus, the role of the 310 -helix as an important intermediate in the mechanism of folding of α-helical proteins may be envisaged [54]. According to a statistical analysis from X-ray diffraction structures of globular proteins [8], the majority of 310 -helices are short (the mean length is 3.3 residues, that is, approximately one turn of helix) and a significant percentage of 310 -helices occur as an N- or C-terminal extension to an α-helix. However, a few relatively long

506

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

NH HN

H3C H2C

H3C H3C

CH3

H2C

CH3

HN

CO

HN

CO

Aib

(S)-Iva

H2C

CH3 CH3

HN CO (S)-(αMe)Nva

H3C HC

N H

CH3

HN CO (S)-(αMe)Val

HN

CO

(S)-ATANP

Figure 15.5. Chemical formulae of the C α -methylated α-amino acid residues Aib, (S)-Iva, (S)-(αMe)Nva, (S)-(αMe)Val, and (S)-ATANP discussed in this section.

(7–12 residues) 310 -helical segments were also detected [57]. Conformational energy calculations demonstrated that Aib (α-aminoisobutyric acid or C α,α -dimethylglycine, Figure 15.5), the prototype of achiral, C α -tetrasubstituted α-amino acids, can effectively promote the onset of 310 -helices in short (or relatively short) peptides due to steric interactions involving the gem-dimethyl groups linked to the α-carbon (Thorpe–Ingold effect) [50–53, 58]. Since 1978, by taking advantage of the incredibly high crystallinity of peptides rich in Cα -tetrasubstituted residues, hundreds of X-ray diffraction structures of 310 -helical peptides were solved. The two longest 310 -helical sequences known to date are the homo-deca- and undecapeptides -(Aib)10,11 -. The minimal main-chain length required for a peptide heavily based on Aib residues to form an α-helix in the crystal state corresponds to seven residues (interestingly, about 13 amino acids are required to induce an α-helix in the solid state for a peptide with protein residues only). By contrast, there is no critical main-chain length dependence for 310 -helix formation; that is, incipient 310 -helices are formed at the lowest possible level (an N α -acylated tripeptide). An N α - and C-blocked -(Aib–l-Ala)3 - peptide gives a regular 310 -helix, but an -(Aib–lAla)4 - peptide gives a predominant α-helix. In peptides of eight or more residues, the α-helix is preferred over the 310 -helix if the percentage of Aib residues does not exceed 50%. However, one or two 310 -helical residues may be observed at either end of the αhelical stretch (the short bits of 310 -helix tighten up the ends of the α-helix by moving the related peptide groups nearer to the axis). This conformational preference was also found in many members of the family of peptaibiotics (membrane-active, Aib-rich, naturally occurring peptides exhibiting antibiotic activity) [59]. The average number of α-helical residues in undeca- and longer peptides is seven (two turns). The average parameters for 310 - and α-helices observed in studies at atomic resolution are listed in Table 15.1 [50]. The number of residues per turn (3.24) is intermediate between those of the theoretical 310 -helix and the theoretical/experimental α-(3.613 -)helix. In a perfect 310 -helix,

TAB L E 15.1. Average Parameters for Right-Handed 310 - and α-Helices [50] Parameter ϕ ψ N· · ·O=C H-bond angle Rotation (per residue) Axial translation (per residue) Residues per turn Pitch

310 -Helix

α-Helix

−57◦ −30◦ 128◦ 111◦ ˚ 1.94 A 3.24 ˚ 6.29 A

−63◦ −42◦ 156◦ 99◦ ˚ 1.56 A 3.63 ˚ 5.67 A

507

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

the side chains on successive turns are exactly eclipsed since there is an integer number of residues per turn. However, the experimentally observed, slightly fractional number of residues per turn does not line up side chains, thereby inducing a slightly staggered, energetically more favorable, disposition. This property may have some implications if a relatively long, amphiphilic 310 -helix needs to be built. On the contrary, the α-helix, with its largely fractional number of amino acids per turn, requires two turns (seven residues or a “heptad” repeat) to position two side chains exactly one on top of the other on the same helical face. Distinct hydrophobic/hydrophilic faces, in turn, are of paramount importance for a correct construction of peptide α- and 310 -helix coiled coils. Finally, it is particularly worth mentioning that a fully developed, stable 310 -helix in solution requires only about eight Cα -tetrasubstituted α-amino acid residues, but this figure is considerably higher (≈20 amino acids) for an α-helix based on protein amino acids under the same experimental conditions. In 1996, the CD spectrum of Ac-[(S )-(αMe)Val]8 -OtBu [(αMe)Val, C α methylvaline; OtBu, tert-butoxy] was monitored in TFE solution and proposed as the standard pattern exhibited by a right-handed 310 -helix [60, 61] (Figure 15.6). This proposal was substantiated by independent X-ray diffraction and 2D-NMR analyses. In particular, this spectrum is characterized by a negative Cotton effect at 207 nm accompanied by a negative shoulder centered near 222 nm. The ellipticity ratio R = [θ ]222 /[θ ]207 is 0.4. It was gratifying to note that these experimental findings were

2

[θ]R x 10–3 (deg x cm2 x dmol–1)

0

–2

–4

–6

–8

Figure 15.6. CD spectrum of the 310 -helical –10 180

200

220 Wavelength (nm)

240

260

homo-octapeptide Ac–[(S)-(αMe)Val]8 –OtBu in TFE solution. (Redrawn from references 60 and 61.)

508

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40

[θ] x 10–3 (deg x cm2 x dmol–1)

20

0

–20

–40

Figure 15.7. Calculated CD spectrum for a 160

180

200

220

Wavelength (nm)

240

260

48-residue 310 -helix peptide with ϕ = −60◦ ,

ψ = −30◦ . (Redrawn from reference 26.)

in good agreement with the theoretical curves published a few years before by Manning and Woody [26] (Figure 15.7). Additional CD calculations of 310 -helix were reported by Bode and Applequist [62], but their method does not treat the n → π ∗ transition. Moreover, the ellipticity at about 195 nm is positive, albeit only slightly. It is relevant to stress the point that the overall shape of this spectrum closely resembles that of an unordered peptide. However, the two spectra can be readily distinguished from the position of the π → π ∗ band (well above 200 nm for the 310 -helix). Whether the α- and 310 -helices can be distinguished by CD has been a long-standing question [11]. In this connection, it should be noted that although the R value can be useful to discriminate between a 310 -helix and an α-helix (R ≈ 1) when used in combination with data from other physicochemical techniques, an uncritical interpretation based solely on this CD parameter as a diagnostic can be risky. Indeed, the R value can be affected by the chemical structure (e.g., C α -methylation) of the peptide building blocks, by environmental (solvent, temperature) changes, and 3D-structural factors, such as incomplete α-helix formation, coexistence of diastereoisomeric (right- and left-handed) helices, and intramolecular C=O· · ·H–N H-bond lengths (including those of bifurcated H-bonds) [49, 63–66]. Finally, it was emphasized that the 222 nm n → π ∗ CD band may deserve a theoretical reexamination [43].

509

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

40

[θ]T x 10–3 (deg x cm2 x dmol–1)

20

0 MeOH 0

–20

MeOH 1

–40

Figure 15.8. CD spectra of the

HFIP 0

homo-heptapeptide alkylamide Ac-[(S)-(αMe)Val]7 -NHiPr, where NHiPr is

HFIP 1

isopropylamino, in methanol (MeOH) and HFIP solutions. Repeated cycles of 310 -helix/α-helix conversion were performed, the order of

HFIP 2 –60 200

220 Wavelength (nm)

240

solvent switches being HFIP 0, MeOH 0, HFIP 1, MeOH 1, HFIP 2. (Redrawn from reference 67.)

The difference in length between the more elongated peptide 310 -helix and the ˚ more compact α-helix is about 0.4 A/residue. This property makes the 310 -/α-helix reversible conversion very promising as a molecular switching tool between the N- and C-terminal functions of a peptide backbone. Using homo-peptides of various main-chain length (for the N α -acetyl, isopropyl amide homo-heptamer, see Figure 15.8), all based on the strongly helicogenic, C α -tetrasubstituted α-amino acid (S )-(αMe)Val, two of us (F.F. and C.T.) have shown that a well-defined, solvent-controlled, reversible 310 -/αhelix transition takes place even in a homo-oligomer as short as a terminally blocked hexapeptide [67–69]. Homo-peptide sequences blocked as a urethane or an acetamide at the N-terminus and as a methyl ester or an N -alkylamide at the C-terminus are all appropriate. Analogous molecular spring properties of related Aib/(αMe)Val peptides were also assessed by CD [70–73]. The next step of the research in this area was the first CD characterization of a water-soluble 310 -helical peptide [74–78]. To this goal, two terminally blocked heptapeptides were prepared, each containing: (1) two residues of ATANP, 2-amino-3[1(1,4,7-triazacyclononane)]-(S )-propanoic acid (Figure 15.5), a chiral Cα -trisubstituted α-amino acid with excellent water-solubilizing properties; and (2) five residues of

510

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Cα -tetrasubstituted α-amino acids, either the achiral Aib or the chiral (S )-Iva (Figure 15.5) and (S )-(αMe)Nva [(αMe)Nva, C α -methyl norvaline; Figure 15.5]. The CD patterns of the Aib- and Iva-containing peptides 1 and 2, respectively, in water are similar (Figure 15.9) and are characterized by a negative band at 202 nm and a negative shoulder at ∼222 nm. In both cases the R ratio is 0.20. Not surprisingly, the negative ellipticity values of peptide 2 almost double those of peptide 1. We believe that this observation is strictly related to the presence in peptide 2 of five chiral (S )-Iva residues as opposed to five achiral Aib residues in peptide 1. More specifically, it is our contention that this experimental finding is mainly associated with the concomitant occurrence of a non-negligible amount of left-handed 310 -helix in the conformational equilibrium mixtures of peptide 1, characterized by less than 30% of chiral (S ) residues. Indeed, there is no evidence in the literature for a different ability of the structurally related Aib and Iva residues to support a helical conformation. If the CD spectrum of peptide 2 in water is compared with that of the (αMe)Val homo-octapeptide in TFE solution, then: (i) The shapes of the curves are very close, despite an ∼5-nm blue shift for peptide 2 in water. (ii) Although both R values would be in the range expected for a 310 -helix, that of peptide 2 is appreciably reduced as a consequence of a concomitant higher intensity of the negative π → π ∗ band and a lower intensity of the negative n → π ∗ band. Thus, on the basis of this chirospectroscopic analysis, we are confident that both peptides 1 and 2 would be folded in the 310 -helical structure in water. The CD curves of the two peptides are only marginally modified by variations of pH, temperature, or concentration. Finally, it is worth mentioning that recently Basu and co-workers [66] have discussed the CD pattern of a partially 310 -helical, water-soluble, terminally blocked Aib–l-Ala–l-Lys 12-mer peptide. In a sequential peptide the alternation of an l-Pro residue, which disrupts the conventional H-bonding schemes observed in helices (lacking the N–H donor group), and a helix-forming residue such as Aib may give rise to a novel helical structure, called the β-bend ribbon spiral [79, 80]. This structure may be considered a variant of the 310 -helix, having approximately the same helical fold of the peptide chain and being stabilized by intramolecular C=O· · ·H–N H-bonds of the β-turn type. The complete characterization of this peptide conformation, which may be of relevance in the development of models for peptaibiotics and for the numerous (l-Pro–X)n (with X = l–Pro) segments found in globular and fibrous proteins, was achieved by X-ray diffraction analyses of terminally blocked (l-Pro–Aib)n sequential peptides (Figure 15.10). The repeating -l-Pro–Aib- dipeptide units show on the average the sequence of backbone torsion angles (φi , ψi and φi +1 , ψi +1 ) − 78◦ , −10◦ and −54◦ , −40◦ . The mean heli˚ and p = 6.29 A. ˚ Typically, the value for the cal parameters are n = 3.43, d = 1.94 A, -l-Pro–Aib- ω torsion angles deviate significantly (| ω| > 10◦ ) from the ideal trans 180◦ value. The required energy for this structural change is partially regained by the formation of acceptable intramolecular C=O· · ·H–N H-bonds. By investigating the terminally blocked -(l-Pro–Aib)4 - octapeptide, we were able to determine the diagnostic CD signature for this unusual ordered peptide conformation [81] (Figure 15.11), which adds physical evidence for the similarity of the β-bend ribbon structure in solution to that of the 310 -helix. The small differences observed between the spectra of these two conformations may be in part attributed to the high percentage (50%) of tertiary amide (Xxx–l-Pro) chromophores present in the former. Also, according to the Xray diffraction analyses, some of the l-Pro–Aib amide bonds in these peptides are significantly nonplanar. This phenomenon can lead to reasonably marked CD consequences, at least in the amide n → π ∗ region, that are independent of inter-residue coupling.

511

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

0

[θ]R x 10–3 (deg x cm2 x dmol–1)

1

–3

–6

2

–9

Figure 15.9. CD spectra of the 310 -helical peptides Ac–Aib–(S)-ATANP–(Aib)2 –(S)-

200

220 Wavelength (nm)

240

ATANP-(Aib)2 -OMe (1), where OMe is methoxy, and Ac-(S)-Iva-(S)-ATANP-[(S)-Iva]2 -(S)-ATANP[(S)-Iva]2 -OMe (2) in water. (Redrawn from reference 75.)

Figure 15.10. X-Ray diffraction structure of the β-bend ribbon spiral forming heptapeptide pBrBz-Aib–(L-Pro–Aib)3 –OMe, where pBrBz is para-bromobenzoyl. The three intramolecular Hbonds of the β-turn (C10 ) type are indicated by dashed lines, O atoms are larger and gray, N and H atoms are white. (Redrawn from reference 80.)

512

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30

[θ]T x 10–3 (deg x cm2 x dmol–1)

0

–30

–60

–90

Figure 15.11. CD spectrum of the β-bend –120 200

220 Wavelength (nm)

240

260

ribbon spiral forming, sequential hexapeptide Ac–(L-Pro-Aib)3 –OMe in TFE solution. (Redrawn from reference 81.)

15.4. β-SHEETS The pleated-sheet β-structure is the second most common ordered secondary structure in peptides and proteins. Both types (parallel- or antiparallel-chain disposition) of βstructures were first proposed by Pauling and Corey in 1951 [82]. The antiparallel-chain β-structure is more widespread than its parallel-chain counterpart. In the former, the directionality of the interstrand H-bonds is optimal, but this can be counterbalanced by side-chain–side-chain interactions and chain topology considerations. In the parallelchain β-structure, the side-chain Cβ atoms of the residues in register and in adjacent ˚ (Figure 15.12). In contrast, strands repeat themselves regularly at a distance of 4.5 A in the antiparallel-chain β-structure these separations strictly alternate between a longer ˚ and a shorter distance (3.5 A). ˚ This short distance in the antiparalleldistance (5.7 A) chain β-structure may be responsible for the statistical observation [83] that sterically hindered amino acids (e.g., the β-branched Val and Ile residues) cannot be easily accommodated and therefore strongly prefer the parallel-chain β-structure. The backbone φ, ψ torsion angles for the two types of pleated-sheet β-structures are partially extended and quite close to each other (φ ≈ −140◦ ; ψ ≈ 135◦ ). The differences of their φ, ψ torsion angles (by 30–45◦ only) from those of the fully extended structure (2.05 -helix;

513

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

NNN

NCN

4.5 Å

4.5 Å

5.7 Å

7.0 Å

3.5 Å

6.9 Å

CCC CN C (a)

(b)

Figure 15.12. Representations of the parallel (a) and antiparallel (b) chain dispositions of the β-sheet structure. Relevant inter- and intrastrand distances are reported.

see Chapter 16, this volume) explain their slightly wavy appearance. Particularly effective β-structure-forming residues are those with sterically demanding, β-branched side chains (Val, Ile, and Thr) and those which can form side-chain to main-chain intra-strand H-bonds (Ser, Thr, Cys) [84]. While Pro is a β-structure breaker, one of the preferred structures for Gly is the antiparallel-chain β-structure. The pleated-sheet β-structure can be either of the intra- or the intermolecular type. The antiparallel-chain, intramolecular type is also termed β-hairpin (two strands and one β-turn). If the strands and turns are multiple, then the resulting structure is termed “cross-β” or “β-meander.” The β-sheet structures may deviate substantially [85] from the nearly planar structures originally proposed [82]. Indeed, both antiparallel- and parallel-chain sheets display an offset of one strand with respect to its neighbor. The overall twist for sheets comprised of l-amino acids is typically left-handed. The β-structure, due to its poor solubility, is responsible for a variety of neurodegenerative “conformational diseases,” such as those characterized by fibril formation and amyloid deposits [86]. This last property of the β-sheet conformation has also made its CD characterization rather difficult. CD spectra for three polypeptides in the β-sheet conformation are shown in Figure 15.13 [87–89]. The β-sheet CD spectrum is characterized by a negative n → π ∗ band near 217 nm and a pair of exciton-split π → π ∗ bands, positive near 195 nm and negative near 175 nm (the latter not shown in Figure 15.13). In contrast to the α-helix, β-sheet CD spectra show much greater variability. This is understandable, in view of the variability of β-sheets, which can be parallel or antiparallel and vary in length and number of strands, as well as in the extent of twist [85]. Is it possible to distinguish parallel and antiparallel β-sheets by CD? It was proposed [90, 91] in the late 1970s that the two types of β-sheet could be distinguished by their crossover wavelength, λco , that is, the wavelength at which the CD passes through zero between the 195-nm positive band and the 175-nm negative band. A λco ∼ 178 nm was reported for films of Boc-(l-Ala)7 -OMe, an antiparallel β-sheet according to IR absorption [92, 93]. For films of Boc-(l-Val)7 -OMe, a parallel β-sheet by IR absorption criteria, λco ∼ 192 nm. Films of Boc–(Xxx)7 –OMe {Xxx = l-Leu, l-Cha (β-cyclohexylalanine), l-Nva (norvaline), l-Nle (norleucine)} exhibited intermediate λco values, suggesting that

514

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40

[θ] x 10–3 (deg x cm2 x dmol–1)

c

a 20

b

0 b

c

Figure 15.13. CD spectra of polypeptide models forming a β-sheet conformation: (a) poly(L-Lys), pH 11.1, 15 min at 52◦ C followed by cooling to 25◦ C; (b) poly(L-Lys) in 1% SDS

a –20 200

220 Wavelength (nm)

240

mixture; (c) copoly(L-Lys– L-Leu) in 0.1 M NaF, pH 7. (Redrawn from reference 45.)

peptides of these amino acids formed mixed β-sheets. However, conformational energy calculations [94, 95] predicted that β-sheets with small, linear side chains, (e.g., l-Ala), are only slightly twisted, whereas sheets with side chains branched at Cβ (e.g., l-Val), have a strong twist. Calculations of the CD for slightly twisted and strongly twisted β-sheets showed that the extent of twist affects the CD spectrum more than the parallel versus antiparallel character. Thus, the λco parameter depends on the degree of twist as well as the relative chain orientation and probably cannot be used as a reliable marker of the latter. In fact, theory indicates that the CD of parallel and antiparallel β-sheets are similar. The protein pectate lyase C (pelC) has more than 30% parallel β-sheet, no antiparallel sheet, and very little α-helix, and the sheet is only slightly twisted [96]. The CD spectrum of pelC [97] strongly resembles that of poly(l-Lys) in the β-form (Figure 15.13) in peak positions and relative amplitudes. Poly(l-Lys) β-sheet is antiparallel by IR absorption criteria and, because of the linear side chains, is expected to have a small twist. These observations support the conclusion that CD differences between parallel and antiparallel β-sheets with comparable twists are minor. There is much interest in peptides that model β-sheets. Intramolecular β-sheets are of greatest interest because β-sheet-to-coil transitions [98] in such peptides can provide thermodynamic parameters describing the effects of amino acid composition and sequence

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

on stability, analogous to the numerous studies of the α-helix. Until recently, β-sheet models have been intermolecular [90, 91, 99, 100], but two-stranded (β-hairpins) and three-stranded β-sheets are now being studied [101]. Evidence for β-hairpins in several model peptides has come from NMR, but the CD spectra show no clear proof for β-structure [102–104]. Contributions of aromatic side chains and β-turns may obscure the β-hairpin signature. Alternatively, NOEs can detect contacts characteristic of the β-hairpin that are present in one part of the molecule while the rest of the molecule is not in the hairpin conformation. By contrast, CD detection requires that all or nearly all of the molecule be in the hairpin. Detection of such “nascent” 3D-structures by NMR and not by CD has been documented for α-helices [105]. Peptides expected to form β-hairpins that show distinct negative bands in the 215 to 220-nm region have been reported [102, 106, 107]. For such peptides, the content of hairpins has been estimated by both NMR and CD, with reasonable agreement. A model hexadecapeptide was reported to be 55% β-hairpin (CD), 41–47% (NOE intensities), and 47% (chemical shifts) [107]. A peptide designed [108] to form a three-stranded β-sheet (betanova) has been investigated intensively. NMR supported the presence of the β-sheet, and it was claimed that CD also corroborated the β-hairpin structure, although no spectra were published. CD at 217 nm and fluorescence indicated a cooperative thermal transition, which was interpreted as a two-state transition. The UV resonance Raman spectrum of betanova was consistent with β-structure but showed no indication of a cooperative thermal transition [109]. Boyden and Asher [109] presented the CD spectrum, which was consistent with an unordered conformation, rather than β-structure, even at 0◦ C. This finding was rationalized [108, 109] as resulting from aromatic side-chain contributions (betanova has one Tyr and one Trp out of 20 residues), but this explanation seems inconsistent with resonance Raman evidence for a molten globule [110] in which the side chains are not in a well-defined conformation. The Serrano group [111] reexamined betanova and, using NMR chemical shifts rather than NOEs, revised their estimates of β-sheet content sharply downward, from ∼80% [108] to ∼10%. Keiderling and co-workers [112] reported a detailed study of betanova. Using CD, IR absorption, and F¨orster resonance energy transfer (FRET), they showed that betanova is predominantly unordered, even at 5◦ C. Both CD and IR absorption gave an estimated β-sheet content of 20–26%. FRET between a donor and acceptor pair at opposite ends ˚ at 5◦ C, which barely increased to 46 of the peptide gave an end-to-end distance of 45 A ◦ ˚ A at 80 C. By contrast, the NMR structure reported by Kortemme et al. [108] gave this ˚ Kuznetsov et al. [112] also studied another three-stranded β-sheet condistance as 21 A. D D struct, P– P, designed by Schenck and Gellman [113] to include two type-II β-turns, stabilized by -d-Pro–Gly- sequences. The original publication showed that the CD spectrum, NOEs, and chemical shifts are all consistent with the targeted structure. Kuznetsov et al. [112] found 42–59% β-sheet by CD and IR absorption. FRET measurements gave ˚ in good agreement with the 30–35 A ˚ estimated from an end-to-end distance of 31 A, structural models of the three-stranded β-sheet. In contrast to the α-helix, there is no generally accepted method to calculate the βsheet content from CD spectra. One suggestion [114] uses the difference [θ ]195 − [θ ]217 taking a value of 50–55 × 103 deg cm2 dmol−1 , derived from the data in Figure 15.13, to represent 100%. As long as the sheet is not strongly twisted, this is a reasonable method, but for Boc–(l-Val)7 –OMe in TFE this difference is ∼104 × 103 deg cm2 dmol−1 [115]. Solubility is a major problem with peptides that form β-sheet. PEGs were attached to the N-terminus of peptides to improve solubility [30,116–119]. Several series of

515

516

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

PEG-based homo-oligopeptides with hydrophobic side chains were examined by CD. The prevailing conformation is the β-sheet structure. However, an interesting finding is that an α-helix to β-sheet transition is observed in the longest (l-Met)n oligomers upon deblocking of the N-terminal protection under acidic conditions [30]. Also, the β-sheet structure of these oligopeptides can be disrupted by incorporation of a guest l-Pro residue [119]. Remarkably, these results, and the analogous ones for oligopeptides not linked to PEG, using Aib, l-Pro, or a d-residue [120], anticipated those recently reported for the important field of amyloid-forming peptides [121–125]. Decapeptides with the sequences H–(l-Val)5 –(l-Ala)2 –(l-Val)3 –NH–PEGM and H–(l-Ala)5 –(l-Val)3 –(l-Ala)2 –NH–PEGM, where PEGM is polyethylene glycol monomethyl ether (Mr ≈ 5 × 103 ) were studied by CD in water and TFE [126]. In both solvents, H–(l-Ala)5 –(l-Val)3 –(l-Ala)2 –NH–PEGM gave CD spectra with a characteristic β-sheet pattern (negative maxima at 216 nm in water and 215 nm in TFE; positive maximum at ∼190 nm in water). By contrast, H–(l-Val)5 –(l-Ala)2 –(lVal)3 –NH–PEGM exhibited a CD spectrum consistent with an unordered conformation in water and a mixture of unordered and α-helix conformations in TFE. The critical main-chain length for forming a β-sheet was extensively investigated. The first monodisperse homo-oligopeptide series examined was (l-Ile)n [99]. In TFE, using CD the critical main-chain length for full development of the β-sheet was found at the heptamer level (Figure 15.14). However, the secondary structure of this oligomer can be disrupted by dilution or addition of a more polar solvent. Subsequently, a large variety of peptide series based on hydrophobic α-amino acids was studied in organic solvents and found to adopt the β-sheet structure [127–129]. Interestingly, the CD spectrum of the β-sheet structure of a terminally-protected, all-d-(Ala)7 -homo-oligomer was found to be the mirror image of that recorded for its enantiomeric peptide [130]. Dilution, heating, and co-solvent addition experiments clearly indicated that the rank order of stability for the homo-heptapeptides examined is l-Ile > l-Val > l-Cys(Me) (S -methyl cysteine) > l-Ala > l-Nva > l-Nle > l-Met [129]. To the same end, peptides of the type H–l-Val–Xxx–(l-Val)3 –NH–PEGM, where Xxx = l-Val, l-Leu, l-Ile [126], and (Xxx–Yyy)n –Zzz–NH–PEGM [131], where Xxx is a hydrophilic amino acid, Yyy is a hydrophobic amino acid, or vice versa, and Zzz is Gly or l-Ser, were prepared and studied. At or above the critical main-chain length, the CD spectra are characteristic of β-sheets. The critical main-chain lengths ranged from 6 to more than 11 residues, depending on the sequence and the solvent. Generally, the bulkier the side chain, the shorter the critical main-chain length. Interestingly, IR absorption criteria indicate that all of the peptides with the host–guest block sequence adopt an antiparallel β-sheet conformation, in contrast to the demonstration [91, 92] that l-Val and l-Ile homo-oligomers assume a parallel β-sheet conformation beyond the critical main-chain length. The preference [126] for antiparallel sheets in the PEGMprotected peptides was attributed to avoidance of steric clashes that would arise between the protecting groups in a parallel β-sheet.

15.5. β-TURNS AND γ-TURNS In 1968 Venkatachalam [132] and Geddes et al. [133], independently, proposed three types (trans I–III) of the 1 ← 4 intramolecularly H-bonded, folded peptide conformation (also called C10 form or β-turn), where there is an H-bond between the C=O group of residue 1 and the N–H group of residue 4 (Figure 15.15). The type-I β-turn has

517

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

40 7 8

30

[θ] x 10–3 (deg x cm2 x dmol–1)

20

10

2 3 4

0

5 6

–10

7 8

–20

Figure 15.14. CD spectra of the Boc–(L-Ile)n –OMe (n = 2–8) homo-oligopeptides in TFE solution (peptide

–30 200

220 Wavelength (nm)

240

concentration: ∼2 mg/ml). (Redrawn from reference 99.)

Figure 15.15. Four types of the 1 ← 4 intramolecularly H-bonded (β-turn) peptide conformation. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white.

518

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

φ2 = −60◦ , ψ2 = −30◦ , and φ3 = −90◦ , ψ3 = 0◦ , while the type-II β-turn has φ2 = −60◦ , ψ2 = 120◦ and φ3 = 80◦ , ψ3 = 0◦ . The two nonhelical turns are interrelated by a rotation of 180◦ at the second peptide moiety. The type-III β-turn has φ2,3 = −60◦ and ψ2,3 = −30◦ , thus forming one turn of the 310 -helix. Obviously, turns of type-I’, II’, and III’ also exist, where the prime superscript indicates that the given turn is the mirror image of the corresponding unprimed one, except of course for the positions of the Cβ and other atoms of the side chains. For the H-bond to remain intact, only certain sidechain groups may be accommodated at Cα (2) and Cα (3). All three types of the trans C10 folded conformation were found in crystals of synthetic and naturally occurring, linear and cyclic oligopeptides according to X-ray diffraction analyses. In globular proteins, the average frequency of trans β-turns is as high as 30%. The most frequently occurring residues in β-turns are l-Pro and l-Ser in the second position, and l-Asn, l-Asp, and Gly in the third position. β-Turns in proteins are often stabilized by antiparallel β-sheets (forming β-hairpins). All three types of β-turns were authenticated in solution as well. An additional type of β-turn (VIa) is that having the central amide group in the cis conformation (ω = 0◦ ) [134] (Figure 15.5). Its characteristic backbone torsion angles are φ2 = −60◦ , ψ2 = 120◦ , and φ3 = −90◦ , ψ3 = 0◦ . Model building studies indicate that, because of the dimension of the pseudo-ring generated by the 1 ← 4 intramolecular Hbond, only the cis peptide structure with both C=O and N–H bonds of the central amide group pointing outwards exists. The occurrence of this structure, although relatively rare, was demonstrated in the crystal state and in solution as well. γ -Turns (also called C7 forms or 1 ← 3 intramolecularly H-bonded, folded peptide conformations) (Figure 15.16) are far less abundant in peptides and proteins than β-turns. In γ -turns, a large change in direction occurs over two amides and a single Cα atom [135, 136]. γ -Turns are pseudo-ring structures that are stabilized by an intramolecular H-bond between the H(3) and O(1) atoms. The trans amide groups lie in two planes, which make an angle of about 115◦ . When R in-NH-CHR-CO- is not an H atom, two different conformers (equatorial and axial) can exist, which are represented on the usual Ramachandran φ, ψ map by two centrosymmetric points, the coordinates of which [for an (l)-residue] are φ = −75◦ , ψ = 65◦ (for the equatorial or inverse γ -turn form) and φ = +75◦ , ψ = −65◦ (for the axial or classical γ -turn form). While the H-bond is strongly bent, it has a normal H· · ·O distance and it still makes a sizable contribution to the stabilization of the folded structure. Actually, some variations of the ω value (| ω| ≈ 10◦ ) are required for the stabilization of the 1 ← 3 intramolecularly H-bonded peptide conformations. However, the small energy of torsion rotation is more than compensated for by the energy of the H-bond [137]. X-ray diffraction analyses and several spectroscopic techniques unambiguously demonstrated the existence of these two types

Figure 15.16. The axial (γ -turn) and equatorial (inverse γ -turn) 1 ← 3 intramolecularly H-bonded peptide conformations. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white.

519

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

of γ -turn conformations. All types of β- and γ -turns were discussed in detail in several review articles [138–142]. Unlike the α-helix and β-sheet, there is no unique CD signature for β-turns because of the range of conformers included in this structural category. The CD was predicted [143] for each of 15 β-turn conformations and variants thereof [132, 134]. Various spectral patterns were predicted, but the predominant one had a weak negative n → π ∗ band at 220–230 nm and had strong π → π ∗ bands, positive at 200–210 nm and negative at 180–190 nm (Figure 15.17). This pattern, called a class B spectrum, resembles that of a β-sheet, but is red-shifted by ∼10 nm. Certain types of β-turns have a relatively high probability of giving other spectral types. For example, some type-II -turns have an α-helix-like (class C) spectrum. Such turns are favored by a heterochiral -d–l- sequence, such as the -d-Phe–l-Pro- sequence in gramicidin S. The prediction that type-II β-turns have a class B spectrum has been verified by CD studies of cyclic peptides with these types of turns [87, 144–146]. However, a cyclic peptide with a type-I turn was observed [146] to have an α-helix-like CD spectrum (class C), rather than the expected [143] class B spectrum. Cyclic peptides with the -d-Xxx–l-Pro- sequence have been found to have a class C spectrum [145–149], in accordance with predictions. Similarly, an -l-Pro–d-Xxx- sequence requires a type-II variant that is predicted to have a relatively high probability of exhibiting a class C

spectrum (a left-handed α-helix spectrum), and such a spectrum is observed for peptides with the -l-Pro-d-Ala- sequence [150].

30

[θ] x 10–3 (deg x cm2 x dmol–1)

20

10

0

–10

–20

–30

Figure 15.17. Calculated CD spectrum for 180

200

220

Wavelength (nm)

240

Venkatachalam’s type-II β-turn (class B spectrum). (Redrawn from reference 143.)

520

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The most definitive study of β-turn CD is that of Bandekar et al. [151], who studied cyclic tripeptides of the type cyclo(l-Ala–Xxx–Aha), where Aha is 6-aminohexanoic acid. With Xxx = l-Ala, a type-I turn is obtained, and the CD spectrum is α-helix-like, in agreement with Gierasch et al. [146]. When Xxx = d-Ala, a type-II turn is required and the CD spectrum is of class B. Thus, of the two most common types of turns, type-I is expected to give a class C spectrum, whereas type-II will show a class B spectrum. The result for the type-II turn agrees with the predictions of Woody [143], but that for the type-I turn does not. In addition, calculations for the low-energy conformers [151] gave results in conflict with experiment. However, Sathyanarayanan and Applequist [152] applied the classical dipole interaction model and obtained qualitative agreement with experiment for these two β-turn models (Figure 15.18). Studies utilizing CD, NMR, and molecular dynamics (MD) simulations have examined 14 cyclic and linear peptides [153–156]. The fractions of type-I and type-II β-turns were determined from measured NOEs and MD simulations. Convex constraint analysis [153] was used to obtain a set of basis CD pectra for type-I, type-II, and type-II β-turns [155, 156], which are shown in Figure 15.19. The CD analysis gave good agreement with the NMR and MD results. Spectra 1 and 4 are both attributed to type-I turns. They have the same sign pattern and are of class C, in accord with expectations [151], but spectrum

[θ] x 10–3 (deg x cm2 x dmol–1)

III

I II

Figure 15.18. Calculated CD spectra for Ac–(L-Ala)2 –NHMe (-NHMe, methylamino) in type-I, II, and III β-turn conformations. It should be noted that, as the atom dipole interaction model of reference 152 does not include the

160

180

200 Wavelength (nm)

220

240

n → π ∗ transition, the 220-nm region is not reproduced reliably. (Redrawn from reference 152.)

521

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ] x 10−3 (deg x cm2 x dmol−1)

40

1

2

20

0 4 –20

–40

Figure 15.19. The four component CD curves for

3 180

200 220 Wavelength (nm)

240

convex constraint analysis of CD spectra from cyclic β-turn model peptides. (Redrawn from reference 155.)

4 differs strongly in relative amplitudes from that of an α-helix. In fact, it resembles the spectrum of a 310 -helix [60, 61, 75]. The authors suggested that spectrum 1 is that of a “strained” type-I turn and spectrum 4 is that of an unstrained turn. However, the type-I spectrum is like those seen with cyclic hexapeptides [146] and tripeptides [151], and at least the hexapeptides do not appear to be strained. It seems more likely that spectrum 1 is that of a canonical type-I turn and spectrum 4 is that of a type-III turn—that is, one turn of a 310 -helix. Spectrum 2 in Figure 15.19 is related to type-II β-turns, is of class C , and differs from the class B spectrum expected for type-II turn in that the n → π ∗ band is weakly positive, rather than negative. Only two peptides in the sample showed predominantly type-II turns, and these have the turns at the -l-Pro-d-Ser- and -l-Val-d-Ser- sequences. The former is analogous to the peptides studied by Ananthanarayanan and Shyamasundar [150], which has a class C spectrum, and the steric bulk of the Val may favor a similar type-II variant for the latter peptide. Thus, these peptides may be type-II by NMR criteria, but differ enough from the canonical type-II turns to exhibit different CD spectra. Component 3 is a class C spectrum and correlates with type-II turns. It is therefore consistent with expectations. Extensive CD studies were performed on selected synthetic sequences of tropoelastin [157–159]. This fibrous protein is known to fold into a spiral structure heavily based on -l-Pro–Gly- type-II β-turns [160]. Not unexpectedly, the spectra resemble those of the type-II turn. The CD curve of a peptide as short as a terminally blocked dipeptide amide, Ac–[d-(αMe)Pro]2 –NHi Pr [(αMe)Pro, C α -methyl proline; NHi Pr, isopropylamide], in acetonitrile solution was reported and proposed as the reference spectrum for a type-III’ β-turn [161] (Figure 15.20). This dipeptide is unique in that it is exclusively based on the conformationally very restricted (αMe)Pro residue. As a result, it is rigidly folded in this structure in the crystal state and 100% folded in this same conformation in CDCl3 , according to an X-ray diffraction, NMR, and FT-IR absorption investigation. The CD properties of γ -turns were not well characterized. In the original definition of the γ -turn [135], the ϕ, ψ angles at the Cα -atom located at the central residue of the turn are (68◦ , −61◦ ), which are near those of the C7 ax conformation—that is, in a region that is unfavorable for an l-amino acid residue. The inverse γ -turn, with

522

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

[θ]T x 10–3 (deg x cm2 x dmol–1)

30

20

10

0

Figure 15.20. CD spectrum of the terminally blocked, type-III β-turn forming,

200

220 Wavelength (nm)

240

Ac[D-(αMe)Pro]2 –NHiPr dipeptide amide in acetonitrile solution. (Redrawn from reference 161.)

the corresponding torsion angles of (−68◦ , 61◦ ), is near the C7 eq conformation and is accessible to all l-amino acid residues. Therefore, the inverse γ -turn is the only form that is likely to be encountered in most peptides, except for a Gly or a d-amino acid central residue. It was reported [162] that the inverse γ -turn is a dominant conformer in Ac-l-Pro-NHMe in carbon tetrachloride and is populated to a significant extent in chloroform. CD measurements, performed in the latter solvent down to 220 nm, revealed a strong, negative band in the n → π ∗ region (Figure 15.21). Thus, an inverse γ -turn is expected to give rise to a strong, negative CD band at 220–230 nm and a γ -turn should produce a large, positive n → π ∗ band. A similar CD curve was calculated for the pentapeptide Ac–(Gly)2 –l-Ala–(Gly)2 –NH2 with the l-Ala residue in the inverse γ -turn conformation [163] (Figure 15.22). For reviews covering in-depth the CD properties of β- and γ -turn peptides, see references 45, 141, 143, and 164–166.

15.6. POLY(L-PRO) HELICES AND COLLAGEN TRIPLE HELIX Amide bonds are usually found in the trans conformation (ω = 180◦ ) in linear peptides, whereas cis amide bonds (ω = 0◦ ) are observed in constrained situations such as those

523

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

0

[θ] x 10–3 (deg x cm2 x dmol–1)

–10

–20

–30

–40

–50

Figure 15.21. CD spectrum of the inverse 200

220

240

Wavelength (nm)

260

γ -turn forming Ac– L-Pro–NHMe in chloroform solution. (Redrawn from reference 162.)

occurring in cyclic peptides, particularly if they are characterized by a small ring size. A trans amide bond in a secondary amide is energetically more stable than a cis amide bond by ∼2 kcal/mol, which explains the overwhelming occurrence in linear peptides of the trans bonding in the crystal state and its large preponderance in solution. However, the energy difference between the two conformations markedly decreases in tertiary amides. Thus, not surprisingly, cis peptide bonds reported in the literature for linear peptides in almost all cases involve a tertiary amide in an -Xxx–Yyy- sequence where Yyy is an l-Pro or an N -alkylated (Sar, l-MeAla, peptoid unit, etc.) residue. Pro–Pro bonds generally adopt the trans conformation, but, in some instances, particularly when the peptide is short or the sequence is syndiotactic (l–d or d–l), a cis conformation does occur [167]. Interestingly, the homopolymer poly(l-Pro) is dimorphic in that, under appropriate experimental conditions, the “all-trans” peptide bond conformation (type-II, PII ) may exhibit a transition (mutarotation) to the “all-cis” peptide bond conformation (type-I, PI ) [168–175] (Figure15.23). The PI and PII helices show quite similar ϕ, ψ values (semi -extended conformations). The ϕ, ψ values are about −65◦ , 155◦ . The classical PII conformation is a threefold helix, which, with appropriate side-chain replacements (e.g., an -OH function at position 4 of the ring), may be endowed with an amphiphilic character. The transition from PI to PII implies a remarkable increase in the long dimension of the 3D-structure (Figure 15.23). As for the ϕ, ψ torsion angles in l-Pro-based peptides, the former is almost invariant (∼ − 65◦ ) owing to the restrictions imposed by the five-membered pyrrolidine ring

524

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4

[θ] x 10–3 (deg x cm2 x dmol–1)

0

–4

–8

Figure 15.22. Calculated CD spectra for 180

200

220

Wavelength (nm)

240

Ac-(Gly)2 -L-Ala-(Gly)2 -NH2 with L-Ala in the inverse γ -turn conformation. (Redrawn from reference 163.)

(Ni to Ci α cyclization). The ψ values accessible to l-Pro residues correspond either to the semi -extended region mentioned above (trans conformation) or to the 310 -/αhelical region (cis conformation). This latter 3D-structure, usually stabilized by 1 ← 4 (or 1 ← 5) intramolecular C=O· · ·H–N H-bonds, cannot be formed by l-Pro residues only. Pro residues, in fact, can exclusively occur at the first two (three) positions of a 310 (α-) helix, respectively, because any following l-Pro residue would act as a helix breaker in that it lacks the H-bonding donor (N–H) fuctionality. The presence of a single l-Pro in the middle of a helix normally induces a “kink” in the structure. In the large majority of published examples, the semi -extended conformation is that adopted by l-Pro, indicating that this residue has an intrinsic propensity to be in the PI or PII structure. This finding is especially verified in the longest homo-oligo(l-Pro)n and is related to unfavorable steric interactions originating between the δ-carbon of an l-Proi residue and the β-carbon of an l-Proi −1 residue if both are folded in an α-/310 -helical conformation. As for the role of the Pro pyrrolidine ring, there is a weak correlation between the type of ring puckering and the backbone φ angle [176, 177], with rings in the up (χ1 < 0◦ ) conformation preferring less negative φ(∼ − 60◦ ) and in the down (χ1 > 0◦ ) conformation preferring more negative φ (∼ − 70◦ ) values. Moreover, the flexibility of the pyrrolidine ring is expected to expand the range of φ, ψ values available to the l-Pro residue, leading to a minimization of unfavorable short-range interactions.

525

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

Figure 15.23. A -(Pro)10 - stretch in the PI poly(L-Pro) (a) and PII poly(L-Pro) (b) conformations. O atoms are larger and gray, N

(a)

(b)

atoms are white.

In summary, the available 3D-structural evidence strongly supports the view that short oligo(l-Pro)n spacers and templates cannot be uncritically viewed as “rigid rods,” as usually considered, but rather that the cis  trans (ω torsion angle) and cis  trans

(ψ torsion angle) equilibria, typical of the Pro residue, may severely hamper reliable conclusions from the experimental data [178–189]. Interestingly, calculations have shown that even long (l-Pro)n peptides may be quite flexible with a defined tendency to fold in a “worm-like” chain characterized by multiple bends [190].

526

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Collagen, the most widespread fibrous protein, is known to be a triple-helical coiled coil in which each of the three strands has a left-handed PII poly(l-Pro) helical conformation. The three strands wrap parallel and in register around a common helical axis forming a right-handed superstructure and are held together by interchain H-bonds. The closepacked nature of the triple helix strictly requires a Gly residue in every third position. These conditions are achieved by the repetitive “consensus” triplet -(Gly-l-Pro-Xxx)n - in which Xxx is any amino acid that can adopt the semi -extended conformation {including an additional l-Pro or an l-Hyp [4(R)-hydroxy-(S )-proline] residue} [191–195]. Due to its elongated nature, no intrachain C=O· · ·H–N H-bonds are possible for this structure, even if secondary amides do occur in the sequence. A stable collagen-type triple helix requires at least five triplets. l-Pro residues can be replaced by acyclic, N -alkylated analogues with the l-configuration [196]. The PI conformation is found only in the solid state and in solvents of low polarity—for example, higher alcohols (1-propanol). PII is the conformation authenticated in water and in halogenated alcohols (e.g., TFE) [197]. The CD spectra of both forms are shown in Figure 15.24. The PII spectrum is very nonconservative and characterized by a weak, positive n → π ∗ band at 226 nm and a strong, negative π → π ∗ band at 206 nm. An additional, much weaker negative band (not shown in Figure 15.24) is observed in the vicinity of 165 nm. 60

I

[θ] × 10–3 (deg × cm2 × dmol–1)

30

0

–30

II

–60

Figure 15.24. CD spectra of PI (I) and PII (II) 180

200

220 Wavelength (nm)

240

poly(L-Pro) in 1-propanol solution and in water, respectively. (Redrawn from reference 197.)

527

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

Since poly(l-Pro) contains tertiary amides, the transition energies for the n → π ∗ and π → π ∗ transitions are shifted to longer wavelengths relative to secondary amides. In general, both transitions are red-shifted by ∼10 nm. Also, the ground-state charge distribution is different for tertiary amides. The major difference is that secondary amides contain a highly polar N–H group, with nitrogen being negatively charged and hydrogen positively charged. In Xxx-Pro tertiary amides the hydrogen is replaced with a methylene group, which carries only a small positive total charge, mirrored by a decrease in the negative charge on the amide nitrogen. Nothing about the PII -helix requires the Pro ring; and, in fact, any of the natural amino acids can fit into the PII -helix. This was recognized with the work of Tiffany and Krimm [198, 199], who pointed out the strong resemblance of the CD spectra of ionized poly(l-Glu) and poly(l-Lys) to that of poly(l-Pro). On this basis, they proposed that ionized poly(l-Glu) and poly(l-Lys), used at the time as models of the unordered conformation, must have significant amounts of the PII conformation. This hypothesis is now supported by a large body of evidence, as discussed in Section 15.7. Short stretches of the PII -helix are found in some globular proteins [173], some of which are completely devoid of Pro. Quantum mechanical calculations have not succeeded in accounting for the CD spectrum of poly(l-Pro) in the PII conformation [200], although the classically based dipole interaction model gives reasonable results [201]. Calculations for poly(l-Ala) in the PII conformation reproduce the weak positive n → π ∗ band and the strong negative band near 200 nm, as shown in Figure 15.25. However, calculations using the PII poly(l-Pro) conformation from X-ray fiber diffraction [202] give poor agreement with experiment (Figure 15.25). It is not clear whether this result is due to differences between solution and solid-state structures (φ,ψ differences and conformational heterogeneity) or to the parameters used to treat the Pro side chains. In contrast to PII poly(l-Pro), PI poly(l-Pro) displays a more conservative CD spectrum [197] (Figure 15.24). It is unusual in that it has a strong, positive band at longer

[θ] × 10–3 (deg × cm2 × dmol–1)

20

10

0

–10

–20

–30 170

180

190

200 210 220 Wavelength (nm)

230

240

250

Figure 15.25. Calculated CD spectra [200] for PII poly(L-Ala)20 with (φ, ψ) = (−60◦ , +160◦ ) ( and the poly(L-Pro)20 (– – –) structure of Sasisekharan [202].

)

528

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

wavelengths (∼215 nm) with a slightly weaker, negative counterpart at shorter wavelengths (∼195 nm). The n → π ∗ transition is presumably responsible for the very weak, negative band near 235 nm. Calculated CD spectra [26] for chains of 10 residues in length agree well with experiments (Figure 15.26). The positions of the two strong features are accurately predicted. However, the calculations underestimate the intensity of the positive 215 nm band and they predict a weak, negative band for the n → π ∗ transition. As opposed to peptides forming α-helices and β-sheets, the CD spectra of (l-Pro)n homo-oligomers are found to exhibit the qualitative features of PII poly(l-Pro)n even for n = 3, although with substantially reduced amplitude and shifted wavelengths [203, 204]. The intrinsic helix-length dependence of the PII -helix CD is not qualitatively different from that of the α-helix but, for the latter, short helices are stable only if nucleated [37, 43]. According to experimental CD analyses, the terminal unprotected (l-Pro)n oligomers show a solvent-dependent conformational change, beginning at the trimer level [205] (Figure 15.27). The intensities of the Cotton effects increase with main-chain length and reach the characteristic values of the polymer when n ≈ 20. Rather surprisingly, the corresponding N-protected oligomers do not show mutarotation on change of the solvent (only the all-trans PII -helix is formed by these peptides). It is also interesting to note that oligo (l-Pro)n peptides, when dissolved in 1-propanol, adopt first a PII -helix, which

[θ] x 10–3 (deg x cm2 x dmol–1)

40

20

0

–20

180

200

220 Wavelength (nm)

240

260

Figure 15.26. Calculated CD spectrum for PI poly(L-Pro)10 . (Redrawn from reference 26.)

529

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

60

37 0

40

[θ] x 10–3 (deg x cm2 x dmol–1)

[θ] x 10–3 (deg x cm2 x dmol–1)

18 12 8 20 5 0

2 6 12

–20

–10

2 3

–20

4 5

–30 15 –40

40 31

37

–50

–40 200

220 Wavelength (nm)

240

180

200 220 Wavelength (nm)

240

Figure 15.27. CD spectra of terminally unprotected homo-oligo (L-Pro)n (n = 2–40): (left) in a 9:1 1-propanol/water mixture (PI helix); (right) in aqueous solution (PII helix). (Redrawn from reference 205.)

slowly interconverts to the PI -helix typically found in this low-polarity solvent [206] (Figure 15.28). Many other valuable and stimulating review articles, as well as theoretical and experimental papers, were published on the CD properties of oligo and poly(l-Pro) helices (including those generated by the related ring-substituted and ring-contracted Pro analogues). A large selection [24, 26, 45, 165, 207–247] is provided in the list of references. Because each of the three strands of the collagen triple helix is based on the PII poly(l-Pro) conformation, it is not surprising that the experimental CD spectrum of this self-associated fibrous protein would be qualitatively similar to that of its monomeric building block [199, 248]. In particular, the CD bands of collagen, though closely related in shape to those of PII poly(l-Pro), have larger intensity values. The experimentally observed overall blue shift of the CD spectrum of collagen relative to that of PII poly(lPro) is paralleled by calculations and depends on the ratio of secondary versus tertiary peptide bonds [24]. In natural collagen, the weak, positive band is seen at 220 nm and the negative band at 197 nm [248] (Figure 15.29). Partially denaturated collagen was found to give CD spectra with lower intensity, red-shifted crossover point, and higher 197 nm/220 nm ellipticity ratio. A ratio of 8.5 between these two bands was recommended as a sensitive measure of the purity of a natural collagen sample. Numerous CD studies of oligomeric and polymeric (Gly-l-Pro–Xxx) peptide triplets, capable of triple helix formation, were performed [196, 242, 249–253]. Figure 15.30 shows the CD spectrum of a polymeric peptide triplet, poly(Gly–l-Pro–l-Nva), in ethylene glycol solution. In monodisperse (Gly–l-Pro–Xxx) oligomers, the stability of the triple helix conformation is governed by peptide main-chain length and concentration, as well as by temperature, the nature of the Nα -blocking group, and the solvent. The central l-Pro of the triplet can

530

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

14 days

10

3 days

[θ] x 10–3 (deg x cm2 x dmol–1)

2 days 1 day

0

–10

2 min

–20

Figure 15.28. Time-dependent evolution of 200

220 Wavelength (nm)

240

the CD spectrum of (L-Pro)13 in 1-propanol. (Redrawn from reference 206.)

be replaced by an acyclic, N -alkylated residue. The ratio of 220 nm/197 nm ellipticities, the inverse of the ratio mentioned above, was found to be useful in establishing the onset of the triple helix in solution. The critical value of 0.15 is typically achieved at room temperature in water for oligopeptides as long as at least five triplets.

15.7. UNORDERED CONFORMATION Most small peptides exist in aqueous solution as ensembles of conformers, rather than having a well-defined conformation. Such peptides are called unordered and have a characteristic CD spectrum [88, 199, 254, 255] with a strong negative band near 197 nm and a weak band approximately at 220 nm (Figure 15.31). The latter band may be a negative shoulder on the short-wavelength band. The distribution of the ensemble in (φ, ψ) space depends upon the peptide sequence, the solvent, and the temperature. The PII -helix (see Section 15.6) has been demonstrated to be an important component of the ensemble in many (perhaps most) peptides [208, 209, 211]. The importance of the PII conformation was first recognized by Tiffany and Krimm [199, 255, 256], who pointed out the remarkable resemblance of the CD spectra of ionized poly(l-Glu) and poly(l-Lys) to the poly(l-Pro) spectrum (Figure 15.32) (the ∼10−nm difference in peak wavelengths results from secondary versus tertiary amides). On the basis of this resemblance, Tiffany and Krimm proposed that “unordered” poly(l-Glu) and poly(l-Lys) must

531

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ] × 10–3 (deg × cm2 × dmol–1)

0

–20

–40

–60

–80

Figure 15.29. CD spectrum of calfskin 180

200

220

Wavelength (nm)

240

collagen (10 mM phosphate buffer, pH 3.5; 15◦ C). (Redrawn from reference 248.)

have a significant amount of local order in the form of short stretches of PII helix, interspersed with less ordered conformations. Drake et al. [257] showed that the CD spectrum of charged poly(l-Lys) exhibits a sharp isodichroic point over a nearly 200◦ temperature range, demonstrating a two-state equilibrium. The spectrum of the low-temperature form is that of a PII helix, whereas that of the high-temperature form has a substantially diminished negative band near 197 nm and a weak negative shoulder near 220 nm. The high-temperature limit must be that of a truly unordered polypeptide with a broad ensemble of conformers from PII , α-helix and β-sheet regions of the Ramachandran map. The vibrational circular dichroism (VCD) spectra of charged poly(l-Glu) and poly(l-Lys) also closely resemble that of poly(l-Pro) [258, 259], providing further convincing proof for Tiffany and Krimm’s proposal. This evidence has been reviewed [208, 209, 211]. Recent studies of small oligopeptides suggest that the conformational distribution in such oligopeptides is much more homogeneous than previously thought. SchweitzerStenner and co-workers [260–265] have used CD in conjunction with IR absorption, polarized Raman, and VCD to analyze the conformation of di- and tripeptides and found sequence-dependent variations in the (φ, ψ) angles of the central residue, but a conformation in the PII region was detected to be prevalent in most tripeptides investigated. These studies also suggest a surprisingly narrow range of conformers [265]. The same conclusion has been achieved for longer l-Ala-rich peptides [266, 267]. For example, a

532

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10

[θ] x 10–3 (deg x cm2 x dmol–1)

0

–10

–20

–30

–40

Figure 15.30. CD spectrum of 200

220 Wavelength (nm)

240

poly(Gly– L-Pro– L-Nva) [average P (triplet) ≈ 120] at 20◦ C in ethylene glycol. (Redrawn from reference 250.)

combination of NMR and molecular dynamics simulations indicate ∼80–85% PII conformation at each of the α-carbons of (l-Ala)n (n = 3 − 7) [267]. A PII -like conformation appears to prevail even at the level of N -acetyl amino acids [268]. These compounds have a very weak and positive n → π ∗ band near 220 nm and a strong negative π → π ∗ band between 195 and 200 nm (Figure 15.33). By contrast, amino acid amides have only a broad and weak positive CD in the 180 to 200-nm region, attributable to a single perturbed amide chromophore. An l-Ala-rich heptadecapeptide with l-Pro at the central residue has been studied by Park et al. [21]. The CD of this peptide is typical of unordered peptides and shows a positive band near 217 nm at low temperatures, indicative of substantial PII content. The temperature-dependence of the CD of this peptide in 8 M guanidinium chloride and the CD data of Drake et al. [257] for poly(l-Lys) were fit to a two-state thermodynamic model with temperature-independent CD for each component and a temperature-independent

H. The fit gave [θ ]222 = +9580 deg cm2 dmol−1 for the low-temperature form, PII , and [θ ]222 = −5560 deg cm2 dmol−1 for the high-temperature form, the truly unordered peptide. These data provide a measure of the PII content of this peptide, which can be applied, with due reservations, to others: fPII = ([θ ]222 + 5560)/15,140

533

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ]T x 10–3 (deg x cm2 x dmol–1)

0

c

–10

b –20

–30

Figure 15.31. CD spectra of poly(L-Lys) in aqueous solution at pH 5.7 (a), the 20-mer S-peptide of ribonuclease A in water at 25◦ C (b), and the Ac-/-NH2 terminally blocked, 11-mer XAO (where X = α, β-diaminobutyric

a –40

200

220

240

Wavelength (nm)

acid or Dap, A = alanine, O = ornithine) in aqueous solution (pH 7.0) at 55◦ C (c). (Redrawn from references 88, 277, and 266, respectively.)

Bienkiewicz et al. [269] used the data of Park et al. [21] for the short-wavelength negative band to provide another method for PII estimation: fPII = ([θ ]200 + 9100)/(−36,600) The CD of unordered peptides generally shows a nearly linear dependence on temperature in the range usually investigated [28, 270]. The slope and intercept of the linear relationship need to be evaluated for each peptide and solvent condition. The implications of the PII component of the unordered peptides for the analysis of helix–coil equilibria were examined by Park et al. [21]. Helix–coil transitions are commonly monitored by CD, generally at 222 nm, and have been assumed to be two-state equilibria between helix and the unordered conformation. The [θ ]222 of the unordered form has frequently been taken to be zero. Park et al. [21] pointed out that we are really dealing with a three-state equilibrium between α-helix, PII -helix, and truly unordered conformations. They proposed that the PII + unordered conformation should be modeled by a peptide with the same sequence as the peptide under investigation, but with the central residue replaced by l-Pro. In the 1970s, the concept of the peptide chromophore parameter was introduced and defined as two amino acid residues, l-Xxx–l-Xxx, forming the peptide bond [271, 272].

534

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10

[θ] x 10−3 (deg x cm2 x dmol−1)

0 –10 –20 –30 –40 –50 –60 150

160

170

180

190

200

210

220

230

240

250

Wavelength (nm)

Figure 15.32. CD spectra of poly(L-Glu) (• • •) at pH 8 (coil) and poly(L-Pro) (o o o). (Redrawn from references 247 and 248, respectively.)

[θ] x 10−3 (deg x cm2 x dmol−1)

10

0

–10

–20

Figure 15.33. CD spectra of L-Ala –30 180

190

200

210 220 230 Wavelength (nm)

240

250

260

derivatives. Ac– L-Ala–OH (— —), H– L-Ala–NH2 (–·–·–), and Ac-L-Ala-NH2 (– – –). (Redrawn from reference 268.)

As a result, in a CD study the total molar ellipticity of an unordered polypeptide chain is represented at any wavelength by the equation [θ ]T = i ni [θ ]R , where ni is the number of peptide chromophores of type i and [θ ]R is the intrinsic (internal or terminal) peptide chromophore ellipticity. To calculate the most relevant l-Xxx–l-Xxx internal [θ ]R values for each homo-dipeptide sequence, the total molar ellipticity values of N- and C- protected homo-tripeptides were subtracted from those of

535

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ]R x 10–3 (deg x cm2 x dmol–1)

0

0 a b

e

c

–10

–10

d

Figure 15.34. CD spectra of the -L-Ala– L-Ala–20

–20

200

220

240

Wavelength (nm)

200

220

240

(a), -L-Leu– L-Leu- (b), -L-Nva– L-Nva- (c) where Nva is norvaline, -L-Val– L-Val- (d), and -L-Ile– L-Ile- (e) internal peptide chromophores in HFIP (1,1,1,3,3,3-hexafluoroisopropanol) solution at room temperature. (Redrawn from reference 272.)

Wavelength (nm)

[θ]R x 10–3 (deg x cm2 x dmol–1)

0

–10

a

–20

b

Figure 15.35. CD spectra of the -L-Ala– L-Alainternal peptide chromophore in water at 65◦ C (a) (redrawn from reference 272); and calculated

–30

200

220 Wavelength (nm)

240

from -(Gly)2 –(L-Ala)2 –(Gly)2 - in water (b) (Redrawn from reference 271.)

the corresponding homo-tetrapeptides. The [θ ]R values thus obtained represent rigorously only those of the corresponding homo-polymers in an unordered conformation. In particular, it was found that: (i) In alcoholic solution the CD spectra of the internal peptide chromophores of aliphatic hydrocarbon-containing peptides, each of the l-configuration, show a weak, negative Cotton effect at about 222 nm accompanied by a significantly more intense, negative Cotton effect just below 200 nm (Figure 15.34). (ii) Within the

536

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ROH family of solvents, the effect of solvent polarity on the internal [θ ]R values is of minor significance. (iii) The corresponding CD spectrum for the l-Ala–l-Ala internal peptide chromophore in aqueous solution at 65◦ C, where the population of higher energy conformers is increased by heating, is similar (Figure 15.35) to that in alcohols. It is worth pointing out that the CD spectra for the l-Lys–l-Lys and l-Glu–l–Glu internal peptide chromophores show a weak, positive dichroism at about 215 nm, but only in the pH region where their side chains are ionized. Not surprisingly, this band is barely discernible in the case of the l-Glu(OMe)–l-Glu(OMe) internal peptide chromophore [273]. It is evident that these data on the internal peptide chromophores fit beautifully with all of the conclusions described above for the CD properties of the unordered peptide conformation. Definitive confirmation of these conclusions, particularly in aqueous solution, can be extracted from the results published by Goodman, Toniolo, Mutter, and their co-workers on the CD spectra of a variety of monodisperse homo-oligopeptide series [29, 99, 116, 120, 274–276] and from the related data on two classical peptides [163, 264, 266, 277–281]. Furthermore, selected stimulating readings on this exciting topic may be found in references 210, 217, 268, and 282–292.

REFERENCES 1. D. W. Sears, S. Beychok, Physical Principles and Techniques of Protein Chemistry, Vol. C, S. J. Leach, ed., Academic Press, New York, 1973, pp. 445–593. 2. W. C. Johnson, Jr., Meth. Biochem. Anal . 1985, 31 , 61–163. 3. W. C. Johnson, Jr., Annu. Rev. Biophys. Chem. 1988, 17 , 145–166. 4. R. W. Woody, Meth. Enzymol . 1995, 246 , 34–71. 5. S. M. Kelly, T. J. Jess, N. C. Price, Biochim. Biophys. Acta-Proteins, Proteom. 2005, 1751 , 119–139. 6. G. N. Ramachandran, V. Sasisekharan, Adv. Protein Chem. 1968, 23 , 283–438. 7. L. Pauling, R. B. Corey, H. R. Branson, Proc. Natl. Acad. Sci. USA 1951, 37 , 205–211. 8. D. J. Barlow, J. M. Thornton, J. Mol. Biol . 1988, 201 , 601–619. 9. G. Holzwarth, P. Doty, J. Am. Chem. Soc. 1965, 87 , 218–228. 10. J. A. Schellman, P. Oriel, J. Chem. Phys. 1962, 37 , 2114–2124. 11. R. W. Woody, I. Tinoco, Jr., J. Chem. Phys. 1967, 46 , 4927–4945. 12. W. J. Moffit, J. Chem. Phys. 1956, 25 , 467–478. 13. I. Tinoco, Jr., R. W. Woody, D. F. Bradley, J. Chem. Phys. 1963, 38 , 1317–1325. 14. J. R. Parrish, Jr., E. R. Blout, Biopolymers 1971, 10 , 1491–1512. 15. G. N´emethy, D. C. Phillips, S. J. Leach, H. A. Scheraga, Nature 1967, 214 , 363–365. 16. J. M. Scholtz, R. L. Baldwin, Annu. Rev. Biophys.Biomol. Struct. 1992, 21 , 95–118. 17. C. A. Rohl, R. L. Baldwin, Biochemistry, 1997, 36 , 8435–8442. 18. S. Marqusee, R. L. Baldwin, Proc. Natl. Acad. Sci. USA 1987, 84 , 8898–8902. 19. N. R. Kallenback, E. J. Spek, Meth. Enzymol . 1998, 295 , 26–41. 20. Y.-H. Chen, J. T. Yang, Biochem. Biophys. Res. Commun. 1971, 44 , 1285–1291. 21. S. H. Park, W. Shalongo, E. Stellwagen, Protein Sci . 1997, 6 , 1694–1700. 22. Y.-H. Chen, J. T. Yang, H. M. Martinez, Biochemistry 1972, 11 , 4120–4131. 23. Y.-H. Chen, J. T. Yang, K. H. Chau, Biochemistry 1974, 13 , 3350–3359. 24. V. Madison, J. Schellman, Biopolymers 1972, 11 , 1041–1076.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

25. P. J. Gans, P. C. Lyu, M. C. Manning, R. W. Woody, N. R. Kallenbach, Biopolymers 1991, 31 , 1605–1614. 26. M. C. Manning, R. W. Woody, Biopolymers 1991, 31 , 569–586. 27. R. W. Woody, J. Polym. Sci., part D: Macromol. Rev . 1977, 12 , 181–320. 28. C. A. Rohl, A. Chakrabartty, R. L. Baldwin, Protein Sci . 1996, 5 , 2623–2637. 29. M. Goodman, A. S. Verdini, C. Toniolo, W. D. Phillips, F. A. Bovey, Proc. Natl. Acad. Sci. USA 1969, 64 , 444–450. 30. C. Toniolo, G. M. Bonora, S. Salardi, M. Mutter, Macromolecules 1979, 12 , 620–625. 31. C. Toniolo, G. M. Bonora, H. Anzinger, M. Mutter, Macromolecules 1983, 16 , 147–149. 32. J. A. Schellman, Compt. Rend. Trav. Lab. Carlsberg, Ser. Chim. 1955, 29 , 230–259. 33. D. S. Kemp, J. H. Rothman, Tetrahedron Lett. 1995, 36 , 4019–4022. 34. D. S. Kemp, J. H. Rothman, Tetrahedron Lett. 1995, 36 , 4023–4026. 35. W. Maison, E. Arce, P. Renold, R. J. Kennedy, D. S. Kemp, J. Am. Chem. Soc. 2001, 123 , 10245–10254. 36. P. Wallimann, R. J. Kennedy, D. S. Kemp, Angew. Chem. Int. Edit. 1999, 38 , 1290–1292. 37. M. Siedlicka, G. Goch, A. Ejchart, H. Sticht, A. Bierzynski, Proc. Natl. Acad. Sci. USA 1999, 96 , 903–908. 38. D.-H. Chin, R. W. Woody, C. A. Rohl, R. L. Baldwin, Proc. Natl. Acad. Sci. USA 2002, 99 , 15416–15421. 39. R. W. Woody, A. Koslowski, Biophys. Chem. 2002, 101 , 535–551. 40. L. B. Clark, J. Am. Chem. Soc. 1995, 117 , 7974–7986. 41. M. Gobbo, L. Biondi, F. Filira, F. Formaggio, M. Crisma, R. Rocchi, C. Toniolo, Q. B. Broxterman, J. Kamphuis, Lett. Pept. Sci . 1998, 5 , 105–107. 42. P. Wallimann, R. J. Kennedy, J. S. Miller, W. Shalongo, D. S. Kemp, J. Am. Chem. Soc. 2003, 125 , 1203–1220. 43. B. Heitmann, G. E. Job, R. J. Kennedy, S. M. Walker, D. S. Kemp, J. Am. Chem. Soc. 2005, 127 , 1690–1704. 44. G. E. Job, R. J. Kennedy, B. Heitmann, J. S. Miller, S. M. Walker, D. S. Kemp, J. Am. Chem. Soc. 2006, 128 , 8227–8233. 45. R. W. Woody, Houben-Weyl: Methods of Organic Chemistry, Vol. E22b, M. Goodman, A. Felix, L. Moroder, C. Toniolo, eds., Thieme, Stuttgart, Germany, 2003, pp. 739–765. 46. K. Groebke, P. Renold, K. Y. Tsang, T. J. Allen, K. F. McClure, D. S. Kemp, Proc. Natl. Acad. Sci. USA 1996, 93 , 4025–4029. 47. C. B. Marshall, A. Chakrabartty, P. L. Davies, J. Biol. Chem. 2005, 280 , 17920–17929. 48. R. S. Farmer, K. L. Kiick, Biomacromolecules 2005, 6 , 1531–1539. 49. Z. Dang, J. D. Hirst, Angew. Chem. Int. Edit. 2001, 40 , 3619–3621. 50. C. Toniolo, E. Benedetti, Trends Biochem. Sci . 1991, 16 , 350–353. 51. C. Toniolo, M. Crisma, F. Formaggio, C. Peggion, Q. B. Broxterman, B. Kaptein, J. Incl. Phenom. Macrocycl. Chem. 2005, 51 , 121–136. 52. M. Crisma, F. Formaggio, A. Moretto, C. Toniolo, Biopolymers (Pept. Sci.) 2006, 84 , 3–12. 53. I. L. Karle, P. Balaram, Biochemistry 1990, 29 , 6747–6756. 54. K. A. Bolin, G. L. Millhauser, Acc. Chem. Res. 1999, 32 , 1027–1033. 55. T. Ohyama, A. Hiroki, K. Kobayashi, R. Katakai, Peptide Chemistry 1996 , C. Kitada, ed., Protein Research Foundation, Osaka, Japan, 1997, pp. 365–368. 56. A. Freiberg, M. P. Machner, W. Pfeil, W.-D. Schubert, D. W. Heinz, R. Seckler, J. Mol. Biol . 2004, 337 , 453–461. 57. A. V. Kajava, B. Kobe, Protein Sci . 2002, 11 , 1082–1090.

537

538

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

58. C. Toniolo, M. Crisma, F. Formaggio, C. Peggion, Biopolymers (Pept. Sci.). 2001, 60 , 396–419. 59. C. Toniolo, H. Br¨uckner, Peptaibiotics: Fungal Peptides Containing α-Dialkyl α-Amino Acids, Wiley-VCH, Weinheim, Germany, 2009. 60. C. Toniolo, A. Polese, F. Formaggio, M. Crisma, J. Kamphuis, J. Am. Chem. Soc. 1996, 118 , 2744–2745. 61. A. Polese, F. Formaggio, M. Crisma, G. Valle, C. Toniolo, G. M. Bonora, Q. B. Broxterman, J. Kamphuis, Chem. Eur. J . 1996, 2 , 1104–1111. 62. K. A. Bode, J. Applequist, J. Phys. Chem. 1996, 100 , 17825–17834. 63. T. S. Sudha, E. K. S. Vijayakumar, P. Balaram, Int. J. Pept. Protein Res. 1983, 22 , 464–468. 64. S. Aravinda, S. Datta, N. Shamala, P. Balaram, Angew. Chem. Int. Edit. 2004, 43 , 6728–6731. 65. N. H. Andersen, Z. Liu, K. S. Prickett, FEBS Lett. 1996, 399 , 47–52. 66. R. Banerjee, S. Chattopadhyay, G. Basu, Proteins 2009, 76 , 184–200. 67. A. Moretto, F. Formaggio, B. Kaptein, Q. B. Broxterman, L. Wu, T. A. Keiderling, C. Toniolo, Biopolymers (Pept. Sci.) 2008, 90 , 567–574. 68. A. Moretto, M. Crisma, F. Formaggio, B. Kaptein, Q. B. Broxterman, T. A. Keiderling, C. Toniolo, Biopolymers (Pept. Sci.) 2007, 88 , 233–238. 69. G. Yoder, A. Polese, R. A. G. D. Silva, F. Formaggio, M. Crisma, Q. B. Broxterman, J. Kamphuis, C. Toniolo, T. A. Keiderling, J. Am. Chem. Soc. 1997, 117 , 10278–10285. 70. F. Donald, G. Hungerford, D. J. S. Birch, B. D. Moore, J. Chem. Soc., Chem. Commun. 1995, 313–314. 71. G. Hungerford, M. Martinez-Insua, D. J. S. Birch, B. D. Moore, Angew. Chem. Int. Edit. Engl . 1996, 35 , 326–329. 72. P. Pengo, L. Pasquato, S. Moro, A. Brigo, F. Fogolari, Q. B. Broxterman, B. Kaptein, P. Scrimin, Angew. Chem. Int. Edit. 2003, 42 , 3388–3392. 73. M. Bellanda, S. Mammi, S. Geremia, N. Demitri, L. Randaccio, Q. B. Broxterman, B. Kaptein, P. Pengo, L. Pasquato, P. Scrimin, Chem. Eur. J . 2007, 13 , 407–416. 74. B. Sacc`a, F. Formaggio, M. Crisma, C. Toniolo, R. Gennaro, Gazz. Chim. Ital . 1997, 127 , 495–500. 75. F. Formaggio, M. Crisma, P. Rossi, P. Scrimin, B. Kaptein, Q. B. Broxterman, C. Toniolo, Chem. Eur. J . 2000, 6 , 4498–4504. ˇ 76. C. Toniolo, F. Formaggio, S. Tognon, Q. B. Broxterman, B. Kaptein, R. Huang, V. Setnicka, T. A. Keiderling, I. H. Mc Coll, L. Hecht, L. D. Barron, Biopolymers 2004, 75 , 32–45. 77. F. Formaggio, C. Peggion, M. Crisma, B. Kaptein, Q. B. Broxterman, J.-P. Mazaleyrat, M. Wakselman, C. Toniolo, Chirality 2004, 16 , 388–397. 78. E. Longo, A. Moretto, F. Formaggio, C. Toniolo, Chirality 2011, 23 , 756–760. 79. I. L. Karle, J. Flippen-Andersen, M. Sukumar, P. Balaram, Proc. Natl. Acad. Sci. USA 1987, 84 , 5087–5091. 80. B. Di Blasio, V. Pavone, M. Saviano, A. Lombardi, F. Nastri, C. Pedone, E. Benedetti, M. Crisma, M. Anzolin, C. Toniolo, J. Am. Chem. Soc. 1992, 114 , 6273–6278. 81. G. Yoder, T. A. Keiderling, F. Formaggio, M. Crisma, C. Toniolo, Biopolymers 1995, 35 , 103–111. 82. L. Pauling, R. B. Corey, Proc. Natl. Acad. Sci. USA 1951, 37 , 729–740. 83. C. Toniolo, Macromolecules 1978, 11 , 437–438. 84. E. R. Blout, Poly-α-Amino Acids, Polypeptides and Proteins, M. A. Stahmann, ed. University of Wisconsin Press, Madison, WI, 1962, pp. 275–279. 85. C. Chothia, J . Mol. Biol . 1973, 75 , 295–302. 86. R. Nelson, D. Eisenberg, Curr. Opin. Struct. Biol . 2006, 16 , 260–265.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121.

S. Brahms, J. Brahms, G. Spach, A. Brack, Proc. Natl. Acad. Sci. USA 1977, 74 , 3208–3212. N. J. Greenfield, G. D. Fasman, Biochemistry 1969, 8 , 4108–4116. L. K. Li, A. Spector, J. Am. Chem. Soc. 1969, 91 , 220–222. J. S. Balcerski, E. S. Pysh, G. M. Bonora, C. Toniolo, J . Am. Chem. Soc. 1976, 98 , 3470–3473. M. M. Kelly, E. S. Pysh, G. M. Bonora, C. Toniolo, J . Am. Chem. Soc. 1977, 99 , 3264–3266. M. Palumbo, S. Da Rin, G. M. Bonora, C. Toniolo, Makromol. Chem. 1976, 177 , 1477–1492. C. Toniolo, M. Palumbo, Biopolymers 1977, 16 , 219–224. K.-C. Chou, M. Pottle, G. N´emethy, Y. Ueda, H. A. Scheraga, J. Mol. Biol . 1982, 162 , 89–112. K.-C. Chou, H. A. Scheraga, Proc. Natl. Acad. Sci. USA 1982, 79 , 7047–7051. M. D. Yoder, N. T. Keen, F. Jurnak, Science 1993, 260 , 1503–1507. D. E. Kamen, Y. Griko, R. W. Woody, Biochemistry 2000, 39 , 15932–15943. L. Tilstra, W. L. Mattice, Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 261–283. M. Goodman, F. Naider, C. Toniolo, Biopolymers 1971, 10 , 1719–1730. C. Toniolo, G. M. Bonora, A. Fontana, Int. J. Pept. Protein Res. 1974, 6 , 283–285. M. Ramirez-Alvarado, T. Kortemme, F. J. Blanco, L. Serrano, Bioorg. Med. Chem. 1999, 7 , 93–103. F. J. Blanco, G. Rivas, L. Serrano, Nat. Struct. Biol . 1994, 1 , 584–590. A. R. Viguera, M. A. Jimenez, M. Rico, L. Serrano, J. Mol. Biol 1996, 255, 507–521. V. Sieber, G. R. Moe, Biochemistry 1996, 35 , 181–188. H. J. Dyson, M. Rance, R. A. Houghten, P. E. Wright, R. A. Lerner, J. Mol. Biol . 1988, 201 , 201–217. M. Ramirez-Alvarado, F. J. Blanco, H. Niemann, L. Serrano, J. Mol. Biol . 1997, 273 , 898–912. A. J. Maynard, G. J. Sharman, M. S. Searle, J. Am. Chem. Soc. 1998, 120 , 1996–2007. T. Kortemme, M. Ramirez-Alvarado, L. Serrano, Science 1998, 281 , 253–256. M. N. Boyden, S. A. Asher, Biochemistry 2001, 40 , 13723–13727. O. B. Ptitsyn, Protein Folding, T. E. Creighton, ed., Freeman, New York, 1992, pp. 243–300. M. L. de la Paz, E. Lacroix, M. Ramirez-Alvarado, L. Serrano, J. Mol. Biol . 2001, 312 , 229–246. S. V. Kuznetsov, J. Hilario, T. A. Keiderling, A. Ansari, Biochemistry 2003, 42 , 4321–4332. H. L. Schenck, S. H. Gellman, J. Am. Chem. Soc. 1998, 120 , 4869–4870. L. Zhong, W. C. Johnson, Jr., Proc. Natl. Acad. Sci. USA 1992, 89 , 4462–4465. C. Toniolo, G. M. Bonora, Makromol. Chem. 1974, 175 , 2203–2207. C. Toniolo, G. M. Bonora, M. Mutter, J. Am. Chem. Soc. 1979, 101 , 450–454. M. Mutter, S. Vuilleumier, Angew. Chem. Int. Edit. Engl . 1989, 28 , 535–554. M. Mutter, F. Maser, K.-H. Altmann, C. Toniolo, G. M. Bonora, Biopolymers 1985, 24 , 1057–1074. C. Toniolo, G. M. Bonora, M. Mutter, V. N. R. Pillai, Makromol. Chem. 1981, 182 , 2007–2014. V. Moretto, M. Crisma, G. M. Bonora, C. Toniolo, H. Balaram, P. Balaram, Macromolecules 1989, 22 , 2939–2944. F. Formaggio, A. Bettio, V. Moretto, M. Crisma, C. Toniolo, Q. B. Broxterman, J. Pept. Sci . 2003, 9 , 461–466.

539

540

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

122. C. Soto, M. S. Kindy, M. Baumann, B. Frangione, Biochem. Biophys. Res. Commun. 1996, 226 , 672–680. 123. E. Hughes, R. M. Burke, A. J. Doig, J. Biol. Chem. 2000, 275 , 25109–25115. 124. D. J. Gordon, K. L. Sciarretta, S. C. Meredith, Biochemistry 2001, 40 , 8237–8245. 125. D. T. S. Rijkers, J. W. M. H¨oppener, G. Posthuma, C. J. M. Lips, R. M. J. Liskamp, Chem. Eur. J . 2002, 8 , 4285–4291. 126. K.-H. Altmann, M. Mutter, Int. J. Pept. Protein Res. 1985, 26 , 373–380. 127. M. Goodman, C. Toniolo, P. Pallai, Forum Peptides, B. Castro, J. Martinez, eds., Dohr, Nancy, France, 1986, pp. 146–174. 128. M. Goodman, C. Toniolo, F. Naider, Peptides, Polypeptides and Proteins, E. R. Blout, F. A. Bovey, M. Goodman, N. Lotan, eds., Wiley, New York, 1974, pp. 308–319. 129. C. Toniolo, G. M. Bonora, Peptides: Chemistry, Structure and Biology, R. Walter, J. Meienhofer, eds., Ann Arbor Science, Ann Arbor, MI, 1975, pp. 145–150. 130. C. Toniolo, M. Palumbo, J. Polym. Sci., Polym. Chem. Edit. 1976, 14 , 515–518. 131. K.-H. Altmann, A. Fl¨orsheimer, M. Mutter, Int. J. Pept. Protein Res. 1986, 27 , 314–319. 132. C. M. Venkatachalam, Biopolymers 1968, 6 , 1425–1436. 133. A. J. Geddes, K. D. Parker, E. D. T. Atkins, E. Brighton, J.Mol. Biol . 1968, 32 , 343–358. 134. P. N. Lewis, F. A. Momany, H. A. Scheraga, Biochim. Biophys. Acta 1973, 303 , 211–229. 135. G. N´emethy, M. P. Printz, Macromolecules 1972, 5 , 755–758. 136. B. W. Matthews, Macromolecules 1972, 5 , 818–819. 137. A. S. Kolaskar, A. V. Lakshminarayanan, K. P. Sarathy, V. Sasisekharan, Biopolymers 1975, 14 , 1081–1094. 138. C. Toniolo, CRC Crit. Rev. Biochem. 1980, 9 , 1–44. 139. G. D. Rose, L. M. Gierasch, J. P. Smith, Adv. Protein Chem. 1985, 37 , 1–109. 140. B. L. Sibanda, J. M. Thornton, J. Mol. Biol . 1993, 229 , 428–447. 141. A. Perczel, M. Hollosi, Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 285–380. 142. K. S. Rotondi, L. M. Gierasch, Biopolymers (Pept. Sci.) 2006, 84 , 13–22. 143. R. W. Woody, Peptides, Polypeptides, and Proteins, E. R. Blout, F. A. Bovey, M. Goodman, N. Lotan, eds., John Wiley & Sons, New York, 1974, pp. 338–350. 144. R. Deslauriers, D. J. Evans, S. J. Leach, Y. C. Meinwald, E. Minasian, G. N´emethy, I. D. Rae, H. A. Scheraga, R. L. Somorjai, E. R. Stimson, J. W. Van Nispen, R. W. Woody, Macromolecules 1981, 14 , 985–996. 145. C. A. Bush, S. K. Sarkar, K. D. Kopple, Biochemistry 1978, 17 , 4951–4954. 146. L. M. Gierasch, C. M. Deber, V. Madison, C. H. Niu, E. R. Blout, Biochemistry 1981, 20 , 4730–4738. 147. S. Laiken, M. Printz, L. C. Craig, J. Biol. Chem. 1969, 244 , 4454–4457. 148. A. C. Gibbs, L. H. Kondejewski, W. Gronwald, A. M. Nip, R. S. Hodges, B. D. Sykes, D. S. Wishart, Nat. Struct. Biol . 1998, 5 , 284–288. 149. S. Lee, H. Mizuno, H. Nakamura, Y. Kodera, T. Kato, N. Go, N. Izumiya, FEBS Lett. 1984, 174 , 310–313. 150. V. Ananthanarayan, N. Shyamasundar, Biochem. Biophys. Res. Commun. 1981, 102 , 295–301. 151. J. Bandekar, D. J. Evans, S. Krimm, S. J. Leach, S. Lee, J. R. McQuie, E. Minasian, G. N´emethy, M. S. Pottle, H. A. Scheraga, E. R. Stimson, R. W. Woody, Int. J. Pept. Protein Res. 1982, 19 , 187–205. 152. B. K. Sathyanarayanan, J. Applequist, Int. J. Pept. Protein Res. 1986, 27 , 86–94. 153. A. Perczel, M. Hollosi, G. Tusnady, G. D. Fasman, Protein Eng. 1991, 4 , 669–679.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

154. A. Perczel, M. Hollosi, B. M. Foxman, G. D. Fasman, J. Am. Chem. Soc. 1991, 113 , 9772–9784. 155. A. Perczel, G. D. Fasman, Protein Sci . 1992, 1 , 378–395. 156. A. Perczel, M. Hollosi, P. Sandor, G. D. Fasman, Int. J. Pept. Protein Res. 1993, 41 , 223–236. 157. A. M. Tamburro, B. Bochicchio, A. Pepe, Biochemistry 2003, 42 , 13347–13362. 158. B. Bochicchio, N. Floquet, A. Pepe, A. J. P. Alix, A. Tamburro, Chem. Eur. J . 2004, 10 , 3166–3176. 159. A. M. Tamburro, Nanomedicine 2009, 4 , 469–487. 160. C. M. Venkatachalam, D. W. Urry, Macromolecules 1981, 14 , 1225–1229. 161. M. De Poli, A. Moretto, M. Crisma, C. Peggion, F. Formaggio, B. Kaptein, Q. B. Broxterman, C. Toniolo, Chem. Eur. J . 2009, 15 , 8015–8025. 162. V. Madison, K. D. Kopple, J. Am. Chem. Soc. 1980, 102 , 4855–4863. 163. Z. Liu, K. Chen, A. Ng, Z. S. Shi, R. W. Woody, N. R. Kallenbach, J. Am. Chem. Soc. 2004, 126 , 15141–15150. 164. R. W. Woody, The Peptides, Vol. 7, V. J. Hruby, ed., Academic Press, New York, 1985, pp. 15–114. 165. N. Sreerama, R. W. Woody, Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 601–619. 166. E. Vass, M. Hollosi, F. Besson, R. Buchet, Chem. Rev . 2003, 103 , 1917–1954. 167. E. Benedetti, A. Bavoso, B. Di Blasio, V. Pavone, C. Pedone, C. Toniolo, G. M. Bonora, Biopolymers 1983, 22 , 305–317. 168. P. M. Cowan, S. McGavin, Naure (London) 1955, 176 , 501–503. 169. W. Traub, U. Shmueli, Naure (London) 1963, 198 , 1165–1166. 170. M. W. MacArthur, J. M. Thornton, J. Mol. Biol . 1991, 218 , 397–412. 171. K. A. Williams, C. M. Deber, Biochemistry 1991, 30 , 8919–8923. 172. A. Yaron, F. Naider, Crit. Rev. Biochem. Mol. Biol . 1993, 28 , 31–81. 173. A. A. Adzhubei, M. J. E. Sternberg, J. Mol. Biol . 1993, 229 , 472–493. 174. M. P. Williamson, Biochem. J . 1994, 297 , 249–260. 175. R. Berisio, S. Loguercio, A. De Simone, A. Zagari, L. Vitagliano, Protein Pept. Lett. 2006, 13 , 847–854. 176. B. K. Ho, E. A. Coutsias, C. Seok, K. A. Dill, Protein Sci . 2005, 14 , 1011–1018. 177. R. Berisio, L. Vitagliano, L. Mazzarella, A. Zagari, Protein Sci . 2002, 11 , 262–270. 178. L. Stryer, R. P. Haugland, Proc. Natl. Acad. Sci. USA 1967, 58 , 719–726. 179. S. F. Sneddon, C. L. Brooks, III, J. Am. Chem. Soc. 1992, 114 , 8220–8225. 180. A. K. Mishra, R. Chandrasekar, M. Faraggi, M. H. Kappler, J. Am. Chem. Soc. 1994, 116 , 1414–1422. 181. H. Tamiaki, K. Nomura, K. Maruyama, Bull. Chem. Soc. Jpn 1994, 67 , 1863–1871. 182. S. Saxena, Y. H. Freed, Chem. Phys. Lett. 1996, 251 , 102–110. 183. W. S. Aldridge, III, B. J. Hornstein, S. Serrano, D. M. Dattelbaum, J. R. Schoonover, T. J. Meyer, J. Org. Chem. 2006, 71 , 5186–5190. 184. R. B. Best, K. A. Merchant, I. V. Gopich, B. Schuler, A. Bax, W. A. Eaton, Proc. Natl. Acad. Sci. USA 2007, 104 , 18964–18969. 185. S. Doose, H. Neuweiler, H. Barsch, M. Sauer, Proc. Natl. Acad. Sci. USA 2007, 104 , 17400–17405. 186. H. Sahoo, D. Roccatano, A. Hennig, W. M. Nau, J. Am. Chem. Soc. 2007, 129 , 9762–9772. 187. M. Cordes, A. K¨ottgen, C. Jasper, O. Jacques, H. Boudebous, B. Giese, Angew. Chem. Int. Edit. 2008, 47 , 3461–3463.

541

542

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

188. Z. Yang, D. Kise, S. Saxena, J. Phys. Chem. 2010, B 114, 6165–6174. 189. E. S. Price, M. S. DeVore, C. K. Johnson, J. Phys. Chem. 2010, B 114, 5895–5902. 190. B. Schuler, E. A. Lipman, P. J. Steinbach, M. Kumke, W. A. Eaton, Proc. Natl. Acad. Sci. USA 2005, 102 , 2754–2759. 191. G. N. Ramachandran, G. Kartha, Nature (London) 1954, 174 , 269–270. 192. A. Rich, F. H. C. Crick, Nature (London) 1955, 176 , 915–916. 193. A. V. Persikov, J. A. M. Ramshaw, B. Brodsky, Biopolymers (Pept. Sci.) 2000, 55 , 436–450. 194. L. Vitagliano, R. Berisio, L. Mazzarella, A. Zagari, Biopolymers 2001, 58 , 459–464. 195. K. Okuyama, C. Hongo, R. Fukushima, G. Wu, H. Narita, K. Noguchi, Y. Tanaka, N. Nishino, Biopolymers (Pept. Sci.) 2004, 76 , 367–377. 196. M. Goodman, M. Bhumralkar, E. A. Jefferson, J. Kwak, E. Locardi, Biopolymers (Pept. Sci.) 1998, 47 , 127–142. 197. S. N. Timasheff, H. Susi, R. Townend, L. Stevens, M. J. Gorbunoff, T. F. Kumosinski, Conformation of Biopolymers, G. N. Ramachandran, ed., Academic Press, New York, 1967, pp. 173–196. 198. M. L. Tiffany, S. Krimm, Biopolymers 1968, 6 , 1767–1770. 199. M. L. Tiffany, S. Krimm, Biopolymers 1969, 8 , 347–359. 200. R. W. Woody, J. Am. Chem. Soc. 2009, 131 , 8234–8245. 201. J. Applequist, Biopolymers 1981, 20 , 2311–2322. 202. V. Sasisekharan, Acta Crystallogr. 1959, 12 , 897–903. 203. H. Okabayshi, T. Isemura, S. Sakakibara, Biopolymers 1968, 6 , 323–330. 204. R. K. Dukor, T. A. Keiderling, Biopolymers 1991, 31 , 1747–1761. 205. M. Rothe, H. Rott, J. Mazanek, Peptides 1976 , A. Loffet, ed., Editions Universit´e de Bruxelles, Bruxelles, Belgium, 1976, pp. 309–318. 206. S. Kakinoki, Y. Hirano, M. Oka, Peptide Science 2004 , Y. Shimohigashi, ed., Japanese Peptide Society, Osaka, Japan, 2005, pp. 453–456. 207. E. R. Blout, Fundamental Aspects and Recent Developments in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli, P. Salvadori, eds., Heyden, London, U.K., 1973, pp. 352–372. 208. R. W. Woody, Adv. Biophys. Chem. 1992, 2 , 37–79. 209. Z. Shi, R. W. Woody, N. R. Kallenbach, Adv. Protein Chem. 2002, 62 , 163–240. 210. R. W. Woody, Monatsh. Chem. 2005, 136 , 347–366. 211. Z. S. Shi, K. Chen, Z. Liu, N. R. Kallenbach, Chem. Rev . 2006, 106 , 1877–1897. 212. E. S. Pysh, J. Mol. Biol . 1967, 23 , 587–599. 213. E. S. Pysh, J. Chem. Phys. 1970, 52 , 4723–4733. 214. J. Applequist, K. A. Bode, J. Phys. Chem. 1999, B 103, 1767–1773. 215. F. A. Bovey, F. P. Hood, Biopolymers 1967, 5 , 325–326. 216. E. W. Ronish, S. Krimm, Biopolymers 1974, 13 , 1635–1651. 217. C. Moore, G. D. Fasman, Chemtracts-Biochem. Mol. Biol . 1993, 4 , 67–74. 218. S. K. Brahmachari, M. Bansal, V. S. Ananthanarayanan, V. Sasisekharan, Macromolecules 1979, 12 , 23–28. 219. N. Helbecque, M. H. Loucheux-Lefebvre, Int. J. Pept. Protein Res. 1982, 19 , 94–101. 220. K. A. Thomasson, J. Applequist, Biopolymers 1991, 31 , 529–535. 221. M. A. Young, E. S. Pysh, J. Am. Chem. Soc. 1975, 97 , 5100–5103. 222. N. Sreerama, R. W. Woody, Biochemistry 1994, 33 , 10022–10025. 223. M. K¨umin, L.-S. Sonntag, H. Wennemers, J. Am. Chem. Soc. 2007, 129 , 466–467.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

224. M. K¨umin, S. Schweizer, C. Ochsenfeld, H. Wennemers, J. Am. Chem. Soc. 2009, 131 , 15474–15482. 225. D. G. McCafferty, C. A. Slate, B. M. Nakhle, H. D. Graham, Jr., T. L. Austell, R. W. Vachet, B. H. Mullis, B. W. Erickson, Tetrahedron 1995, 51 , 9859–9872. 226. L. Crespo, G. Sanclimens, B. Montaner, R. Perez-Tomas, M. Royo, M. Pons, F. Albericio, E. Giralt, J. Am. Chem. Soc. 2002, 124 , 8876–8883. 227. G. Sanclimens, L. Crespo, E. Giralt, F. Albericio, M. Royo, J. Org. Chem. 2005, 70 , 6274–6281. 228. M. A. Kelly, B. W. Chellgren, A. L. Rucker, J. M. Troutman, M. G. Fried, A.-F. Miller, T. P. Creamer, Biochemistry 2001, 40 , 14376–14383. 229. R. Schweitzer-Stenner, F. Eker, A. Perez, K. Griebenow, X. Cao, L. A. Nafie, Biopolymers (Pept. Sci.) 2003, 71 , 558–568. 230. R. Boni, A. S. Verdini, Macromolecules 1973, 6 , 517–520. 231. C. G. Overberger, M. J. Han, J. Polym. Sci., Polym. Chem. Edit. 1975, 13 , 2251–2258. 232. K.-D. Ahn, C. G. Overberger, J. Polym. Sci., Polym. Chem. Edit. 1983, 21 , 1699–1715. 233. W. W.-Y. Yang, C. G. Overberger, C. M. Venkatachalam, J. Polym. Sci., Polym. Chem. Edit. 1983, 21 , 1751–1761. 234. O. Tiba, C. G. Overberger, J. Polym. Sci., Polym. Chem. Edit. 1987, 25 , 2941–2952. 235. F. Rabanal, M. D. Ludevid, M. Pons, E. Giralt, Biopolymers 1993, 33 , 1019–1028. 236. E. C. Petrella, L. M. Machesky, D. A. Kaiser, T. D. Pollard, Biochemistry 1996, 35 , 16535–16543. 237. M. Mutter, T. W¨ohr, S. Gioria, M. Keller, Biopolymers (Pept. Sci.) 1999, 51 , 121–128. 238. A. Zhang, Y. Guo, Chem. Eur. J . 2008, 14 , 8939–8946. 239. L. Tooke, L. Duitch, T. J. Measey, R. Schweitzer-Stenner, Biopolymers 2010, 93 , 451–457. 240. E. Beausoleil, W. D. Lubell, Biopolymers 2000, 53 , 249–256. 241. P. Zhang, L. Polavarapu, J. Huang, T. Li, Chirality 2006, 19 , 99–106. 242. M. Wakahara, T. Nakamura, M. Oka, T. Hayashi, M. Hattori, Y. Hirano, Peptide Science 2001 , H. Aoyagi, ed., The Japanese Peptide Society, Osaka, Japan, 2002, pp. 289–292. 243. M. Yuge, S. Kakinoki, M. Kitamura, Y. Hirano, M. Oka, Peptide Science 2005 , T. Wakamiya, ed., The Japanese Peptide Society, Osaka, Japan, 2006, pp. 403–406. 244. M. Yuge, S. Kakinoki, M. Kitamura, Y. Hirano, M. Oka, Peptide Science 2006 , H. Ishida, H. Mihara, eds., The Japanese Peptide Society, Osaka, Japan, 2006, pp. 82–83. 245. M. Yuge, S. Kakinoki, Y. Hirano, M. Oka, Peptide Science 2006 , H. Ishida, H. Mihara, eds., The Japanese Peptide Society, Osaka, Japan, 2006, pp. 84–85. 246. L. Tterlikkis, F. M. Loxsom, W. Rhodes, Biopolymers 1973, 12 , 675–684. 247. W. C. Johnson, Jr., I. Tinoco, Jr., J. Am. Chem. Soc. 1972, 94 , 4389–4390. 248. D. D. Jenness, C. Sprecher, W. C. Johnson, Jr., Biopolymers 1976, 15 , 513–521. 249. F. R. Brown, III, J. Carver, E. R. Blout, J. Mol. Biol . 1969, 39 , 307–313. 250. G. M. Bonora, C. Toniolo, Biopolymers 1974, 13 , 1067–1078. 251. J. Kwak, A. De Capua, E. Locardi, M. Goodman, J. Am. Chem. Soc. 2002, 124 , 14085–14091. 252. G. A. Kinberger, J. P. Taulane, M. Goodman, Tetrahedron 2006, 62 , 5280–5286. 253. F. W. Kotch, I. A. Guzei, R. T. Raines, J. Am. Chem. Soc. 2008, 130 , 2952–2953. 254. W. C. Johnson, Jr., I. Tinoco, Jr., J. Am. Chem. Soc. 1972, 94 , 4389–4390. 255. M. L. Tiffany, S. Krimm, Biopolymers 1968, 6 , 1379–1382. 256. M. L. Tiffany, S. Krimm, Biopolymers 1972, 11 , 2309–2316. 257. A. F. Drake, G. Siligardi, W. A. Gibbons, Biophys. Chem. 1988, 31 , 143–146. 258. S. C. Yasui, T. A. Keiderling, J. Am. Chem. Soc. 1986, 108 , 5576–5581.

543

544

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

259. M. G. Paterlini, T. B. Freedman, L. A. Nafie, Biopolymers 1986, 25 , 1751–1765. 260. F. Eker, X. L. Cao, L. Nafie, R. Schweitzer-Stenner, J. Am. Chem. Soc. 2002, 124 , 14330–14341. 261. F. Eker, K. Griebenow, R. Schweitzer-Stenner, J. Am. Chem. Soc. 2003, 125 , 8178–8185. 262. F. Eker, K. Griebenow, X. L. Cao, L. A. Nafie, R. Schweitzer-Stenner, Proc. Natl. Acad. Sci. USA 2004, 101 , 10054–10059. 263. A. Hagarman, T. Measey, R. S. Doddasomayajula, I. Dragomir, F. Eker, K. Griebenow, R. Schweitzer-Stenner, J. Phys. Chem. 2006, B110 , 6979–6986. 264. R. Schweitzer-Stenner, T. J. Measey, Proc. Natl. Acad. Sci. USA 2007, 104 , 6649–6654. 265. R. Schweitzer-Stenner, J. Phys. Chem. 2009, B113 , 2922–2932. 266. Z. S. Shi, C. A. Olson, G. D. Rose, R. L. Baldwin, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2002, 99 , 9190–9195. 267. J. Graf, P. H. Nguyen, G. Stock, H. Schwalbe, J. Am. Chem. Soc. 2007, 129 , 1179–1189. 268. I. Gokce, R. W. Woody, G. Anderluh, J. H. Lakey, J. Am. Chem. Soc. 2005, 127 , 9700–9701. 269. E. A. Bienkiewicz, A.-Y. M. Woody, R. W. Woody, J. Mol. Biol . 2000, 297 , 119–133. 270. J. M. Scholtz, H. Qiang, E. J. York, J. M. Stewart, R. L. Baldwin, Biopolymers 1991, 31 , 1463–1470. 271. R. S. Lord, D. J. Cox, Biopolymers 1973, 12 , 2359–2373. 272. C. Toniolo, G. M. Bonora, Can. J. Chem. 1976, 54 , 70–76. 273. S. Mammi, M. Goodman, J. Pept. Sci . 2005, 11 , 273–277. 274. M. Goodman, F. Naider, R. Rupp, Bioorg. Chem. 1971, 1 , 310–328. 275. C. Toniolo, G. M. Bonora, M. Mutter, Makromol. Chem. 1975, 176 , 2547–2558. 276. G. M. Bonora, C. Toniolo, M. Mutter, Polymer 1978, 19 , 1382–1386. 277. E. Scoffone, F. Marchiori, R. Rocchi. A. Scatturin, A. M. Tamburro, Peptides 1969 , E. Scoffone, ed., North-Holland, Amsterdam, The Netherlands, 1971, pp. 223–230. 278. J. W. Nelson, N. R. Kallenbach, Proteins: Struct., Funct., Genet. 1986, 1 , 211–217. 279. B. Zagrovic, J. Lipfert, E. J. Sorin, I. S. Millett, W. F. van Gunsteren, S. Doniach, V. S. Pande, Proc. Natl. Acad. Sci. USA 2005, 102 , 11698–11703. 280. J. Makowska, S. Rodziewicz-Motowidło, K. Bagi´nska, J. A. Vila, A. Liwo, L. Chmurzy´nski, H. A. Scheraga, Proc. Natl. Acad. Sci. USA 2006, 103 , 1744–1749. 281. J. Makowska, S. Rodziewicz-Motowidło, K. Bagi´nska, M. Makowski, J. A. Vila, A. Liwo, L. Chmurzy´nski, H. A. Scheraga, Biophys. J . 2007, 92 , 2904–2917. 282. K. Chen, Z. Liu, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2004, 101 , 15352–15357. 283. K. Chen, Z. Liu, C. Zhou, Z. Shi, N. R. Kallenbach, J. Am. Chem. Soc. 2005, 127 , 10146–10147. 284. Z. Shi, K. Chen, Z. Liu, A. Ng, W. C. Bracken, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2005, 102 , 17964–17968. 285. K. Bagi´nska, J. Makowska, W. Wiczk, F. Kasprzykowski, L. Chmurzy´nski, J. Pept. Sci . 2008, 14 , 283–289. 286. P. Salvador, A. Asensio, J. J. Dannenberg, J. Phys. Chem. 2007, B111 , 7462–7466. 287. B. W. Chellgren, T. P. Creamer, Biochemistry 2004, 43 , 5864–5869. 288. A. M. Firestine, V. M. Chellgren, S. J. Rucker, T. E. Lester, T. P. Creamer, Biochemistry 2008, 47 , 3216–3224. 289. V. Setnicka, J. Hlavacek, M. Urbanova, J. Pept. Sci . 2009, 15 , 533–539. 290. I. Jakli, A. Perczel, J. Pept. Sci . 2009, 15 , 738–752. 291. M. Narita, N. Ohkawa, S. Nagasawa, S. Isokawa, Int. J. Pept. Protein Res. 1984, 24 , 129–134. 292. A. Rath, A. R. Davidson, C. M. Deber, Biopolymers (Pept. Sci.) 2005, 80 , 179–185.

16 ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS Claudio Toniolo and Fernando Formaggio

16.1. INTRODUCTION Peptidomimetics, originally also termed pseudopeptides or peptide surrogates, are peptide analogues that can contain natural (but noncoded) or synthetic amino acids. They are generated by modifying existing peptides to alter their properties (e.g., enzymatic stability, lipophilicity, biological activity). As a result, they are relevant for the development of compounds with novel therapeutic profiles. In recent years they have found wide application in medicinal chemistry as biostable, bioavailable, and often potent mimetics of naturally occurring peptides. Initial synthetic efforts were centered on modifications of the peptide side chains, or involved amino acid additions, deletions, or substitutions only, but more recently the main interest of peptide chemists from academia and industrial laboratories as well have focused mainly on backbone modifications. Several review articles [1–14] have dealt with the molecular design of specific, receptor-selective peptidomimetic ligands and with their challenging synthetic issues and intriguing conformational preferences, but none of them has ever treated in detail the electronic circular dichroism (ECD) properties of the chromophores characterizing this class of peptide analogues. In this overview, for the first time the results of the published ECD studies of peptidomimetics have been summarized and critically discussed. We have restricted our attention to the most relevant and popular subclass of peptidomimetics, namely those with modifications in the -NH–CH(R)–CO- backbone. For the ECD properties of unmodified peptides and poly-α-amino acids in their classical secondary structures, the reader is referred to Chapter 15 in this volume.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

545

546

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Oi

Hi

Ci′

Ni Ci

C′i−1

Oi−1

α

Ni+1

Figure 16.1. Intramolecular steric repulsions that H

induce warping of the i → i intramolecularly

R

Hi+1

H-bonded (C5 ) peptide conformation.

16.2. FULLY EXTENDED PEPTIDE CONFORMATION (2.05 -HELIX) The fully extended polypeptide conformation (φi = ψi = 180◦ ) (2.05 -helix) was proposed at an early stage in 3D-structural studies of fibrous proteins. In this form, H-bonding takes place between the N–H groups of one chain and the C=O groups of the chains on either side, thus making a planar sheet held together by interchain H-bonds directed approximately perpendicular to the chain axis. Neighboring sheets are then held together by van der Waals forces. In 1951 Pauling and Corey [15] investigated the possibility of small contractions of the polypeptide chains and proposed precise conformations for parallel and antiparallel pleated -sheet β-forms which better satisfy stereochemical and H-bonding requirements and have chain-repeat lengths nearer to those found experimentally. These authors were also able to show that steric hindrance between adjacent chains prevents the onset of the planar sheet in case the side chain is anything but a H-atom, that is, this conformation could be formed only by poly(Gly). The repeating motif of the fully extended, flat polypeptide conformation is the (i )N–H· · ·O=C (i ) intramolecularly H-bonded form (Figure 16.1). The relative disposition of the two dipoles, Ni –Hi and Ci =Oi , is such that there is obviously some interaction between them. These four atoms, together with the Cαi atom, are involved in a pentagonal cyclic structure. It is for this reason that this conformation is also called the C5 structure. The C5 form was considered in conformational energy calculations and its occurrence in apolar noninteracting solvents was proposed in solution studies using IR absorption and NMR measurements of model peptides [16–18]. Gly derivatives have the highest population of C5 structure if compared to the derivatives of residues carrying a side chain. The influence of the bulkiness of the side chain can easily be explained by considering the intramolecular nonbonded interactions between the group R and the atoms Hi +1 and Oi −1 , which induce a warping of these asymmetric molecules (Figure 16.1). Unequivocal verification of the occurrence of the C5 form was obtained in the crystal state in a few peptides and proteins by X-ray diffraction analysis [18–21]. Among the coded amino acids, Gly was found to be involved in > 99% of the C5 structures, including an unusual stretch of four consecutive Gly residues which therefore forms a short twofold 2.05 -helix. Notably, a variant of the planar C5 form (with Y-conjugation) was reported in an X-ray diffraction investigation of homo-peptides from the natural, but noncoded, achiral Ala (α,β-didehydro Ala) residue [22]. The results of the theoretical and experimental (crystal-state and low-polarity solvents) studies from the Toniolo and Benedetti laboratories, summarized in references 23, and 24, strongly supported the view that the quaternary α-amino acids C α,α -diethylglycine (Deg, Figure 16.2), C α,α -di-n-propylglycine, C α,α -diphenylglycine, and C α,α -dibenzylglycine, all achiral (with Cα,α -symmetrically disubstituted side chains), overwhelmingly tend to adopt the C5 structure. This peptide conformation

547

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

H3C

H2C

CH2

CH3

H3C H2C

C

CH3

H3C H2C

C CO

NH

(CH2)3

Deg

CH3

C CO

NH

(CH2)4

(S)-Beg

NH

CO (S)-Epg

α

Figure 16.2. Chemical formulae for the C -ethylated achiral Deg, and chiral Beg and Epg α-amino acid residues.

Figure 16.3. Average geometrical parameters that characterize the fully-extended, intramolecularly H-bonded C5 structure from a statistical analysis of the known X-ray diffraction structures. (Adapted from reference 23.)

was characterized at atomic resolution (Figure 16.3). The narrowing of the tetrahedral ˚ N· · ·O distance for the N–Cα –C bond angle (τ ) to ∼103◦ and the very short (2.54 A) intramolecular H-bond are the most notable features. In the homo-peptides from these residues the multiple N–H and C=O groups involved in the intramolecularly H-bonded forms are not implicated in any intermolecular H-bonding, possibly because severe steric hindrance from the bulky side chains prevent adjacent peptide chains to approach sufficiently. In conclusion, it became evident that the multiple C5 structure (2.05 -helix) becomes stable enough, at least in homo-peptides, when both side-chain Cβ -atoms of the constituent residue are symmetrically substituted (but not interconnected in a cyclic system). In contrast, when both side-chain Cβ -atoms of the constituent residue are unsubstituted, as in α-aminoisobutyric acid (Aib), or when only one of them is substituted, as in chiral C α -methylated residues, then the 310 -(or α-)helix is the largely preferred conformation [24]. However, since 2000 this previously widely accepted dogma does not hold any more. Tanaka and his colleagues [25, 26] reported that the terminally protected, chiral homotetramer Tfa–[(S )-Beg]4 –OEt (Tfa, trifluoroacetyl; Beg, C α -n-butyl, C α -ethylglycine, Figure 16.2; OEt, ethoxy) adopts a fully planar 2.05 -helix in the crystal state. This surprising finding was subsequently supported by Toniolo et al. [27], who studied Epg (C α -ethyl, C α -n-pentylglycine, Figure 16.2) homo-oligomers in the crystal state. The X-ray diffraction structure of the double C5 conformation of the dipeptide Tfa–[(S )Epg]2 –OtBu (OtBu, tert-butoxy) is shown in Figure 16.4. At this point, the unavoidable conclusion is that an asymmetric Cα,α -disubstituted Gly residue, bearing two different

548

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 16.4. X-Ray diffraction structure of Tfa–[(S)-Epg]2 –OtBu. Intramolecular N–H· · ·O=C Hbonds are represented by dashed lines; F and O atoms are larger and black and gray, respectively, while N and H atoms are white. (Adapted from reference 27.)

side chains each with at least two carbon atoms, may represent an appropriate building block for the construction of a 2.05 -helix. These investigations have also shown that the integrity of this type of helical structure is preserved only in solvents of low polarity and provided that the C-terminal protection of the peptide sequence is devoid of any H-bonding donor group. It is evident that, based on the above discussed data, an analysis of the experimental ECD properties of chiral 2.05 -helices has become an accessible task. To this end, and toward the in-depth elucidation of the ECD spectrum by a theoretical approach, collaborative efforts are currently being conducted [28]. An important requirement for a successful experimental study in solution is that the solubility of the chiral oligopeptide would be sufficiently enhanced in UV-transparent, low-polarity solvents, as compared to that of the peptides prepared and investigated to date. Unfortunately, recent attempts to obtain reliable ECD spectra using the polar solvent 2,2,2-trifluoroethanol (TFE) did not furnish appreciable results [25, 29].

16.3. POLY-N(ALKYL)-α-AMINO ACIDS Poly(Pro) and poly-N (alkyl)-α-amino acids, abbreviated here as poly[N (R)AA], are characterized exclusively by tertiary peptide bonds. In this sense, the latter may be considered acyclic analogues of poly(Pro). The least congested polypeptides of this family, poly[N (alkyl)Gly] or peptoids, will be discussed separately (next section). Among the ECD properties of the poly[N(R)AA] from chiral α-amino acids, only those of polyN (methyl)-α-amino acids have been investigated (Figure 16.5a). Notably, N -methylation is an important modification of the peptide bond [30–33]. It commonly occurs in natural peptides from plants, marine sources, and a variety of microorganisms. Several

549

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

(a)

R N

CH *

CO n

R = CH3

MeAla

R = (CH2)2-COOC2H5

MeGlu(OEt)

CH3 (b)

Figure 16.5. (a) General formulae of the N-methylated homo-poly-α-amino acids discussed here and (b) X-ray diffraction structure of H–(L-MeAla)6 –OH (O atoms larger and gray, N atoms white). (Adapted from reference 33.)

of these compounds possess interesting antibiotic, antitumor, and immunosuppressant activities. Synthetic, N -methylated analogues of bioactive peptides often exhibit high stability to enzymatic degradation. N -methylated peptides are also potent inhibitors of β-sheet structure and, as a result, of amyloid formation. In their pioneering work, Mark and Goodman [34, 35] demonstrated by energy calculations that the homo-polymer poly(l-MeAla), where MeAla is N -methylated Ala, is severely conformationally restricted and is preferentially folded in a right-handed, approximately threefold helix with all peptide bonds in the trans conformation. This unusual, semi -extended, nonintramolecularly H-bonded, secondary structure has backbone φ, ψ torsion angles of about −150◦ , 70◦ . The main factor favoring this conformation for poly(l-MeAla) is the severe steric repulsion between the C α -methyl and N -methyl groups that would occur in other, more canonical peptide conformations. These conclusions were essentially confirmed by subsequent works from other laboratories [36, 37]. Madison and Schellman [38] calculated the ECD spectrum of the most stable, alltrans, conformation for the (l-MeAla)20 homo-oligomer. They found two bands of approximately the same intensity at 221 nm (negative), associated with the n → π * transition of the peptide chromophore, and at 193 nm (positive), associated with a complex pattern of π → π * transitions (Figure 16.6). The crossover point was predicted at 207 nm. Goodman and co-workers [39, 40] reported the experimental ECD spectrum of poly(l-MeAla) in the structure-supporting solvent TFE (Figure 16.7). It shows a broad negative Cotton effect centered at 223 nm followed by a slightly less intense, positive Cotton effect at 192 nm (the crossover point is seen at 204 nm). Thus, the theoretical and experimental ECD results fit nicely. Moreover, their findings on poly(l-MeAla) differ somewhat from those obtained with poly(l-Pro) [38] in that no evidence for an all-cis form was detected upon changing solvent polarity. Furthermore, a few ECD studies were performed in trifluoroacetic acid (TFA), due to the generally poor solubility of poly(l-MeAla). However, the conclusions from these studies are doubtful, as TFA was many years later shown to easily hydrolyze tertiary peptide bonds [32, 41]. Arvidsson and co-workers [33] investigated the ECD properties of the monodisperse homo-oligomer H–(l-MeAla)6 –OH in methanol and aqueous solutions. The spectra parallel closely those described for the corresponding polymer in TFE [39, 40]. In addition, Peggion, Goodman, and their colleagues reported conformational energy calculations and ECD data for poly[l-MeGlu(OEt)] [where MeGlu(OEt) is N -methyl, γ -ethyl glutamate] [42]. Their findings agree well with those for poly(l-MeAla).

550

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20

[θ] × 10–3 (deg × cm2 × dmol–1)

10

0

–10

–20

Figure 16.6. Computed ECD spectrum of the 200

220

240

Wavelength (nm)

260

(L-MeAla)20 homo-oligomer. (Adapted from reference 38.)

However, in contrast to this remarkably closely fitting scenario, quite surprisingly, in their recent X-ray diffraction analysis of the two homo-oligomers H–(l-MeAla)5,6 –OH (Figure 16.5b) Arvidsson and co-workers [33] unambiguously showed that these peptides do not form a helix with all φ, ψ torsion angles of (−135◦ , 65◦ ), but this conformation strictly alternates with that characteristic of a distinct semi -extended conformation (−65◦ , 140◦ ). The former backbone conformation generates a torsion angle between the two MeAla methyl groups [(methyl)C–N–Cα –Cβ ] of about 70◦ , while the torsion angle in the latter conformation, which resembles that of poly(l-Pro) PPII [38], is reduced to approximately 20◦ . At this point, there is a clear need for a theoretical investigation of the ECD properties of this novel poly(l-MeAla), sequentially bis-semi -extended, 3D-structure.

16.4. POLY-PEPTOIDS Peptoids (Figure 16.8a) are polymers of Nα -substituted Gly residues [43]. Therefore, they are generally achiral, unless the nitrogen atom is substituted with a group (R) containing an asymmetric center (Figure 16.8b). Peptoids can be defined as α-peptide mimics in which the side chain R is linked to the α-nitrogen instead of the α-carbon atom. They also represent a variant with respect to the chiral poly-N (alkyl)-α-amino acids, poly[N(R)AA],

551

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

40

[θ] × 10–3 (deg × cm2 × dmol–1)

20

0

–20

Figure 16.7. ECD spectrum of

–40 200

220

240

260

Wavelength (nm)

(a)

O

R

H

H

O

N

*

*

N

O

H O

H

R

H

R*

H

chiral peptide

* O

H O

(Adapted from reference 39.)

H

R N

H

H

N N O

R (b)

H3C

chiral peptoid

N *

H

H

*

R

Nsch C6H11

Figure 16.8. (a) Chemical formulae of chiral peptide and peptoid chains. Starred

O H3C

H

homo-poly[L-MeAla] in TFE solution.

carbon atoms or R groups indicate chiralities. (b) Chemical formulae of the two α-chiral, aliphatic, N-substituents investigated: Nsch

H Nssb

is N-(S)-(1-cyclohexylethyl)glycine and Nssb is N-(S)-(sec-butyl)glycine.

552

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

where AA is any α-amino acid other than Gly, discussed in the preceding section. These compounds are an attractive scaffold for biological applications (e.g., as antimicrobials) because they can be generated using a straightforward, modular synthesis that allows the incorporation of a wide variety of functionalities [44, 45]. Peptoids have been evaluated as tools to study biomolecular interactions, and they also hold significant promise for therapeutic applications due to their enhanced proteolytic stabilities and increased cellular permeabilities relative to α-peptides. The unsubstituted α-carbon of peptoids should allow almost unhindered rotation about the φ, ψ torsion angles. In addition, both cis and trans conformers of the tertiary amide bonds should be accessible at room temperature. However, the substituent on the α-nitrogen is expected to confer 3D-structural properties on peptoids that might seriously limit their flexibility. Here, we will discuss only peptoids with aliphatic N-substituents, in that aromatic chromophores can affect the ECD spectral shapes. Also, we will not treat hybrids of α-amino acids and Nα -substituted Gly residues. The lowest-energy conformation for the simplest model compound of the aliphatic peptoid subclass, Ac–Sar–NMe2 (Ac, acetyl; Sar, sarcosine or N -methylglycine; NMe2 , dimethylamino) corresponds to φ = ±90◦ , ψ = 180◦ for both the trans and cis Ac–Sar amide conformers [44]. In the crystal state, the homo-pentapeptoid H-(Nrch)5 -NH2 [Nrch, N -(R)-(1-cyclohexylethyl)glycine] adopts a left-handed helical conformation with repeating cis amide bonds (Figure 16.9) [46]. The periodicity of this helix is approximately three residues per turn. The C=O groups are aligned with the helix axis. The handedness of the helix is governed by the chirality of the Nα -substituent. The backbone φ, ψ torsion angles are similar to those observed for the PPI poly(Pro) helix, but with opposite signs (obviously, the signs depend on the Nch peptoid chirality). The bands in the ECD spectrum of the α-chiral aliphatic homo-oligopeptoid H–(Nsch)5 –NH2 [Nsch, N -(S )-(1-cyclohexylethyl)glycine] in CH3 CN are relatively weak, reflecting only a partial helical ordering in solution [46, 47]. However, those of the longer homo-oligomers investigated—for example, H–(Nsch)15 –NH2 (Figure 16.10)—show a distinct positive maximum at 210 nm, and two stronger negative maxima at 195 nm and 225 nm, respectively. The ECD bands are spectral characteristics that are typically associated with those of the PPI poly(Pro) helix [48]. Therefore, these

Figure 16.9. X-ray diffraction structure of the Nrch homo-pentapeptoid amide (left, perpendicular; right, parallel to the threefold helix axis). (Adapted from reference 46.)

553

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

4

[θ]R × 10–3 (deg × cm2 × dmol–1)

2

0

–2

–4

–6

–8

Figure 16.10. ECD spectrum of the Nsch

–10 180

200

220 Wavelength (nm)

240

homo-pentadecapeptoid in CH3 CN solution. (Adapted from reference 46.)

ECD findings are consistent with the results from the crystallographic analysis. The shapes of the ECD spectra and their steady increase in intensity with lengthening of the main chain are qualitatively similar for the Nssb [N -(S )-(sec-butyl)glycine] subclass of aliphatic sidechain containing peptoid homo-oligomers. This latter observation points to a more ordered helical fold for the longest oligo-peptoids. ECD also demonstrated that this helical structure is quite stable upon heating. This result is consistent with a 3D-structure predominantly stabilized by steric repulsion rather than by intrachain H-bonding. NMR data on these foldameric series provide strong evidence that the tertiary amide cis –trans isomeric ratio is main-chain length dependent and that at the n = 15 level the cis-amide family of dynamically interconverting conformers is overwhelmingly populated. Introduction of the achiral, water-solubilizing NLys [N -(4-aminobutyl)glycine] and NArg [N -(3-guanidinopropyl)glycine] guest residues into the host Nssb or Nsch homopeptoid chains allowed ECD measurements to be performed in aqueous buffer solutions [49–51]. Under these conditions, the ECD spectra are weak, reminiscent of that of a peptide unordered conformation. However, in methanol (MeOH) solution and in membranemimetic vesicles the curves are more intense and resemble those of PPI poly(l-Pro). Not surprisingly, the intensity of the ECD curves is proportional to the percentage of chiral Nα -substituted residues.

554

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R peptide

CO

NH

CH

R CO NH

R depsipeptide

CO

NH CH

CH

CO NH

R′ CO

O

CH

R

R CH

CO NH

R CO NH

CH

CH

CO

R′ CO

O

CH CO

Figure 16.11. Chemical formulae of peptide and depsipeptide (the latter with strictly alternating amide-ester groups) chains.

16.5. POLY-DEPSIPEPTIDES In the classical definition, a poly-depsipeptide chain is generated by a strict alternation of an α-amino acid and an α-hydroxy acid in a linear sequence (Figure 16.11). Therefore, this kind of sequential polymer is characterized by a repeating dyad with an amide unit followed (or preceded) by an ester unit. Subsequently, this terminology was inappropriately extended to any backbone-modified peptide that incorporates one or more α-hydroxy acid building blocks. It is worth noting that increasing effort is currently being devoted to the conformational analysis of a variety of compounds characterized by the presence of α-(or β-)hydroxy acids. More specifically, these compounds include (a) oligo- and polyesters as biodegradable and biocompatible materials [52–55] and (b) depsipeptides and depsiproteins, in this broader sense, to mimic naturally occurring ion carriers [56, 57] or to check the influence of specific H-bonds on peptide or protein bioactivity and conformation [58–62]. In this section, we will discuss exclusively the chirospectroscopic properties of classical depsipeptides. As stated above, poly-depsipeptides represent a combination of polypeptides and poly-α-esters (for the 3D-structural and chirospectroscopic properties of polypeptides, see Chapter 15, this volume). The 3D-structure of the simplest chiral poly-α-ester, poly(lLac) where Lac is lactic acid, was investigated by energy computations and in the crystal state [63–68]. The results point to a regular or slightly distorted, right-handed, threefold helix. However, according to the optical rotatory dispersion data of Goodman and D’Alagni [69], this homo-polymer, the chemical structure of which rules out any possibility of intramolecular H-bonding, may not have a helical conformation in solution. The UV absorption and related ECD spectra of ester model compounds, oligo- and poly(lLac), and poly-β-esters exhibit a solvent-dependent band near 210 nm, arising from the n → π * transition of the –COOR chromophore [52, 53, 55, 69–73]. The X-ray diffraction structures of the l-Lac dimer and trimer were recently solved [74]. The only available X-ray diffraction structure of a classical depsipeptide, Z–(Aib–Hib)2 –Aib–OMe [Z, benzyloxycarbonyl; Hib, α-hydroxyisobutyric acid; OMe, methoxy] [59] (Figure 16.12), strongly supports the view that the α-hydroxy acid Hib guest unit is easily incorporated into the 310 -helical structure typical of the host α-amino acid Aib chain (in this depsipeptide both building blocks are characterized by two helicogenic gem-methyl groups at the Cα -atom). The resulting helical conformation, in which the scheme of consecutive intramolecular H-bonds is interrupted at alternate positions, is classified as a β-bend ribbon spiral. Two preferred depsipeptide secondary structures were identified from energy calculations on the poly-depsipeptide (l-Ala–l-Lac) [63, 64]. The most stable helix, termed R = 13 (Figure 16.13), has ϕ, ψ torsion angles (−65◦ , −35◦ for l-Ala; −63◦ , −47◦ for l-Lac) similar to those of the standard right-handed polypeptide α-helix. The amide and ester

555

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

Figure 16.12. X-ray diffraction structure of the oligo-pentadepsipeptide Z–(Aib–Hib)2 –OMe. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white. (Adapted from reference 59.)

O O C

C

C

C

N

CH

C O

H

C C

O C

N

H O

C

C

H

O

C N

H

O H O

C

O

C C H

C

C N C

N C

C O

C

C

H C

C H

O

O

C C

H

O

H H

O

C

N

C

C

O

H

H

Figure 16.13. Left, the R = 10 helix, and right, the R = 13 helix, calculated for the polydepsipeptide (L-Ala– L-Lac). (Adapted from reference 64.)

C=O bonds are approximately parallel to the helix axis. Intramolecular H-bonding occurs between ester carbonyl oxygen atoms and amide NH atoms that are separated by three α-carbons. The pseudo-ring thus generated is characterized by 13 atoms. The other helix, termed R = 10 (Figure 16.13), is left-handed with backbone torsion angles 51◦ , −94◦ for l-Ala and −144◦ , 30◦ for l-Lac. The amide C=O and N–H bonds are

556

8 4 0 –4 –8 160 180 200 220 240

12

[θ] × 10–3 (deg × cm2 × dmol–1)

[θ] × 10–3 (deg × cm2 × dmol–1)

[θ] × 10–3 (deg × cm2 × dmol–1)

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8 4 0 –4 –8 160 180 200 220 240

8

4

0

–4

–8 160 180 200 220 240

Wavelength (nm)

Wavelength (nm)

Wavelength (nm)

(a)

(b)

(c)

Figure 16.14. Theoretical ECD spectra of the poly-depsipeptide (L-Ala– L-Lac) R = 13 helix (a), R = 10 helix (b), and the unordered conformation (c, dashed line). The full line in (c) is the spectrum of a mixture of 50% unordered conformation and 50% R = 10 helix. (Adapted from reference 72.)

roughly parallel to the helix axis, whereas the ester C=O bonds are approximately perpendicular. Intramolecular H-bonds are observed between neighboring amide groups, which produce pseudo-ring structures formed by 10 atoms. Comparison of the theoretical (Figure 16.14) and experimental (Figure 16.15) ECD properties suggests that poly(l-Ala–l-Lac) is only about 50% ordered in solution (even in solvents of low polarity) [75–77]. The experimental ECD spectra of two other poly-depsipeptides, poly(l-Val–l-Lac) and poly(l-Ala–l-Hiv) where Hiv is α-hydroxyisovaleric acid, are similar. Not surprisingly, the calculated ECD spectrum for the R = 13 helix is close in shape to the ECD curves of polypeptides in the right-handed α-helix structure. However, it is clearly very different from the experimental poly-depsipeptide ECD spectra. Apparently, these polymers are not folded in the right-handed R = 13 helical conformation to any significant extent in solution.

16.6. α,β-DIDEHYDRO-α-AMINO ACID-BASED POLY-PEPTIDES α,β-Didehydro(unsaturated)-α-amino acids (Figure 16.16), usually represented by the notation AA, have been frequently found in naturally occurring peptides of microbial origin and in a limited number of proteins [78–81]. They are also constituents of an important class of polycyclic peptide antibiotics (lantibiotics). The presence of AA in peptides alters lipophilicity and bioactivity as well as increases resistance to hydrolysis by proteolytic enzymes. AA residues have been incorporated by chemical synthesis in a variety of natural peptide sequences to obtain highly active analogues. This modification has become a useful method to study structure–function relationships in bioactive peptides. The accumulation of three functionalities (the amide α-NH- and α-CO- groups, along with the C=C double bond) at position Cα of a AA residue has remarkable stereochemical and spectroscopic effects [82–86]. In particular, the presence of an sp 2 hybridized carbon (Cα ) atom in the backbone, the altered electronic distribution

557

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

9 2

6

[θ] × 10–3 (deg × cm2 × dmol–1)

1

3

0

3 –3

Figure 16.15. ECD spectra of the

–6

poly-depsipeptides (L-Ala– L-Lac) (1),

200

220

240

Wavelength (nm)

R1

R C

R=H

R = C6H5 R1 = H

C HN

R1 = H

CO

R=H

260

(L-Val– L-Lac) (2), and (L-Ala– L-Hiv) (3) in tetrahydrofuran solution at −50◦ C. (Adapted from reference 75.)

ΔAla ΔZPhe

R1 = C6H5 ΔEPhe

Figure 16.16. Chemical formulae of representative α,β-didehydro()α-amino acid residues, including the two Phe E- and Z- diastereoisomers.

(Y-conjugation) caused by the α –β π -system, and the change in the side-chain rotamer population all contribute significantly to the preferred conformation of the peptide main chain. The 3D-structural flexibility of the -peptide backbone as well as of the specific side chain of the AA residue is restricted on account of the double bond between the Cα and Cβ atoms. Although conjugation requires an extended conformation, the bulkiness of the side chain may play a significant role in the overall conformation of the AA residue. In the -NH–C(=CRR1 )–CO- class (with the R and R1 groups of protein amino acids) all residues exhibit cis (Z )–trans (E ) isomerism around the C=C double bond, except Ala (R = R1 = H) (Figure 16.16) and Val (R = R1 = CH3 ). An example of this type of diastereoisomerism is shown in Figure 16.16 for Phe. In Z Phe, the C=O group is in the trans position with respect to the phenyl group, while in the E Phe it is in the cis position. Most of the conformational studies on AA-rich peptides were carried out on Phe peptides and particularly on peptides based on Z Phe because of the relative ease of

558

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

synthesis of this isomer and the fact that most preparation procedures lead to the Z Phe configuration. The average bond lengths and bond angles of AA residues were calculated from a variety of crystal structures. The shortening of N–Cα and Cα –C single bond lengths and the elongation of the C =O double bond in Phe may be due to partial conjugation of the Cα =Cβ double bond with the adjacent peptide bonds. The intramolecular steric clash between Cδ H and NH of the Phe residue results in a remarkable opening up of the bond angles Cα –Cβ –Cγ and N–Cα –Cβ . From simple model building studies it was observed that the most favorable conformations for the Phe residue are (φ, ψ) ≈ (±60◦ , ±30◦ ) and (±60◦ , ∓150◦ ). Theoretical conformational studies confirmed these findings and suggested that Phe strongly favors β-turns and 310 - (or α-) helical conformations. The near-UV absorption spectrum of peptides containing Phe is characterized by an intense, conformationally useful absorption band at about 280 nm, which has been assigned to a charge-transfer electronic transition from the electron donating styryl groups to the electron-accepting carbonyl group in the Phe moiety [87, 88]. The chromophoric system of the residue, therefore, is essentially the cinnamic moiety C6 H5 –C=C–C=O. The peptides containing two Phe residues show an analogous band, centered around the same wavelength. Not unexpectedly, its intensity is approximately twice as that of their mono-unsaturated counterparts. In addition, a rather strong band at about 220 nm dominates the far-UV absorption spectrum of Phe peptides. It is clear that this latter spectroscopic property, unfortunately, is potentially confusing in terms of a correct peptide conformational assignment. The reasons for the small differences, if any, between the absorptions of the Phe Z- and E-diastereoisomers are not well understood. Induced ECD proved to be an excellent tool to determine the solution conformation of peptides based on the achiral Phe residue, especially in detecting the screw sense of the helices formed. It is principally for this reason that the number of publications on this topic is huge [87,89–126]. The Z Phe-containing peptides show different ECD curves, depending on the main-chain length and on the position in the sequence and number of the Z Phe residues. ECD spectra of mono Z Phe compounds exhibit only low-intensity bands in the near-UV region (Figure 16.17). In general, the shape and sign of these bands are strongly affected by the nature of the nearby chiral α-amino acid. ECD spectra of Z Phe tri- and longer peptides exhibit a broad, relatively intense band at about

[θ] × 10–4 (deg × cm2 × dmol–1)

20 10 1 0 –10 2 –20

Figure 16.17. ECD spectra of the dipeptide

–30

Ac–Z Phe– L-Ala–OH (1) and the tripeptide Ac–(Z Phe)2 – L-Ala–OH (2) in MeOH solution.

240

280 Wavelength (nm)

320

The latter peptide is characterized by two consecutive Z Phe residues. (Adapted from reference 89.)

559

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

[θ] × 10–3 (deg × cm2 × dmol–1)

20

0

–20

Figure 16.18. ECD spectrum of the

240

280 Wavelength (nm)

320

tetrapeptide Ac–(Z Phe)3 – L-Ala–OH in MeOH solution. This peptide is characterized by three consecutive Z Phe residues. (Adapted from reference 90.)

280 nm, corresponding to the main absorption band of the cinnamoyl-like chromophore. The varying intensities of this band are indicative of the different propensities of the peptides to fold into a β-turn. This conclusion is confirmed by the solvent dependence of the intensity of this band. Peptides containing two or more (up to eight, of which 2–4 are consecutive) Z Phe residues show a couplet of intense bands with opposite signs at 300 and 260 nm and a crossover point at ∼280 nm (Figures 16.17–16.22). This ECD pattern is typical of exciton splitting due to the dipole–dipole interactions between the Z Phe chromophores and is a strong indication that the two Z Phe residues are placed in a mutual, fixed disposition within the structure of the molecule, generally a 310 -helix. The (−+) signs (in the direction of decreasing wavelengths) of the couplet correspond to a 310 -helix with a right-handed screw sense, while the (+−) couplet is assigned to the helix with the left-handed screw sense. Comparison of ellipticity of the peptides containing two and three Z Phe residues show similar values, in spite of the presence of an extra Z Phe in the latter case. This should actually be the case if the 310 -helical conformation of the peptide would terminate with a type I, instead of a type III, β-turn. It may also be a result of the change of the helix sense of the last residue or unwinding of the helix, a common observation at the helix C-terminus [127]. In the peptides where the chiral residue(s) are located in internal

560

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30 1

[θ] × 10–3 (deg × cm2 × dmol–1)

2 15 3

0 4

–15

Figure 16.19. ECD spectra of the pentapeptide Boc– L-Ala–Z Phe–Gly–Z Phe– L-Ala–OMe, where Boc is tert-butyloxycarbonyl, in CH3 CN (1),

–30 240

320

280

solutions. (Adapted from reference 92.)

Wavelength (nm)

L-Leu

CH2 Cl2 (2), MeOH (3), and 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) (4)

L-Val

[θ] × 10–3 (deg × cm2 × dmol–1)

40

20

L-Pro L-Ala

0

–20

D-Pro no chiral additive

–40 240

Figure 16.20. ECD spectra of the achiral

D-Leu

280 Wavelength (nm)

320

nonapeptide H–Gly–(Z Phe–Aib)4 –OMe in the presence or absence of added Boc–Xxx–OH (each Xxx residue is indicated). (Adapted from reference 108.)

position(s) of the sequence, the helix is right-handed, while peptide esters (but not peptide amides) where the single chiral residue is at the C-terminus are found to adopt the lefthanded helical sense. If the single chiral residue is, however, at the N-terminus, the peptides tend to adopt 310 -helix conformations of both screw senses in the crystal state. In this case, the ECD results are consistent with the presence of right- and left-handed conformers also in solution, with a prevalence of the more stable right-handed helix. Interestingly, in a few peptides a reversible screw sense inversion of the 310 -helix was detected, depending on solvent (Figures 16.19 and 16.21) and temperature conditions. An ECD band was also observed at 320 nm in some of these compounds [90, 118] that also changes sign with changing solvent polarity (Figure 16.21). One reason for the

561

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

1 2 3 4 5 6 7 8 9

[θ] × 10–3 (deg × cm2 × dmol–1)

300 1 200

2 3

100

5

0

100:0 90:10 80:20 70:30 60:40 50:50 40:60 30:70 20:80

4

9 6

–100

7 8

Figure 16.21. ECD spectra of the decapeptide

–200 260

240

280

300

320

[θ] × 10–3 (deg × cm2 × dmol–1)

Boc– L-Ala–(Z Phe)4 – L-Ala–(Z Phe)3 –Gly–OMe in CHCl3 –MeOH solvent mixtures (curve 1: 100% CHCl3 ). (Adapted from reference 118.)

Wavelength (nm)

10

340

Z

E

0

–10

Figure 16.22. ECD spectra of the hexapeptides Boc–Gly–E Phe– L-Phe–Gly–E Phe– L-Phe–OH (E)

240

280 Wavelength (nm)

320

and Boc–Gly–Z Phe– L-Phe–Gly–Z Phe– L-Phe–OH (Z) in CHCl3 solution. (Adapted from reference 115.)

occurrence of this band may be the weak electronic transition polarized along the short axis of the benzene ring. This contribution to the ECD spectrum suggests that the benzene ring is not free to rotate, as expected, owing to the presence of the Cα =Cβ double bond. Its strong intensity also indicates that the phenyl ring is restricted to a preferred chiral disposition. The usefulness of the Z Phe chromophore for conformational analysis was further corroborated by a series of studies on interactions of achiral, Aib/Z Phe-based, helical peptides with external chiral molecules (“noncovalent chiral domino effect”) [97, 99–114] (Figure 16.20). The ECD properties of the H–L–Glu(l-Lys)–Z Phe–OH amphiphilic dipeptides self-assembled into nanovesicles were investigated [121]. Two hexapeptides, each characterized by two Phe residues (either in the E - or Z configuration) at positions 2 and 5 in the sequence, were studied by ECD spectroscopy [115] (Figure 16.22). Interestingly, it was found that in solvents of low polarity, where their conformational freedom is restricted and a rigid helical structure is attained, the two ECD spectra are almost mirror images. ECD spectra were calculated in the far-

562

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

[θ] × 10–4 (deg × cm2 × dmol–1)

20

10

0

–10

Figure 16.23. Theoretical near-UV ECD spectrum of the dodecapeptide

–20 240

280 Wavelength (nm)

320

Boc–(L-Ala–Z Phe–Aib)4 –OMe in the right-handed 310 -helix with φ, ψ torsion angles −54◦ , −28◦ . (Adapted from reference 99.)

and near-UV regions for helical, Z Phe-containing, peptides [99, 100]. The simulations were performed for various conformers that differ in helix (310 or α) type, helix screw sense, and Z Phe side-chain orientation. An example is reported in Figure 16.23. The theoretical ECD-peptide conformation relationships provided useful guidelines for determination of the helix sense in the Z Phe peptides and for the estimation of their statistically averaged conformations in solution. Other α,β-didehydroaromatic residues, the ECD properties of which were studied both experimentally and theoretically, include α,β-didehydro-(1-pyrenyl)alanine and α,β-didehydro-(4,4 -biphenyl)alanine [104, 114]. The observed ECD curves were interpreted on the basis of the exciton chirality method. The ECD spectra of the terminally blocked dipeptides Ac–l-Pro–l-Xxx–NHMe (NHMe, methylamino), where Xxx is an aliphatic α,β-unsaturated residue, i .e. Z Leu, Val, Z Abu (Abu, α-aminobutyric acid), and E Abu were recorded in different solvents [125, 128]. The curves, although not assignable to any common spectral class, must be attributed to peptides in a preferred type II β-turn conformation as determined for these compounds by use of other spectroscopic techniques. The spectra exhibit a remarkable solvent dependence and suggest an unordered conformation in aqueous solution. In contrast to the aliphatic AA mentioned above, the simplest AA of this subclass, Ala, is known to overwhelmingly prefer the flat, fully extended (C5 ), Y-conjugated conformation as established by solution and X-ray diffraction analyses of its homo-oligomers [22]. This is the reason why the Nα -protected Ala/chiral α-amino acid dipeptide amides investigated by ECD do not fold into a β-turn conformation [125, 127, 129]. It is worth mentioning that the UV absorption curves of the aliphatic AAcontaining peptides do not show separate bands for amide and C=C double bond electronic transitions. As a result, their ECD spectra do not exhibit any conformationally informative Cotton effect above 250 nm, in contrast to the properties of the aromatic AA-containing peptides.

16.7. POLY-β-PEPTIDES Among peptide foldamers, poly-β-peptides have special appeal because β-amino acids represent the next homologs of α-amino acids in the “backbone space” [130–135] (Figure 16.24). β-Amino acids are either achiral (β-alanine) or chiral. In the latter

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

(a)

NH

CH2

(b)

CH2

CO

(d)

Figure 16.24. Chemical formulae of (a) the acyclic, * HN

CH3 NH

(c)

CH2

CH *

CO

CH *

CH2

CO

unsubstituted, achiral residue β-alanine; (b, c) the acyclic, mono-substituted, chiral β 2 -residue β-aminoisobutyric acid (βAib) and β 3 -residue β-aminobutyric acid (βAbu), respectively; (d, e) the

(e) * HN

CH3 NH

* CO

* CO

cyclic, di-substituted, chiral β 2,3 -residues 2-aminocyclopentanecarboxylic acid (ACPC) and 2-aminocyclohexanecarboxylic acid (ACHC), respectively.

compounds, the chiral center can be found at the α-carbon (near the CO group; β 2 AA) or at the β-carbon (near the NH group; β 3 AA) or at both carbons (β 2,3 AA). The β 2,3 amino acids, which bear two chiral atoms, give rise to four diastereoisomers. Similarly to α-peptides, β-peptides are based on amide groups able to form stabilizing intramolecular C=O· · ·H–N H-bonds. A large body of work has laid a sufficiently solid conformational platform for spectroscopic investigations of β-peptides. Early studies of poly-β-peptides showed that these compounds are able to fold into helical structures, although details of the helix geometries were not elucidated [136–139]. Moreover, sheet structures were also proposed. An explosive breakthrough in this field came from the papers originating from the laboratories of Gellman and DeGrado and their collaborators, initially in the late 1990s [131,140–150], followed by those from Seebach and co-workers [133, 134, 151–158]. These articles include a huge amount of experimental ECD work. Results of theoretical ECD analyses have also been reported [140, 152, 159–162]. The synthesis of monodisperse oligomers of defined sequence enabled highresolution 3D-structural studies of β-peptides based on the conformationally constrained cyclic β-amino acids (S , S )- or (R, R)-trans-2-aminocyclohexanecarboxylic acid (ACHC) and (S , S )- or (R, R)-trans-2-aminocyclopentanecarboxylic acid (ACPC). The former compounds adopt the 14-helix conformation in the crystal state as well as in organic solvents. Also, series of β-peptides prepared from acyclic residues with a diverse collection of side chains fold into this type of helical structure. Depending on the chirality of the β-amino acids, either the left-handed or the right-handed 14-helix is generated. β-Peptides rich in β 3 -amino acids derived from naturally occurring (S )-amino acids adopt left-handed 14-helices. The 14-helix (Figure 16.25) is stabilized by H-bonding between an Ni –Hi group and a Ci +2 =Oi +2 group, forming a succession of 14-membered pseudo-rings. This 3D˚ versus 2.2 A ˚ radius) and longer (1.56 A ˚ versus 1.50 A ˚ structure is slightly wider (2.7 A rise per residue) than the α-peptide α-helix. While the α-helix has a 3.6 residue repeat, the 14-helix repeat is approximately every three residues (therefore, the alternative notation 314 -helix is often used), which positions the side chain of each residue directly atop one another along one face of the helix. The φ, θ , ψ backbone torsion angles are about −135◦ , 60◦ , −140◦ . The amide C=O and N–H groups project toward the N- and Ctermini, respectively, resulting in a net dipole opposite to that of the α-helix. The ECD spectra of several β-peptides that adopt the 14-helix, as determined by NMR or crystallography, show a band near 195 nm and a band of opposite sign near 215 nm [131, 134, 141, 146, 148–150, 155, 163, 170, 171, 179, 181, 184] (Figures 16.26 and 16.27). The magnitude of the ellipticity at 215 nm varies somewhat from peptide to peptide. However, it is possible that some or all of these peptides are not fully helical. NMR spectroscopy would not be sensitive to a small amount of nonhelical structure, so

563

564

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 16.25. Left, the idealized 14-helix of the (S, S)-trans-ACHC homo-deca-β-peptide; center, the idealized α-helix of (L-Ala)10 ; right, the idealized 12-helix of the (S, S)-trans-ACPC homo-deca-β-peptide, as viewed along (top) and perpendicular to (bottom) the helix axis. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, N and H atoms are white.

long as it is in rapid exchange with the helix conformation. Consistent with this suggestion, the ellipticity is greatest for those peptides with the most conformationally restricted amino acids, which lead to minimal fraying of the ends of the helices. The contribution of the amide π → π * transition to the ECD spectrum of the 14-helix was calculated. Because of excitonic coupling, this transition is split in two orthogonally polarized bands at 194 and 204 nm. The higher-energy band is in good agreement with experiment. The observed value of 214 nm for the lower-energy band probably reflects the presence of an overlapping amide n → π * transition, centered at a slightly longer wavelength. The intensity of the ECD spectrum of the α-helix is known to depend on main-chain length, becoming more intense as the helix is lengthened. Similar behavior was found for the 14-helix. The ECD spectra of many 10- to 15-residue-long peptides, which were designed to adopt a 14-helical conformation, are more intense than those of their shorter counterparts. For example, the ellipticities of a series of amphiphilic β-peptides were examined in the presence of micelles, which strongly stabilize the 14-helical conformation. Their mean residue ellipticities increase in a length-dependent manner. Systematic conformational searches and molecular dynamics calculations of the ACPC versus the ACHC β-amino acid revealed inherent preferences for different helical conformations [131–135, 140,142–145, 147, 166, 175, 176]. The cyclohexyl ring of ACHC stabilizes the θ torsional angle to a value near ±60◦ , which specifically stabilizes the 14-helical conformation. The smaller ring size of ACPC biases θ toward larger values, generating a different helical form, the 12-helix, as the most favorable conformer (Figure 16.25). The 3D-structure of the 12-helix is stabilized by a series of H-bonds between an amide C=O group at position i and an amide N–H group at position i +3 in the sequence. The helix has approximately 2.5 residues/turn and shows the same polarity as the α-helix, with the amide N–H groups exposed from the N-terminus of the helix.

565

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

8

[θ]R × 10–3 (deg × cm2 × dmol–1)

4

0

–4

–8

Figure 16.26. ECD spectra of the 14-helical –12 200

220 Wavelength (nm)

240

H–(ACHC)n –NH2 (n = 5, dashed line; n = 6, full line) homo-β-oligopeptides in MeOH solution. All ACHC residues have the (S, S)-trans configuration. (Adapted from reference 163.)

The ability to switch between two completely different β-peptide helices by relatively modest alteration of residue structure calls attention to a significant difference between α-amino acids and β-amino acids as building blocks. The chemist can exert much greater control over the intrinsic secondary structural propensity of β-amino acid residues than is possible with α-amino acid residues. The prediction that homo-β-peptides of ACPC should form the 12-helix was born out in experimental studies, in which relatively short oligomers were shown to adopt the 12-helix conformation, both in organic solution and in the crystal state. In organic solvents, the conformation is so stable that it is observed in peptides containing as few as six ACPC residues. However, β-peptides consisting of this amino acid were not soluble in water. To address this limitation, the pyrrolidinyl β-amino acid trans-3-aminopyrrolidine-4-carboxylic acid (APC) was prepared and incorporated into β-peptides along with ACPC residues. ECD (Figure 16.28) studies indicated that oligomers with as few as four residues show substantial populations of the 12-helix in water and that the helical content increases with main-chain length. The ECD curves exhibit a band at about 220 nm, followed by a more intense band of opposite sign near 200 nm. The zero-crossing is seen in the vicinity of 213 nm. Theoretical calculations indicated that the amide π → π * contribution to the ECD spectrum of the 12-helix should be similar in shape to that of the 14-helix but that the sign should be reversed

566

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a) O

H2N

N H

N H

O

O

O

O

N H

N H

N H

COOH

(b)

[θ] × 10–3 (deg × cm2 × dmol–1)

80

40

0

–40 200

220 240 Wavelength (nm)

Figure 16.27. Chemical formula (a) and ECD spectrum (b) of a 14-helical hexapeptide based on acyclic, mono-substituted, β 3 -amino acid building blocks in MeOH solution. (Adapted from reference 134.)

for a given helical handedness and the splitting between the parallel and perpendicular bands should be greater. The experimental spectra observed for a hexamer that forms the 12-helix are consistent with this analysis, showing a band near 205 nm followed by a band of opposite sign at approximately 190 nm. Additionally, a band is observed near 220 nm, which is probably associated with the amide n → π * transition. The presence of a band at 200–205 nm, together with another band near 220 nm, was not observed in other secondary structures of β-peptides and may be diagnostic of the 12-helix. Homo-oligomers from the (R, S )-cis-ACPC β-residue were shown to adopt preferentially a six-strand, extended, polar conformation [164]. This secondary structure is stabilized by intra-residue electrostatic interactions between the N–H and C=O groups, which form weak H-bonds in solution and lead to a six-membered H-bonded pseudoring structure (C6 ) (Figure 16.29). The ECD spectrum of the longest β-peptide of this series, the homo-heptamer, has a negative Cotton effect at 203 nm, the intensity of which decreases with shortening of the β-peptide backbone. β-Peptides with strictly alternating β 2 - and β 3 - (or β 3 - and β 2 -) mono-substituted residues tend to adopt the 10/12-helix conformation with 2.7 residues/turn [131, 134, 153, 154] (Figure 16.30). Depending on whether the sequence begins with a β 3 - or a β 2 -unit, the helix can start with a 10- or a 12-membered ring, respectively. This helix was studied by ECD spectroscopy. The spectrum shows a single, intense peak near 205

567

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

15

[θ]R × 10–3 (deg × cm2 × dmol–1)

10

5

0

Figure 16.28. ECD spectra of the 12-helical Ac–(APC–ACPC)2 – 4 –NH2 sequential β-peptide amides in H2 O solution. All ACPC β-amino acids residues have the (R, R)-trans configuration,

–5 200

220 Wavelength (nm)

(b)

(a) [θ] × 10–3 (deg × cm2 × dmol–1)

while that of all APC β-amino acids is (R, S)-trans. (Adapted from reference 147.)

240

0.0

–0.4

–0.8

H

–1.2

N

C

H

O

NH2 7

–1.6 200

220 240 Wavelength (nm)

Figure 16.29. ECD spectrum (a) of the homo-β-heptapeptide amide (intramolecularly H-bonded into a series of C6 forms, b) based on the (R, S)-cis-ACPC building block. (Adapted from reference 164.)

568

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a) O N H

O N H

O N H

O N H

O N H

O N H

O N H

O N H

O N H

(b)

10

12

[θ] × 10–3 (deg × cm2 × dmol–1)

(c)

80

60

40

20

0 220 240 200 Wavelength (nm)

Figure 16.30. (a) The amino acid sequence of the sequential β3 /β2 -nonapeptide studied. (b) 3D-Structure of the 10/12-helix. The H-atoms, except those of the amides, are omitted for clarity. (c) The ECD spectrum of the nonapeptide in MeOH solution. (Adapted from reference 153.)

nm (Figure 16.30). In this helix, the amides surrounded by methylenes form H-bonds to one another (i , i +2), generating the 10-membered rings, while the 12-atom rings are produced between amides surrounded by side chains (i +1, i +3). In contrast to the uniform alignment of amide bonds with the helical axis for the 14- and 12-helices, there are two types of amide bond orientations in the 10/12-helix. The 10-atom-ring amides are approximately perpendicular to the helical axis, while the 12-atom-ring amides are nearly aligned with the helical axis. This arrangement results in a smaller overall helix dipole compared to that of the other helical conformations. Interestingly, the dramatic solvent-dependent change observed in the ECD spectrum of a β-peptide was suggested to arise from an environmentally induced switch between the 10/12-helix (in H2 O) and the 14-helix (in MeOH). A comparable change in ECD arising from end-group deprotection was similarly rationalized. We have discussed above the detailed geometries and chirospectroscopic properties of the most common helical structures formed by poly-β-peptides. For additional, remarkable scientific contributions on those topics, see references 163–186. Other related, rapidly expanding areas, not mentioned here, include oligomers based on γ - and δ-amino acids, and various types of “hybrid” (α/β-; α/γ -; β/γ -)peptides. Alternating heterochiral oligo-β-peptides were also investigated [165, 167]. β-Strand-like conformations of β-peptides attracted the attention of structural biochemists and ECD theoreticians [169]. It was also found that in a limited number of cases a given ECD pattern can be induced by spatially different 3D-structures [152]. To gain more insight into the relationship between

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

β-peptide conformation and ECD properties, more accurate methods to calculate the ECD spectra for β-peptides are required. It is clear that the field of peptides based in part or fully on β-, γ -, and δ-amino acids is wide open and limited almost exclusively by scientists’ imagination.

ACKNOWLEDGMENTS The authors wish to thank Dr. R. W. Woody for critical reading of the manuscript.

REFERENCES 1. Houben-Weyl: Methods of Organic Chemistry, Vol. E 22c, Synthesis of Peptides and Peptidomimetics, M. Goodman, A. Felix, L. Moroder, and C. Toniolo, Eds., Thieme, Stuttgart, Germany, 2003. 2. P. S. Farmer, E. J. Ariens, Trends Pharmacol. Sci . 1982, 3 , 362–365. 3. A. F. Spatola, Chemistry and Biochemistry of Amino Acids, Peptides, and Proteins, Vol. 7, B. Weinstein, ed., Dekker, New York, 1983, pp. 267–357. 4. C. Toniolo, Int. J. Pept. Protein Res. 1990, 35 , 287–300. 5. J.-L. Fauch`ere, C. Thurieau, Adv. Drug Design 1992, 23 , 127–159. 6. A. Giannis, T. Kolter, Angew. Chem. Int. Edit. 1993, 32 , 1244–1267. 7. J. Gante, Angew. Chem. Int. Edit. 1994, 33 , 1699–1720. 8. R. M. J. Liskamp, Recl. Trav. Chim. Pays-Bas 1994, 113 , 1–19. 9. V. J. Hruby, G. Li, C. Haskell-Luevano, M. Shenderovich, Biopolymers 1997, 43 , 219–266. 10. Bioorg. Med. Chem. 1999, 7 , No. 1 (special issue), J. W. Kelly, ed. 11. Advances in Amino Acid Mimetics and Peptidomimetics, Vol. 2, A. Abell, ed., JAI Press, Stamford, CT, 1999. 12. Acc. Chem. Res. 2008, 41 , No. 10 (special issue), Y.-D. Wu and S. Gellman, eds. 13. A. Grauer, B. K¨onig, Eur. J. Org. Chem. 2009, 5099–5111. 14. G. H¨olzemann, Kontakte (Darmstadt) 1991, 55–63. 15. L. Pauling, R. B. Corey, Proc. Natl. Acad. Sci. USA 1951, 37 , 729–740. 16. S. Mizushima, T. Shimanouchi, M. Tsuboi, R. Souda, J. Am. Chem. Soc. 1952, 74 , 270–271. 17. M. T. Cung, M. Marraud, J. N´eel, Ann. Chim. (Paris) 1972, 7 , 183–209. 18. C. Toniolo, CRC Crit. Rev. Biochem. 1980, 9 , 1–44. 19. S. Hovm¨oller, T. Zhou, T. Ohlson, Acta Crystallogr. D 2002, 58 , 768–776. ˚ 20. Aberg, A. Yaremchuk, M. Tukalo, B. Rasmussen, S. Cusack, Biochemistry 1997, 36 , 3084–3094. 21. H. Birkedal, D. Schwarzenbach, Ph. Pattison, Chem. Commun. 2002, 2812–2813. 22. M. Crisma, F. Formaggio, C. Toniolo, T. Yoshikawa, T. Wakamiya, J. Am. Chem. Soc. 1999, 121 , 3272–3278. 23. C. Toniolo, E. Benedetti, Molecular Conformation and Biological Interactions, P. Balaram and S. Ramaseshan, eds., Indian Academy of Sciences, Bangalore, India, 1991, pp. 511–521. 24. M. Crisma, F. Formaggio, A. Moretto, C. Toniolo, Biopolymers (Pept. Sci.) 2006, 84 , 3–12. 25. N. Imawaka, M. Tanaka, H. Suemune, Helv. Chim. Acta 2000, 83 , 2823–2835. 26. M. Tanaka, Chem. Pharm. Bull. (Tokyo) 2007, 55 , 349–358. 27. C. Toniolo, M. Crisma, F. Formaggio, A. Moretto, C. Peggion, B. Kaptein, Q. B. Broxterman, in Peptides for Youth: The Proceedings of the 20th American Peptide Symposium, S. Del Valle, E. Escher, and W. D. Lubell, eds., Springer Science, New York, 2009, pp. 45–46.

569

570

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

44.

45. 46. 47.

48. 49. 50. 51. 52. 53. 54. 55. 56.

57. 58. 59. 60.

C. Toniolo, F. Formaggio, R. W. Woody, 2012, work in progress. M. Oba, M. Tanaka, M. Kurihara, H. Suemune, Helv. Chim. Acta. 2002, 85 , 3197–3214. E. Schr¨oder, K. Lubke, The Peptides, Vol. II, Academic Press, New York, 1966, pp. 397–423. C. Gilon, M. A. Dechantsreiter, F. Burkhart, A. Friedler, H. Kessler, Houben-Weyl: Methods of Organic Chemistry. Synthesis of Peptides and Peptidomimetics, Vol. E 22c, M. Goodman, A. Felix, L. Moroder, and C. Toniolo, eds., Thieme, Stuttgart, Germany, 2003, pp. 215–271. A. Moretto, M. Crisma, B. Kaptein, Q. B. Broxterman, C. Toniolo, Biopolymers (Pept. Sci.) 2006, 84 , 553–565. S. Zhang, S. Prabpai, P. Kongsaeree, P. I. Arvidsson, Chem. Commun. 2006, 497–499. J. E. Mark, M. Goodman, J. Am. Chem. Soc. 1967, 89 , 1267–1268. J. E. Mark, M. Goodman, Biopolymers 1967, 5 , 809–814. A. M. Liquori, P. De Santis, Biopolymers 1967, 5 , 815–820. W. L. Mattice, Macromolecules 1973, 6 , 855–858. V. Madison, J. Schellman, Biopolymers 1972, 11 , 1041–1076. M. Goodman, M. Fried, J. Am. Chem. Soc. 1967, 89 , 1264–1267. M. Goodman, F. Chen, F. R. Prince, Biopolymers 1973, 12 , 2549–2561. J. R. Spencer, N. G. J. Delaet, A. Toy-Palmer, V. V. Antonenko, M. Goodman, J. Org. Chem. 1993, 58 , 1635–1638. A. Cosani, M. Terbojevich, M. Palumbo, E. Peggion, Macromolecules 1979, 12 , 875–878. R. J. Simon, R. S. Kania, R. N. Zuckermann, V. D. Huebner, D. A. Jewell, S. Banville, S. Ng, L. Wang, S. Rosenberg, C. K. Marlowe, D. C. Spellmeyer, R. Tan, A. D. Frankel, D. V. Santi, F. E. Cohen, P. A. Bartlett, Proc. Natl. Acad. Sci. USA 1992, 89 , 9367–9371. L. S. Richter, D. C. Spellmeyer, E. J. Martin, G. M. Figliozzi, R. N. Zuckermann, Combinatorial Peptide and Nonpeptide Libraries, G. Jung, ed., VCD, Weinheim, Germany, 1996, pp. 387–404. S. A. Fowler, H. E. Blackwell, Org. Biomol. Chem. 2009, 7 , 1508–1524. C. W. Wu, K. Kirshenbaum, T. J. Sanborn, J. A. Patch, K. Huang, K. A. Dill, R. N. Zuckermann, A. E. Barron, J. Am. Chem. Soc. 2003, 125 , 13525–13530. K. Kirshenbaum, A. E. Barron, R. A. Goldsmith, Ph. Armand, E. K. Bradley, K. T. V. Truong, K. A. Dill, F. E. Cohen, R. N. Zuckermann, Proc. Natl. Acad. Sci. USA 1998, 95 , 4303–4308. F. A. Bovey, F. P. Hood, Biopolymers 1967, 5 , 325–326. J. A. Patch, A. E. Barron, J. Am. Chem. Soc. 2003, 125 , 12092–12093. S. L. Seurynck, J. A. Patch, A. E. Barron, Chem. Biol . 2005, 12 , 77–88. S. L. Seurynck-Servoss, M. T. Dohm, A. E. Barron, Biochemistry 2006, 45 , 11809–11818. J. Delsarte, G. Weill, Macromolecules 1974, 7 , 450–453. D. Grenier, R. E. Prud’homme, A. Leborgne, N. Spassky, J. Polym. Sci., Polym. Chem. Ed . 1981, 19 , 1781–1793. C. Aleman, B. Lotz, J. Puiggali, Macromolecules 2001, 34 , 4795–4801. M. Rueping, A. Dietrich, V. Buschmann, M. G. Fritz, M. Sauer, D. Seebach, Macromolecules 2001, 34 , 7042–7048. V. T. Ivanov, I. I. Mikhaleva, Houben-Weyl: Methods of Organic Chemistry, Vol. E 22c, M. Goodman, A. Felix, L. Moroder, C. Toniolo, eds., Thieme, Stuttgart, Germany, 2003, pp. 274–291. M. Dobler, Ionophores and Their Structures, John Wiley & Sons, New York, 1981. H. Oku, K. Yamada, R. Katakai, Biopolymers 2008, 89 , 270–283. M. Crisma, G. Valle, G. M. Bonora, C. Toniolo, G. Cavicchioni, Int. J. Pept. Protein Res. 1993, 41 , 553–560. G. Boussard, M. Marraud, J. N´eel, B. Maigret, A. Aubry, Biopolymers 1977, 16 , 1033–1052.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78.

79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95.

C. N. Pace, Nat. Struct. Mol. Biol . 2009, 16 , 681–682. J. Gao, D. A. Bosco, E. T. Powers, J. W. Kelly, Nat. Struct. Mol. Biol . 2009, 16 , 684–690. M. Goodman, C. Gilon, M. Palumbo, R. T. Ingwall, Israel J. Chem. 1974, 12 , 67–77. R. T. Ingwall, M. Goodman, Macromolecules 1974, 7 , 598–605. P. De Santis, A. J. Kovacs, Biopolymers 1968, 6 , 299–306. D. A. Brant, A. E. Tonelli, P. J. Flory, Macromolecules 1969, 2 , 228–235. A. E. Tonelli, Macromolecules 1992, 25 , 3581–3584. X. Yang, S. Kang, S. L. Hsu, H. D. Stidham, P. B. Smith, A. Leugers, Macromolecules 2001, 34 , 5037–5041. M. Goodman, M. D’Alagni, J. Polym. Chem. B, Polym. Lett. 1967, 5 , 515–521. C. Toniolo, V. Perciaccante, J. Falcetta, R. Rupp, M. Goodman, J. Org. Chem. 1970, 35 , 6–10. L. R. Caswell, M. F. Howard, T. M. Onisto, J. Org. Chem. 1976, 41 , 3312–3316. T. Kimura, T. Fukuda, S. Shimada, H. Matsuda, Chem. Lett. 2004, 33 , 608–609. A. B. Sugiharto, C. M. Johnson, I. E. Dunlop, S. Roke, J. Phys. Chem. 2008, C112 , 7531–7534. M. Lutz, J. van Krieken, Acta Crystallogr. 2010, C66 , o401–o405. R. T. Ingwall, C. Gilon, M. Goodman, Macromolecules 1976, 9 , 802–808. M. Goodman, W. Becktel, R. Katakai, G. Wouters, Makromol. Chem. 1981, Suppl. 4 , 100–115. W. J. Becktel, G. Wouters, D. M. Simmons, M. Goodman, Macromolecules 1985, 18 , 630–634. ¨ U. Schmidt, J. H¨ausler, E. Ohler, H. Poisel, Progress in the Chemistry of Organic Natural Products, Vol. 37, W. Herz, H. Grisebach, and G. W. Kirby, eds., Springer, Berlin, 1979, pp. 251–327. C. H. Stammer, Chemistry and Biochemistry of Amino Acids, Peptides, and Proteins, Vol. 6, B. Weinstein, ed., Dekker, New York, 1982, pp. 33–74. K. Noda, Y. Shimohigashi, N. Izumiya, The Peptides: Analysis, Synthesis, Biology, Vol. 5, E. Gross and J. Meienhofer, eds., Academic Press, New York, 1983, pp. 285–339. U. Schmidt, A. Lieberknecht, J. Wild, Synthesis 1988, 159–172. K. Uma, P. Balaram, Indian J. Chem. 1989, 28B , 705–710. T. P. Singh, P. Narula, H. C. Patel, Acta Crystallogr. 1990, B46 , 539–545. T. P. Singh, P. Jaur, Prog. Biophys. Mol. Biol . 1996, 66 , 141–165. R. Jain, V. S. Chauhan, Biopolymers (Pept. Sci.) 1996, 40 , 105–119. P. Mathur, S. Ramakumar, V. S. Chauhan, Biopolymers (Pept. Sci.) 2004, 76 , 150–161. O. Pieroni, A. Fissi, G. Montagnoli, Biopolymers 1973, 12 , 1445–1449. Y. Shimohigashi, J. W. Dunning, Jr., M. D. Grim, C. H. Stammer, J. Chem. Soc., Perkin Trans. II 1981, 1171–1175. O. Pieroni, G. Montagnoli, A. Fissi, S. Merlino, F. Ciardelli, J. Am. Chem. Soc. 1975, 97 , 6820–6826. O. Pieroni, A. Fissi, S. Merlino, F. Ciardelli, Israel J. Chem.. 1976/77, 15 , 22–28. O. Pieroni, A. Fissi, G. Montagnoli, F. Ciardelli, Biopolymers 1977, 16 , 1677–1686. O. Pieroni, A. Fissi, C. Pratesi, P. A. Temussi, F. Ciardelli, J. Am. Chem. Soc. 1991, 113 , 6338–6340. O. Pieroni, A. Fissi, C. Pratesi, P. A. Temussi, F. Ciardelli, Biopolymers 1993, 33 , 1–10. A. Tuzi, M. R. Ciajolo, G. Guarino, P. A. Temussi, A. Fissi, O. Pieroni, Biopolymers 1993, 33 , 1111–1121. O. Pieroni, A. Fissi, R. M. Jain, V. S. Chauhan, Biopolymers 1996, 38 , 97–108.

571

572

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

96. A. Tuzi, M. R. Ciajolo, D. Picone, O. Crescenzi, P. A. Temussi, A. Fissi, O. Pieroni, J. Pept. Sci . 1996, 2 , 47–58. 97. H. Miyake, H. Kamon, I. Miyahara, H. Sugimoto, H. Tsukube, J. Am. Chem. Soc. 2008, 130 , 792–793. 98. Y. Demizu, N. Yamagata, Y. Sato, M. Doi, M. Tanaka, H. Okuda, M. Kurihara, J. Pept. Sci . 2010, 16 , 153–158. 99. Y. Inai, T. Ito, T. Hirabayashi, K. Yokota, Biopolymers 1993, 33 , 1173–1184. 100. Y. Inai, Y. Kurokawa, T. Hirabayashi, Macromolecules 1999, 32 , 4575–4581. 101. Y. Inai, Y. Kurokawa, T. Hirabayashi, Biopolymers 1999, 49 , 551–564. 102. Y. Inai, K. Tagawa, A. Takasu, T. Hirabayashi, T. Oshikawa, M. Yamashita, J. Am. Chem. Soc. 2000, 122 , 11731–11732. 103. Y. Inai, T. Oshikawa, M. Yamashita, I. Kato, T. Hirabayashi, Chem. Lett. 2001, 472–473. 104. Y. Inai, T. Oshikawa, M. Yamashita, T. Hirabayashi, Y. Kurokawa, Bull. Chem. Soc. Jpn. 2001, 74 , 959–966. 105. Y. Inai, Y. Kurokawa, N. Kojima, J. Chem. Soc., Perkin Trans. 2 , 2002, 1850–1857. 106. Y. Inai, Y. Ishida, K. Tagawa, A. Takasu, T. Hirabayashi, J. Am. Chem. Soc. 2002, 124 , 2466–2473. 107. Y. Inai, H. Komori, Biomacromolecules 2004, 5 , 1231–1240. 108. N. Ousaka, Y. Inai, T. Okabe, Biopolymers 2006, 83 , 337–351. 109. Y. Inai, N. Ousaka, Biopolymers 2006, 82 , 471–481. 110. H. Komori, Y. Inai, J. Phys. Chem. 2006, A110 , 9099–9107. 111. N. Ousaka, Y. Inai, J. Am. Chem. Soc. 2006, 128 , 14736–14737. 112. H. Komori, Y. Inai, J. Org. Chem. 2007, 72 , 4012–4022. 113. N. Ousaka, Y. Inai, R. Kuroda, J. Am. Chem. Soc. 2008, 130 , 12266–12267. 114. N. Ousaka, Y. Inai, J. Org. Chem. 2009, 74 , 1429–1439. 115. R. Latajka, M. Jewginski, M. Makowski, A. Krezel, S. Paluch, Biopolymers 2008, 89 , 691–699. 116. R. Jain, M. Singh, V. S. Chauhan, Tetrahedron 1994, 50 , 907–920. 117. R. M. Jain, K. R. Rajashankar, S. Ramakumar, V. S. Chauhan, J. Am. Chem. Soc. 1997, 119 , 3205–3211. 118. U. A. Ramagopal, S. Ramakumar, P. Mathur, R. Joshi, V. S. Chauhan, Protein Eng. 2002, 15 , 331–335. 119. Rudresh, S. Ramakumar, U. A. Ramagopal, Y. Inai, S. Goel, D. Sahal, V. S. Chauhan, Structure 2004, 12 , 389–396. 120. P. Mathur, U. A. Ramagopal, S. Ramakumar, N. R. Jagannathan, V. S. Chauhan, Biopolymers (Pept. Sci.) 2006, 84 , 298–309. 121. A. Mishra, J. J. Panda, A. Basu, V. S. Chauhan, Langmuir 2008, 24 , 4571–4576. 122. A. K. Padyana, S. Ramakumar, P. Mathur, N. R. Jagannathan, V. S. Chauhan, J. Pept. Sci . 2009, 9 , 54–63. 123. D. H. Rich, P. K. Bhatnagar, J. Am. Chem. Soc. 1978, 100 , 2212–2218. 124. C. Shin, A. Kisuno. Y. Yonezawa, Chem. Lett. 1988, 1469–1472. 125. M. Lisowski, G. Pietrzynski, B. Rzeszotarska, Int. J. Pept. Protein Res. 1993, 42 , 466–474. 126. A. Anantharaman, D. Sahal, J. Med. Chem. 2010, 53 , 6079–6088. 127. C. Toniolo, E. Benedetti, Trends Biochem. Sci . 1991, 16 , 350–353. 128. G. Pietrzynski, B. Rzeszotarska, E. Ciszak, M. Lisowski, Z. Kubica, G. Boussard, Int. J. Pept. Protein Res. 1996, 48 , 347–356. 129. A. Gupta, V. S. Chauhan, Biopolymers 1990, 30 , 395–403. 130. S. H. Gellman, Acc. Chem. Res. 1998, 31 , 173–180.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

131. R. P. Cheng, S. H. Gellman, W. F. DeGrado, Chem. Rev . 2001, 101 , 3219–3232. 132. G. Lelais, D. Seebach, Biopolymers (Pept. Sci.) 2004, 76 , 206–243. ˇ 133. D. Seebach, T. Kimmerlin, R. Sebesta, M. A. Campo, A. K. Beck, Tetrahedron 2004, 60 , 7455–7506. 134. D. Seebach, A. K. Beck, D. J. Bierbaum, Chem. Biodivers. 2004, 1 , 1111–1239. 135. D. Seebach, D. F. Hook, A. Gl¨attli, Biopolymers (Pept. Sci.) 2006, 84 , 23–37. 136. J. M. Fernandez-Santin, J. Aymami, A. Rodriguez-Galan, S. Mu˜noz-Guerra, J. A. Subirana, Nature 1984, 311 , 53–54. 137. J. M. Fernandez-Santin, S. Mu˜noz-Guerra, A. Rodriguez-Galan, J. Aymami, J. Lloveras, J. A. Subirana, E. Giralt, M. Ptak, Macromolecules 1987, 20 , 62–68. 138. J. Bella, C. Aleman, J. M. Fernandez-Santin, C. Alegre, J. A. Subirana, Macromolecules 1992, 25 , 5225–5230. 139. F. Chen, G. Lepore, M. Goodman, Macromolecules 1974, 7 , 779–783. 140. J. Applequist, K. A. Bode, D. H. Appella, L. A. Christianson, S. H. Gellman, J. Am. Chem. Soc. 1998, 120 , 4891–4892. 141. D. H. Appella, L. A. Christianson, I. L. Karle, D. R. Powell, S. H. Gellman, J. Am. Chem. Soc. 1999, 121 , 6206–6212. 142. D. H. Appella, L. A. Christianson, D. A. Klein, M. R. Richards, D. R. Powell, S. H. Gellman, J. Am. Chem. Soc. 1999, 121 , 7574–7581. 143. Y. Hamuro, J. P. Schneider, W. F. DeGrado, J. Am. Chem. Soc. 1999, 121 , 12200–11201. 144. R. P. Cheng, W. F. DeGrado, J. Am. Chem. Soc. 2001, 123 , 5162–5163. 145. D. Liu, W. F. DeGrado, J. Am. Chem. Soc. 2001, 123 , 7553–7559. 146. R. F. Epand, T. L. Raguse, S. H. Gellman, R. M. Epand, Biochemistry 2004, 43 , 9527–9535. 147. X. Wang, J. F. Espinosa, S. H. Gellman, J. Am. Chem. Soc. 2000, 122 , 4821–4822. 148. M. Lee, T. L. Raguse, M. Schinneri, W. C. Pomerantz, X. Wang, P. Wipf, S. H. Gellman, Org. Lett. 2007, 9 , 1801–1804. 149. W. C. Pomerantz, T. L. R. Grygiel, J. R. Lai, S. H. Gellman, Org. Lett. 2008, 10 , 1799–1802. 150. E. Vat, W. C. Pomerantz, M. Geyer, S. H. Gellman, L. Brunsveld, ChemBioChem 2008, 9 , 2254–2259. 151. D. Seebach, J. V. Schreiber, P. I. Arvidsson, J. Frackenpohl, Helv. Chim. Acta 2001, 84 , 271–279. 152. A. Gl¨attli, X. Daura, D. Seebach, W. F. van Gunsteren, J. Am. Chem. Soc. 2002, 124 , 12972–12978. 153. M. Rueping, J. V. Schreiber, G. Lelais, B. Jaun, D. Seebach, Helv. Chim. Acta 2002, 85 , 2577–2593. 154. G. Lelais, D. Seebach, B. Jaun, R. I. Mathad, O. Fl¨ogel, F. Rossi, M. Campo, A. Wortmann, Helv. Chim. Acta 2006, 89 , 361–403. 155. R. I. Mathad, B. Jaun, O. Fl¨ogel, J. Gardiner, M. L¨oweneck, J. D. C. Cod´ee, P. H. Seeberger, D. Seebach, M. K. Edmonds, F. H. M. Graichen, A. D. Abell, Helv. Chim. Acta 2007, 90 , 2251–2273. 156. D. Seebach, S. Abele, K. Gademann, G. Guichard, T. Hintermann, B. Jaun, J. L. Mathews, J. V. Schreiber, L. Oberer, U. Hommel, H. Widmer, Helv. Chim. Acta 1998, 81 , 932–982. 157. M.-O. Ebert, J. Gardiner, S. Ballet, A. D. Abell, D. Seebach, Helv. Chim. Acta 2009, 92 , 2643–2658. 158. J. V. Schreiber, D. Seebach, Helv. Chim. Acta 2000, 83 , 3139–3152. 159. K. A. Bode, J. Applequist, Macromolecules 1997, 30 , 2144–2150. 160. J. Applequist, K. A. Bode, J. Phys. Chem. 2000, A104 , 7129–7132. 161. J. Applequist, J. Phys. Chem. 2000, A104 , 7133–7139.

573

574

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

162. X. Daura, D. Bakowies, D. Seebach, J. Fleischhauer, W. F. van Gunsteren, P. Kr¨uger, Eur. Biophys. J . 2003, 32 , 661–670. 163. A. Hetenyi, I. M. Mandity, T. A. Martinek, G. K. Toth, F. F¨ulop, J. Am. Chem. Soc. 2005, 127 , 547–553. 164. T. A. Martinek, G. K. Toth, E. Vass, M. Hollosi, F. F¨ulop, Angew. Chem. Int. Edit. 2002, 41 , 1718–1721. 165. T. A. Martinek, I. M. Mandity, L. F¨ulop, G. K. Toth, E. Vass, M. Hollosi, E. Forro, F. F¨ulop, J. Am. Chem. Soc. 2006, 128 , 13539–13544. 166. A. Hetenyi, Z. Szakonyi, I. M. Mandity, E. Szolnoki, G. K. Toth, T. A. Martinek, F. F¨ulop, Chem. Commun. 2009, 177–179. 167. G. V. M. Sharma, K. R. Reddy, P. R. Krishna, A. R. Sankar, K. Narsimulu, S. K. Kumar, P. Jayaprakash, B. Jagannadh, A. C. Kunwar, J. Am. Chem. Soc. 2003, 125 , 13670–13671. 168. G. V. M. Sharma, K. R. Reddy, P. R. Krishna, A. R. Sankar, P. Jayaprakash, B. Jagannadh, A. C. Kunwar, Angew. Chem. Int. Edit. 2004, 43 , 3961–3965. 169. S. Chandrasekhar, B. N. Babu, A. Prabhakar, A. Sudhakar, M. S. Reddy, M. U. Kiran, B. Jagadeesh, Chem. Commun. 2006, 1548–1550. 170. A. S. Norgren, P. I. Arvidsson, Org. Biomol. Chem. 2005, 3 , 1359–1361. 171. A. S. Norgren, P. I. Arvidsson, J. Org. Chem. 2008, 73 , 5272–5278. 172. A. Martinez de Ilarduya, C. Alem´an, M. Garc´ıa-Alvarez, F. Lopez-Carrasquero, S. Mu˜nozGuerra, Macromolecules 1999, 32 , 3257–3263. 173. M. G. Garc´ıa-Martin, M. Garc´ıa-Alvarez, A. Martinez de Ilarduya, J. A. Galbis, S. Mu˜nozGuerra, Biopolymers 2005, 77 , 121–127. 174. E. Torres, E. Gorrea, K. K. Burasco, E. Da Silva, P. Nolis, F. Rua, S. Boussert, I. Diez-Perez, S. Dannenberg, S. Izquierdo, E. Giralt, C. Jaime, V. Branchadell, R. M. Ortu˜no, Org. Biomol. Chem. 2010, 8 , 564–575. 175. Y. Imamura, N. Watanabe, N. Umezawa, T. Iwatsubo, N. Kato, T. Tomita, T. Iguchi, J. Am. Chem. Soc. 2009, 131 , 7353–7359. 176. G. L. Simpson, A. H. Gordon, D. M. Lindsay, N. Pramsawar, M. P. Crump, K. Mulholland, B. R. Hayter, T. Gallagher, J. Am. Chem. Soc. 2006, 128 , 10638–10639. 177. E. Vaz, L. Brunsveld, Org. Lett. 2006, 8 , 4199–4202. 178. J. Cheng, T. J. Deming, Macromolecules 2001, 34 , 5169–5174. 179. S. A. Hart, A. B. F. Bahadoor, E. E. Matthews, X. J. Qiu, A. Schepartz, J. Am. Chem. Soc. 2003, 125 , 4022–4023. 180. J. A. Kritzer, J. D. Lear, M. E. Hodsdon, A. Schepartz, J. Am. Chem. Soc. 2004, 126 , 9468–9469. 181. J. A. Kritzer, J. Tirado-Rives, S. A. Hart, J. D. Lear, W. L. Jorgensen, A. Schepartz, J. Am. Chem. Soc. 2005, 127 , 167–168. 182. J. A. Kritzer, N. W. Luedtke, E. A. Harker, A. Schepartz, J. Am. Chem. Soc. 2005, 127 , 14584–14585. 183. J. A. Kritzer, O. M. Stephens, D. A. Guarracino, S. K. Reznik, A. Schepartz, Bioorg. Med. Chem. 2005, 13 , 11–16. 184. D. A. Guarracino, H. R. Chiang, T. N. Banks, J. D. Lear, M. E. Hodsdon, A. Schepartz, Org. Lett. 2006, 8 , 807–810. 185. E. Abraham, C. W. Bailey, T. D. W. Claridge, S. G. Davies, K. B. Ling, B. Odell, T. L. Rees, P. M. Roberts, A. J. Russell, A. D. Smith, L. J. Smith, H. R. Storr, M. J. Sweet, A. L. Thompson, J. E. Thomson, G. E. Tranter, D. J. Watkin, Tetrahedron: Asymmetry 2010, 21 , 1797–1815. 186. C. Fernandes, S. Faure, E. Pereira, V. Th´ery, V. Declerck, R. Guillot, D. J. Aitken, Org. Lett. 2010, 12 , 3606–3609.

17 CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS ´ Klara ´ Bednaˇ ´ rova, ´ and Michaela Vorl´ıcˇ kova´ Jaroslav Kypr, Iva Kejnovska,

17.1. INTRODUCTION Circular dichroism (CD) arises from differential absorption of right-handed and left-handed circularly polarized light by chiral molecules. This phenomenon has been described in detail in previous books (e.g., reference 1) and reviews [2–6; see Chapter 18, this volume]. In nucleic acids there are three sources of chirality. First is the asymmetric sugar (especially position C1 ); this chirality causes monomeric nucleosides to exhibit CD. The second source is the helicity of the secondary structures adopted by nucleic acids. The third source of CD results from long-range tertiary ordering of DNA in some environments. CD of monomeric constituents of nucleic acids and short single-stranded fragments were described previously [2]. The theory of CD is well-developed [1] and complex. Nevertheless, the use of CD spectroscopy to elucidate nucleic acid secondary structure is mainly based on empirical grounds. Conventional CD spectroscopy operates within the spectral range of about 200 nm to 320 nm. For these measurements, conventional spectrometers are used. CD spectroscopy is even more sensitive and informative in the far UV region below 200 nm, but these measurements are difficult to perform and the specialized instruments required are expensive [2]. CD spectra of nucleic acids can also be measured in the infrared region (Chapters 18, 22, and 23, this volume), but here the method is much less sensitive. In this chapter we will focus on CD results obtained in the 200- to 320-nm range, the range mostly used to study secondary structures of nucleic acids.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

575

576

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Nucleic acids (NA) are of two types, DNA and RNA. They mainly differ by the type of constituent sugar, which is deoxyribose in DNA and ribose in RNA. DNA generally contains adenine, guanine, thymine, and cytosine. RNA contains uracil instead of thymine and other noncanonical bases. In vivo, RNA strands are much shorter than DNA and the two molecules also differ in secondary structure. DNA is mostly double-stranded. In RNAs, helices, bulges, loops, and mismatches contribute to complex tertiary structures. CD spectroscopy is more useful in studies of DNA than RNA. Below we will review problems where CD spectroscopy contributed to the understanding of the conformational polymorphism of DNA.

17.2. DENATURED DNA, B-FORM AND THE HELIX-COIL TRANSITION CD spectra of denatured DNAs are weak because the strands are mostly disordered and exhibit little chirality. The B-form helix of natural DNAs (i.e., the basic ordered double-stranded conformation) provides a rather weak CD spectrum containing a positive band around 275 nm and a negative band at around 245 nm (Figure 17.1, left). The bands have about the same intensity and the integral of the spectral curve is close to zero. We call such a spectrum conservative. The base pairs are more or less perpendicular to the long axis of the double helix, which introduces relatively weak chirality into the structure. Furthermore, contributions to the CD spectrum from different regions of the compositionally heterogeneous parts of the molecules mutually compensate. CD spectra of helices formed by synthetic DNA molecules are stronger and do have characteristics dependent on nucleotide sequence [6]. This is caused not only by different chromophores but also by rather different B-family conformations. CD spectra of native and denatured synthetic oligo- and polynucleotides differ from one another, depending on the variants of the B-type conformation adopted. CD spectroscopy thus allows one to follow the helix–coil transition of DNAs (Figure 17.1, right). The temperature dependence monitored at a selected wavelength can be used to characterize the stability of the DNA structure. The plot of CD intensity versus temperature results in a characteristic S-shaped curve reflecting cooperativity of the helix–coil transition. The presence of isoelliptic points in the spectra is indicative of the two-state nature of the transition. The course of the melting enables determination of the melting point (Tm ) and thermodynamic parameters of the studied structure.

17.3. THE B-FORM AND THE HAIRPIN A hairpin arises from folding back of a DNA single strand on itself to form a helix capped at one end with a loop of single-stranded residues. For hairpin formation to occur, the DNA sequence must contain at least approximately a dyad symmetry. The B-form doublestranded structure-to-hairpin transition is also reflected by CD spectroscopy. We illustrate this with a (G + C)-rich DNA fragment (Figure 17.2). For this sequence, electrophoresis confirmed that the duplex-to-hairpin transition occurred [6]. It is noteworthy that the transition is often irreversible, indicating a large kinetic barrier between the two-stranded form and the hairpin.

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

Figure 17.1. B-form DNA spectrum and the helix–coil transition. Left: CD spectrum of B-form DNA from calf thymus (42% G + C). The spectrum was measured in 10 mM sodium acetate, pH 7. Spectra of the natural DNAs were measured on a Roussel–Jouan dichrograph, Model CD 185. Spectra of the polynucleotides in this and in the following figures were measured on the Jobin–Yvon dichrograph Mark VI and, unless stated otherwise, in 1-cm cells (absorption ∼ 0.7) at room temperature. CD in all figures is expressed as ε (in M−1 cm−1 ), molarity being related to nucleoside residues in the DNA samples. Sketches of the particular DNA structures were taken from the NDB database. Right: CD spectra of poly(dA-dT) in 10 mM sodium acetate, pH 7 at 19◦ C (dash). The spectra reflecting a helix-coil transition were taken at 21.9◦ C (dash–dot), 24.3◦ C (long dash) and 26.6◦ C (solid line). Insert: Temperature-induced changes in poly(dA-dT) monitored at (circles) 220 nm and (squares) 262 nm.

17.4. THE A-FORM AND THE B–A TRANSITION The A-form of DNA provides a much stronger CD spectrum than the B-form (Figure 17.3). The A-form spectrum is dominated by a strong positive band at 260 nm and a strong negative band at 210 nm. These spectral features probably originate from base-pair tilting that is characteristic of the A-form. A low-ionic strength DNA solution (∼1–5 mM) can be transformed from the B-form to the A-form by addition of ethanol [7] or other agents. Higher salt concentrations result in precipitation of DNA or cause its psi condensation (see below). The B–A transition is highly cooperative, is reversible, and has fast kinetics, and the CD spectra recorded during the transition intersect in isoelliptic points. RNA provides a CD spectrum similar to the A-form DNA of the

577

578

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 17.2. Temperature-induced irreversible duplex-hairpin transition. The spectrum of d(C6 –G6 ) was measured in 1mM sodium phosphate, 0.3 mM EDTA, pH 7. Left: CD spectra measured at 0◦ C corresponding to duplex before denaturation (dashed line) and hairpin after denaturation (solid line). Right: Spectral changes induced by increasing (open symbols) and decreasing (filled symbols) temperatures monitored at 260 nm (squares) and 280 nm (triangles). CD spectra were measured in 0.1 cm cells. (Redrawn from reference 6.)

corresponding nucleotide sequence (Figure 17.3). CD spectroscopy has proven to be a useful tool to study the B–A transition in various DNAs.

17.5. THE B/A CONFORMATION OF (dC)n · (dG)n DNA SEQUENCES The (dC)n · (dG)n sequences of DNA adopt an unusual B-conformation. A CD spectrum of the self-complementary sequence d(C6 –G6 ) is shown in Figure 17.2. It has a distinct positive CD band at 260 nm, similar to the A-form, even in the absence of alcohol. Combined CD spectroscopy, molecular dynamics, and NMR studies showed that this sequence adopts an A-like structure with B-like sugar puckering [8]. This and similar sequences undergo a cooperative and two-state transition to A-form induced by alcohols. Even the reverse self-complementary sequences d(Gn –Cn ) have A-like features. The (dG)n sequence in the 5 half of the molecule is A-like, whereas the 3 (dC)n half is B-like [9]. These findings extended our understanding of the conformational polymorphism of DNA.

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

Figure 17.3. B–A transition of DNA. Left panel: CD spectra reflecting trifluorethanol-induced B–A transition of d(GCGGCGACTGGTGAGTACGC) duplex with its complementary strand: 0% TFE (dashed), 80% TFE (solid) lines. Insert: The transition monitored at 266 nm. Right panel: CD spectra of RNA of the same sequence (U instead of T) duplexed with a complementary DNA strand: 0% TFE (dashed), 80% TFE (solid) lines. (Redrawn from reference 6.)

17.6. THE Z-FORMS AND THE B–Z TRANSITION Like the B–A transition, the B–Z transition was first detected in solution by CD spectroscopy [10]. The Z-form is a left-handed helix. The transition is highly cooperative, and it has a slow kinetics caused by the need for base-pair flipping. It is specific for alternating pyrimidine–purine, namely dC–dG, sequences, and it is facilitated by the presence of methyl5 dC [11]. The CD spectrum of the Z-form is more or less a mirror image of that of the B-form (Figure 17.4), although variants of the Z-form (called Z ) have rather different spectra [6]. The B–Z transition is also reflected by UV absorption spectroscopy (Figure 17.4, upper insert).

17.7. GUANINE QUADRUPLEXES Guanine quadruplexes are interesting alternative arrangements to the Watson–Crick double helix. These structures are based on guanine tetrads held together by Hoogsteen

579

580

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 17.4. CD spectra reflecting trifluorethanol-induced B–Z transition of poly(dG–dC) duplex. Spectra were measured in 59% TFE (dashed) and 67% TFE (solid) lines. TFE was added to the DNA dissolved in 1 mM sodium phosphate, 0.3 mM EDTA, pH 7, and CD spectra were measured at 0◦ C. Inserts: The transition monitored at 291 nm. UV absorption spectra measured at the same conditions as the CD measurements. (Redrawn from reference 5.)

hydrogen bonds. The tetrads are stacked on top of one another, and a cation is inserted in the cavity between the neighboring tetrads. Thus, the stability of guanine quadruplexes is sensitive to the cation type. There are several types of guanine quadruplexes. The quadruplexes can be built by one, two, or four DNA molecules and the strand orientation can be parallel, antiparallel, or hybrid parallel–antiparallel. CD spectroscopy can distinguish among particular guanine quadruplex topologies (Figure 17.5). The spectrum of parallel quadruplexes contains a dominating positive CD band at 260 nm; antiparallel quadruplexes are characteristic by a positive CD band at 295 nm and a negative one around 260 nm. The bands at 260 and 295 nm probably reflect populations of anti and syn glycosidic angles of the dG residues in the particular quadruplex arrangements. The 260-nm band is similar in shape to that of the A-form, which indicates similar base stacking [12]. The quadruplex, however, has a positive band at 210 nm where the A-form has a negative band (compare Figures 17.3 and 17.5). Quadruplex formation is also reflected by UV absorption spectroscopy (Figure 17.5, insert). Study of quadruplex structures is important as the human genome contains thousands of sequences prone to formation of guanine quadruplexes under favorable conditions. CD spectroscopy, especially

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

Figure 17.5. CD spectra of guanine quadruplexes. Left: Time-dependent formation of a parallelstranded quadruplex of d(G4 ) stabilized by 16 mM K+ : spectrum immediately after K+ addition (dashed) and after 24 h (solid line). Right: Na+ -induced formation of an antiparallel bimolecular quadruplex of d(G4 T4 G4 ): 1 mM Na+ (dashed), 500 mM NaCl (solid line). Oligonucleotides were dissolved in 1 mM sodium phosphate, 0.3 mM EDTA, pH 7, thermally denatured (5 min at 90◦ C), and slowly cooled before measurements. Insert: UV absorption spectra of d(G4 T4 G4 ) measured at the same conditions as the CD measurements. (Redrawn from reference 6.)

in combination with gel electrophoresis, has already contributed significantly to our knowledge of various quadruplex topologies adopted by, for example, the sequences in human telomeres [13].

17.8. CYTOSINE QUADRUPLEXES DNA strands rich in cytosine also form quadruplexes. These consist of two parallel homoduplexes connected through hemi-protonated C · C+ pairs [14]. The two homoduplexes are mutually intercalated in an antiparallel orientation (Figure 17.6). The cytosine quadruplexes provide a characteristic CD spectrum dominated by a positive band at 290 nm (Figure 17.6). Formation of cytosine quadruplexes requires acidic pH, which is needed to protonate cytosine. Interestingly, some cytosine quadruplexes are stable even at pH 7.

581

582

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 17.6. Acid-induced transition of a C-rich oligonucleotide into a cytosine quadruplex. CD spectra and UV absorption spectra (insert) reflecting the acid-induced transition of a human telomere DNA fragment d[(C3 TAA)3 C3 ] into an intercalated cytosine quadruplex (CD dependence measured at 287 nm). The oligonucleotide was dissolved in 10 mM potassium phosphate, 0.1 M KCl, pH 7.1 (dashed line); the pH value was changed to pH 5 by addition of dilute HCl and the CD spectrum was measured (solid line). pH values were determined using a Sentron Red-Line electrode and a Sentron Titan pH meter.

17.9. DNA STRANDS RICH IN GUANINE AND ADENINE DNA strands rich in guanine and adenine can form various ordered structures that melt cooperatively. The conformers may be either (a) antiparallel duplexes, which provide CD spectra similar to B-form DNA [15], (b) parallel duplexes formed under physiological salt conditions and neutral pH, or (c) ordered single strands induced by acid pH [16], ethanol [17], or even dimethylsulfoxide [18]. The parallel duplex provides a CD spectrum with a dominating positive CD band at 260 nm, like the A-form of DNA or parallel guanine quadruplex, suggesting similar guanine–guanine stacking in various G-rich DNAs [19]. The CD spectrum of the ordered single strand is very similar to the spectrum of the homoduplex, indicating that the homoduplex arises though a dimerization of the ordered single strands [17].

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

17.10. EXTENSIVE CHANGES IN THE CIRCULAR DICHROISM OF POLY(dA-dT) Poly(dA-dT) provides a more or less B-like CD spectrum under physiological conditions, but the CD spectrum changes drastically at molar concentrations of CsF or alcohol in the presence of cesium ions as shown in Figure 17.7 [20]. We call this conformation the X form [21]. The X form is characterized by a very deep negative band in the longwavelength part of the CD spectrum. Remarkably, the changes are specific for CsF. For example, NaCl does not induce these changes. Ethanol induces a transition of poly(dAdT) into the A-form if sodium is the counterion, but the X-form is induced if cesium is the counterion [21]. The CsF-induced changes are accompanied by large changes in the phosphorus NMR spectrum of poly(dA-dT) (Figure 17.7, insert) [20]. Two widely separated resonances appear, as they do with the Z-form, but the X-form is not Z-form because assignment of the two resonances is opposite. A structure of (dA-dT)3 observed by X-ray that contains Hoogsteen base pairing [22] may be the crystal counterpart of the X-form.

17.11. POLY(AMINO2 dA-dT) AND POLY(dG-METHYL5 dC) Poly(amino2 dA-dT) and poly(dG-methyl5 dC) are much more similar DNAs than their designations might indicate (Figure 17.8, inserts). Both contain alternating purine–pyrimidine sequences, and both have amino groups in the double helix minor groove and methyl groups in the major groove. Poly(dG-methyl5 dC) undergoes the B–Z transition even under physiological conditions [11]. Poly(amino2 dA–dT), unlike poly(dA-dT), also undergoes a salt-induced transition under physiological conditions, and the resulting structure has features of both the A-form and X-form [23]. It is clear from these studies that both the amino group in the minor groove and the methyl group in the major groove strongly influence conformational behavior of the alternating purine–pyrimidine DNAs.

17.12. PSI DNA Under some conditions (PEG, ethanol with moderate salt concentrations, polylysine, etc.), DNA condenses into so-called psi DNA, a form that can be seen in the electron microscope. In the condensates, a long-range structure introduces chirality. In the CD spectra, psi DNA is characterized by huge positive or negative amplitudes and signal beyond 300 nm, where DNA bases no longer absorb light. The CD signal originates from differential scattering of light by helical arrangement of the condensates [24]. Psi DNA is an ordered structure, and it is considered to be a simple approximation of the DNA organization in chromosomes.

17.13. CONCLUSIONS This review summarized the significant role played by CD spectroscopy in the history of nucleic acid studies. We focused on problems where relevant CD studies have essentially been finished (B-form, A-form, Z-form) as well as on those that remain open to further

583

584

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

2

–3.6 –4.0 –4.4 –4.8 NMR chemical shift (ppm)

31P

Δe [M–1 m–1]

0

–2

–2 Δe275 nm –4

–4 –6

–6

–8

–10 –8 50 240

260

280

55 60 65 Methanol [%]

300

Wavelength [nm]

Figure 17.7. Methanol-induced transition of poly(dA–dT) into an X-form. CD spectra of poly(dA–dT) measured in 0.5 mM sodium phosphate, 3 mM CsCl, and 0% (dot), 57% (dash dot dot), 60% (short dash), 62% (dash dot), and 64% (solid line) methanol (v/v). Temperature was 0◦ C. Insert: The transition monitored by ε at 275 nm. The conditions are as above, but 0.002 mM CaCl2 was present in the methanol solution. Top: 31 P NMR spectrum of the X-form of poly(dA-dT) induced by 6.3 M CsF. (Redrawn from reference 21.)

studies like those of B/A-DNA, quadruplexes, X-form, structures of poly(amino2 dA-dT) or (G + A)-rich DNA sequences. CD spectroscopy does not provide molecular structures at atomic resolution, but it has many advantages: •

It is a simple, fast, and relatively cheap method. It is extremely sensitive and therefore requires small amounts of material. This assures solubility of the sample even under extreme conditions. • It works over a wide range of DNA concentrations, from one order of magnitude higher than those used for NMR measurements to very dilute solutions. • Long as well as short DNA molecules can be studied. •

585

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

(a)

Figure 17.8.

(b)

MgCl2 -induced B–Z and B–X transitions of poly(dG–methyl5 dC) and

2

poly(amino dA–dT). Polynucleotides were dissolved in 0.6 mM potassium phosphate, 0.03 mM EDTA, pH 6.8. Left: CD spectra taken over the course of the B–Z transition of poly(dG–methyl5 dC) measured 1, 4, 17, and 88 min (dashed to solid lines, respectively) after addition of MgCl2 to 0.05 mM concentration. Right: CD spectra reflecting the B–X transition of poly(amino2 dA–dT) in 0, 0.028, 0.056, 0.070, and 0.190 mM MgCl2 (from thin to solid lines as concentration is increased). Top: Sketches of G·methyl5 C and amino2 A · T base pairs. (Redrawn from reference 6.)

•

The method is used empirically, and dozens of molecules can be compared in a single study. • CD can be measured under various conditions of temperature, pH, organic solvents, salt types, and concentrations, which enables researchers to map the whole conformational space of the studied molecule. • Gradual changes within a single DNA conformation and cooperative changes between discrete conformational states can be distinguished. The two types of changes have distinct physical properties and biological relevance. Not a single band, but instead the whole spectrum, should be followed and taken into account in the interpretation of the CD spectra. In combination with simple UV absorption spectroscopy and gel electrophoresis, CD spectroscopy is a powerful complementary method to X-ray diffraction and NMR spectroscopy in studies of DNA.

586

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ACKNOWLEDGMENTS The authors thank Professor B. Woody for his valuable advice and comments. This work was supported by grant IAA500040903 and IAA100040701 kindly provided by the Grant Agency of the Academy of Sciences of the Czech Republic.

REFERENCES 1. N. Berova, K. Nakanishi, R. Woody, Circular Dichroism, Principles and Applications, 2nd ed., Wiley-VCH, New York 2000. 2. W. C. Johnson, CD of nucleic acids, in Circular Dichroism, Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. Woody, eds., Wiley-VCH, New York, 2000, pp. 703–718. 3. J. C. Maurizot, Circular dichroism of nucleic acids: Nonclassical conformations and modified oligonucleotides, in Circular Dichroism, Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. Woody, eds., Wiley-VCH, New York, 2000, pp. 719–739. 4. D. M. Gray, R. L. Ratliff, M. R. Vaughan, Meth. Enzymol . 1992, 211 , 389–406. 5. M. Vorlickova, J. Kypr, V. Sklenar, Nucleic acids: Spectroscopic methods, in Encyclopedia of Analytical Science, 2nd ed., P. J. Worsfold, A. Townshend, C. F. Poole, eds., Elsevier, Oxford, 2005, 6 , pp. 391–399. 6. J. Kypr, I. Kejnovska, D. Renciuk, M. Vorlickova, Nucl. Acids Res. 2009, 37 , 1713–1725. 7. V. I. Ivanov, L. E. Minchenkova, E. E. Minyat, M. D. Frank Kamenetskii, A. K. Schyolkina, J. Mol. Biol . 1974, 87 , 817–833. 8. L. Trantirek, R. Stefl, M. Vorlickova, J. Koca, V. Sklenar, J. Kypr, J. Mol. Biol . 2000, 297 , 907–922. 9. R. Stefl, L. Trantirek, M. Vorlickova, J. Koca, V. Sklenar, J. Kypr, J. Mol. Biol . 2001, 307 , 513–524. 10. F. M. Pohl, T. M. Jovin, J. Mol. Biol . 1972, 67 , 375–396. 11. M. Behe, G. Felsenfeld, Proc. Natl. Acad. Sci. USA 1981, 78 , 1619–1623. 12. J. Kypr, M. Fialova, J. Chladkova, M. Tumova, M. Vorlickova, Eur. Biophys. J . 2001, 30 , 555–558. 13. M. Vorlickova, J. Chladkova, I. Kejnovska, M. Fialova, J. Kypr, Nucl. Acids Res. 2005, 33 , 5851–5860. 14. M. Gueron, J. L. Leroy, Curr. Opin. Struct. Biology 2000, 10 , 326–331. 15. M. Ortiz-Lombardia, R. Eritja, F. Azor´ın, J. Kypr, I. Tejralova, M. Vorl´ıcˇ kova, Biochemistry 1995, 34 , 14408–14415. 16. N. G. Dolinnaya, J. R. Fresco, Progr. Nucl. Acids Res. 2003, 75 , 321–347. 17. M. Vorlickova, I. Kejnovska, J. Kovanda, J. Kypr, Nucl. Acids Res. 1999, 27 , 581–586. 18. J. Kypr, M. Vorlickova, Biopolymers 2001, 62 , 81–84. 19. J. Kypr, M. Vorlickova, Biopolymers 2002, 67 , 275–277. 20. M. Vorlickova, J. Kypr, V. Sklenar, J. Mol. Biol . 1983, 166 , 85–92. 21. M. Vorlickova, J. Kypr, J. Biomol. Struct. Dynam. 1985, 3 , 67–83. 22. N. G. A. Abrescia, A. Thompson, T. Huynh-Dinh, J. A. Subirana, Proc. Natl. Acad. Sci. USA 2002, 99 , 2806–2811. 23. M. Vorlickova, J. Sagi, A. Szabolcs, A. Szemzo, L. Otvos, J. Kypr, J. Biomol. Struct. Dynam. 1988, 6 , 503–510. 24. C. Bustamante, M.F., M. F. Maestre, I. Tinoco, Jr., J. Chem. Phys. 1980, 73 , 4273–4281.

18 ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES Roberto Corradini, Tullia Tedeschi, Stefano Sforza, and Rosangela Marchelli

18.1. INTRODUCTION Peptide nucleic acids (PNAs) are a class of oligonucleotide analogues with a polyamide backbone, depicted in Figure 18.1, first described by Nielsen et al. in 1991 [1, 2]. These molecules have been largely utilized in DNA and RNA recognition in a variety of applications, ranging from modulation of gene expression [3] to diagnostic methodologies [4], and have been proposed as robust materials for micro- and nanofabrication [5]. The interesting features of this class of compounds are: (a) the ability to interact with complementary sequences of DNA and RNA forming very stable duplexes, with thermal stability higher than the corresponding duplexes formed by DNA oligonucleotides [6], and even more stable PNA–PNA duplexes [7]; (b) the high sequence selectivity of the duplex formation, which again is superior to DNA oligonucleotides in the recognition of even a single mispairing of the bases [8, 9]; (c) the ability of poly-pyrimidine PNA to form PNA–DNA–PNA triplexes of remarkable stability, which can produce strandinvasion of duplex DNA [10]; and (d) the high stability in biological fluids, due to their unnatural skeleton, which prevents degradation by nucleases and proteases [11]. In their simplest version, PNAs have an achiral structure; yet it has been demonstrated that even in the achiral PNA–PNA duplexes they form helical (hence chiral) structures [12]. If a chiral bias is introduced, a preference for both right- or left-handed structures can be obtained. Thus these molecules are also a very interesting case of simplified nucleic acid structures in which the helicity can be modulated by an appropriate design. In this chapter we summarize the results on the electronic CD properties of these very interesting artificial biopolymers. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

587

588

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

B

B O

O *

O O

O

P O

*

*

N H

n

O

N

DNA

n

*

Figure 18.1. Structure of peptide nucleic acid (PNA) outlining their similarity to DNA.

PNA

The formation of PNA-containing structures with nucleic acids is governed by the Watson–Crick base pairing for double helix, and by both Watson–Crick and Hoogsteen hydrogen bonding in the case of triplex formation, as illustrated in Figure 18.2a. A peculiar feature of PNA is that both antiparallel and parallel complexes with DNA and RNA can be formed, conventionally described as depicted in Figure 18.2b. Although the H-bonding interactions, as well as the stacking interactions between adjacent base pairs, are the same occurring in DNA and RNA structures, PNA have a preferred helical structure (named P-form) which differs from that of DNA significantly, with 18 base pairs per turn and a smaller twist angle, as inferred from PNA–PNA solidstate studies (Figure 18.2c) [12]. The available PNA:DNA crystal structures [13, 14] show similar characteristics, with a pitch of 15.5 base pairs and a twist angle (23◦ ) significantly smaller than both B- and A-DNA (36◦ and 32.7◦ , respectively). Therefore, conformational analysis based on CD spectra by analogy with known DNA structures (A, B, Z, etc.) should be avoided or considered only tentative. A series of other techniques, such as NMR or induced circular dichroism, should be used in order to produce experimental evidence for the conformation proposed.

(a)

H N T

O H

N

O Hoogsteen H-bonds

H

N

N A N

N

H N

N C+ + N H O

O H

N

T

N

H N G

N

H

O

Hoogsteen H-bonds N

O Watson-Crick H-bonds

H

N

N

N

N H

(b)

N

H H

C

N

O Watson-Crick H-bonds

(c) Parallel

N-Term

C-Term PNA

5′

N-Term

3′ Antiparallel

3′

PNA

DNA

C-Term PNA 5′

DNA PNA

Figure 18.2. (a) Type of interactions formed by PNA: Watson–Crick for duplex formation and Hoogsteen in PNA–DNA–PNA triplexes. (b) Parallel and antiparallel orientation of PNA–DNA duplexes. (c) PNA:PNA duplex in the solid state, forming an helical P-helix. (Data from Protein Data Bank, Code 1PUP, from reference 12.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

Several different problems have been addressed using CD spectroscopy: (a) preorganization of PNA single strand; (b) formation of PNA–PNA duplexes and their preferred helical features; (c) formation of duplexes or triplexes with naturally occurring nucleic acids by conformational changes detected by CD; (d) kinetics of duplex and triplex formation; (e) thermal stability of the duplexes and triplexes and conformational changes occurring as a function of temperatures; (f) ability of achiral dyes to bind to PNA-containing duplexes. In this chapter we will show the techniques and the most significant results in this field, with some specific examples illustrating the approaches mostly used for the study of PNA properties. Related subjects on the study of biopolymers by Electronic Circular Dichroism are reported in a previous book [15, Chapter 26: DNA–Drug Interactions] and in Chapters 14 (Electronic Circular Dichroism of Proteins), 15 (Electronic Circular Dichroism of Peptides), 17 (Circular Dichroism Spectroscopy of Nucleic Acids), 19 (Circular Dichroism of Protein–Nucleic Acid Interactions), and 20 (Drug and Natural Product Binding to Nucleic Acids Analyzed by Electronic Circular Dichroism) from the current volume.

18.2. CIRCULAR DICHROISM AS A TOOL FOR STUDYING PREORGANIZATION: SINGLE STRAND PNAS AND PNA–PNA DUPLEXES Electronic circular dichroism (ECD) has been used since early studies on PNA properties mostly by Norden and co-workers, who showed how the formation of PNA–PNA and PNA–DNA duplexes and of PNA–DNA–PNA triplexes could be followed by variation of the CD signal [7, 16]. Since then, other authors have used CD for studying the conformation of PNA analogues containing chiral moieties or the complexation of both chiral and achiral PNA with nucleic acids. Achiral standard single-strand PNAs and PNA–PNA duplexes do not show circular dichroism. However, since early studies, PNAs have been usually synthesized with a llysine moiety linked at the C-terminus, inserted for increasing their solubility in aqueous media; thus they were actually chiral [1]. Furthermore, following the successful use of PNA in many applications, a variety of molecules appeared in the literature which were designed in taking the polyamidic PNA chain as a model, and introducing either rings or substituents in their backbone, as illustrated in Figure 18.3 [17]. Many modifications contain chiral moieties and can be studied by electronic CD. Though the main interest for PNA resides in the possibility to bind DNA and RNA, the study of single-stranded PNA and of PNA–PNA double helices has provided precious insights into the “preorganization” of PNA—that is, their preference for one helix handedness—which, in turn, affect their ability to bind natural nucleic acids. Of particular relevance has been the introduction of chiral monomers derived from amino acids in the polyamide backbone (model F in Figure 18.3), for which a systematic study based on CD spectroscopy has allowed us to obtain a clear-cut rationale, which is now commonly used for the design of PNA with special properties [18]. The introduction of amino acids at the end of an achiral PNA strand does not induce circular dichroism signal in the nucleobase region, due to the flexibility of these molecules (see P1 in Figure 18.4b). Furthermore, the presence of both E and Z isomers of the tertiary amide connecting the nucleobase to the backbone, as demonstrated by NMR spectra of single-stranded PNA in solution, generates a series of 2n (n = number of PNA monomers) different stereoisomers [19], which also contributes to a random arrangement of nucleobases in this simple version of chiral PNAs.

589

590

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

Base

Ring-closure

O

O O *

N

*

n*

N H

O

O *

N

N H

Base

(c)

Base

n*

N

N H

n* R2

R2

Substituents Base Base

(d)

(e)

O *

N H

Base O

*

N

H N

(f) O

N n*

n* n

O

R1 *

N H 5 (γ)

N

2(α α)

O n*

R2

Figure 18.3. Examples of modifications of the PNA backbone described in the literature. (a) Principal strategies for inducing preorganization of PNA. (b) Example of a six-member ring structure involving the aminoethyl group. (c) Example of a five-member ring structure involving the aminoethyl group. (d) Example of a five-member ring built on the glycine part. (e) Substitution of the aminoethyl moiety with cyclic 1,2-diamines. (f) Introduction of substituents on C2, C5, or both in the aminoethylglycine backbone.

Conjugation of a peptide is a common strategy to modulate the properties of PNA in biomedical applications, especially cellular uptake. PNA–peptide chimeras can be synthesized easily using the same strategies for both the PNA and the peptide segments. Also, these compounds as single strands do not show significant circular dichroism bands in the nucleobase region, though showing typical CD of the peptide moiety [20]. Introduction of functional groups in the backbone induces a certain degree of preorganization in the ssPNA, and these products have CD signals. The PNAs bearing a C2-modified lysine monomer at the C-terminus showed no CD signals in the single strands, whereas PNAs containing C2-modified lysine monomers, either d- or l-, in the middle of a strand of achiral monomers showed a spectrum with alternate maxima in the 270–280 nm, 250–260 nm, 240–245 nm, and 220 nm regions (Figure 18.4a), with opposite signs for the two enantiomers and molar ellipticity (calculated with respect to bases) on the order of ±2000 deg/M cm for the 250–260 nm band. The modified PNAs that contain monomers derived from other amino acids have been found to have similar features, with the exception of the 270–280 nm band, which was not always present. Since these data were obtained on PNA bearing portions of self-complementary sequences, it cannot be excluded that such features are related to partially formed self-pairing PNA complexes, as observed in the solid state [21]. A much higher degree of preorganization was observed for the ssPNA containing C5-modified monomers. Ly and co-workers reported the strong preference of l-Ser- and l-Ala-derived C5 (γ )-chiral PNA monomers for a right-handedness, using NMR and molecular modeling [22]. PNA strands containing increasing numbers of l-Ser monomers showed a CD signature (Figure 18.4b) similar to that obtained for C2-modified chiral PNAs derived from d-amino acids, with similar intensity ([θ ]260 = 2000–4000 deg M−1 cm−1 per base). This suggests a similarity in the preorganization of the PNA. The synthesis of PNA containing cyclic structures was reported by several authors as a tool to induce a preorganization of the PNA strand, thus minimizing the entropy loss

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(b)

(a) Molar Ellipticity (Deg M−1 cm−1)

7000 Base

6000

O

5000 *

4000 3000 2000

N H

O

N n*

N3N+

1000 0 –1000 –2000 –3000 220

240

260

280

300

wl (nm)

(c) (i)

(ii)

Figure 18.4. Examples of CD spectra of single-stranded PNA. (a) Chiral PNA containing one C2-modified monomer based on D-lysine in the middle of the strand (molecular ellipticity calculated for nucleobase residue, data from reference 33). (b) PNA containing a C5-modified L-Ser residue (indicated as γ T) obtained changing the position and the number of chiral residues, measured in sodium phosphate buffer at 2 μM strand concentration. (Reprinted with permission from reference 22, copyright 2006 American Chemical Society.) (c) Example of PNA containing monomers derived from cyclic structures: PNA containing one T monomer based on D- (Dt.T*) or L-trans-4-aminoproline (Lt.T*) at the N-terminus; (i) (Lt.T*)-AT2 AT2 AT2 , (ii) (Dt.T*)-AT2 AT2 AT2 . (Reprinted from reference 24, copyright 1999, with permission from Elsevier.)

during PNA–DNA or PNA–RNA duplex formation [23]. This resulted in a much more structured CD signal of the corresponding PNA, although common patterns could not be inferred. As an example, we report in Figure 18.4c the CD spectra of two enantiomers of PNA containing one chiral monomer derived from d- or l-trans-hydroxyproline at the N-terminus, which showed an intense, yet peculiar, CD signal. As a general conclusion, every new structure implies a different type of preorganization and a preferred conformation and therefore has its own CD signature [24]. Although the presence of one terminal amino acid is not enough to give rise to significant CD effects for single-strand PNAs, two complementary antiparallel PNA strands with a C-terminal lysine give rise to well-defined CD spectra upon mixing, very similar to those obtained in the case of DNA–DNA duplexes, which suggest the formation of a duplex with an helical arrangement [7]. The intensity of the CD spectrum reaches a maximum when the stoichiometry of the two PNA strands is at 1:1 ratio. The effect of different amino acids linked to the Cterminus, as well as the effect of different nucleobases placed at the C-terminus close to

591

592

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the aminoacidic moiety, has been thoroughly investigated [25]. With a l-lysine at the Cterminus, the different nucleobases gave always CD spectra with the same sign. However, the intensity of the CD signal was found to be dependent upon the nucleobase closest to the C-terminus. In particular, either a guanine or a cytosine gave a well-defined circular dichroism spectrum, and in the case of cytosine the subsequent base should preferably be a purine base. The higher thermal stability of a G–C base pair, as compared to an A–T base pair, is thought to be associated to the structural stability required for the amino acid to induce a preferred handedness. A typical CD spectrum of an antiparallel PNA–PNA duplex with a terminal l-amino acid is characterized by alternate maxima, in the 270–280-nm, 250–260 nm, 240–245 nm, and 220–230 nm regions; the molar ellipticity for the 250–260 nm band depends both on the length of the PNA and on the type of amino acid used for induction, as shown in Figure 18.5, with differences [θ ]275 − [θ ]255 per base in the range 2000 to 10,000 deg M−1 cm−1 . The development of the CD spectrum upon PNA mixing is quite slow (few minutes) as compared to the appearance of the hypochromicity effect (seconds), suggesting that the base-pairing process occurs quite fast, followed by a slower reorganization with the emergence of a preferred helical handedness [25]. Replacement of the l-lysine by other l-amino acids gives essentially the same pattern, but of different intensity. In general, all the tested amino acids linked at the C-terminus with the same configuration (Lys, Leu, Phe, Ala, Glu) induced the same CD pattern, irrespective of the side chain, with the only exception of l-Glu at pH 5, whose CD spectrum appeared to be the mirror image of l-Glu (and of all the other l-amino acids) at pH 7. The presence of a positively charged side chain was found to give rise to stronger CD signals, as inferred from a systematic study involving PNA with His, Lys, Arg, and N ε -acetylated-Lys [26]. On the base of antiparallel PNA–PNA duplexes formed by PNAs of different length (4-, 6-, 8-, 10-, 12-mer) a leveling of helicity was proposed to occur from 10-mer to 12-mer [25]. A subsequent systematic study completed this series with oligomers up to 19-mer, showing that propagation of the helicity can increase the CD signal beyond this limit. Furthermore, the process of generation and propagation of the helical structure was found to have a complex behavior, due to multiple-state conformations, depending on

(b) 6

2

4

0 220 –2

240

260

280 300 Wavelength (nm)

–4 –6 –8

10-Arg 10-Lys 10-His

2 CD (mdeg)

CD (mdeg)

(a) 4

0 220 –2 –4 –6

–10

–8

–12

–10

240

260

280 300 Wavelength (nm) 17-Arg 17-Lys 17-His

Figure 18.5.

Example of CD spectra of duplexes formed by an achiral PNA and its complementary PNA strand bearing an amino acid residue at C-terminus. (a) 10-mer (HAGTGATCTAC-aa-NH2 /H-GTAGATCACT-NH2 ) and (b) 17-mer (H-CTGTGACAGTGATCTAC-aa-NH2 / H-GTAGATCACTGTCACAG-NH2 ). The CD intensity is a function of both the amino acid used

(indicated in the graph) and the length of the duplex. Spectra were recorded in water at pH 7.0, with a PNA strand concentration of 5 μM. (Reprinted with permission from reference 26, copyright 2010, American Chemical Society.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

the amino acid, the length of the duplex (Figure 18.5), and the solvent used [26]. This is in line with the observation that, in the solid state, PNA–PNA duplexes bearing a chiral amino acid at the C-term were found to form equal amounts of right- and left-handed helices, whereas in solution different external constraints give rise to a prevalence of one helix handedness [27]. The experimental CD spectrum of the antiparallel PNA–PNA duplex with a l-lysine at the C-terminus was found to be quite similar to the theoretical CD spectrum calculated for a canonical B-form DNA of the same sequence, supporting a right-handed structure for the antiparallel PNA–PNA duplex [8]. The CD spectrum predicted for a helical stack of base pairs was calculated using the Schellman matrix quantum mechanical method [25, 28, 29]. However, it was later shown by Rasmussen et al. [12] that the PNA–PNA structure is significantly different from canonical forms such as B-DNA. The antiparallel PNA–PNA helices were found to be very wide, largely pitched, and with the base pairs perpendicular to the helix axis. Thus, PNA molecules were found to have a unique structure very different from standard B- or A-helix, suitably called P-helix. The actual helical sense was later demonstrated by several experimental evidences to be left-handed (vide infra). Interestingly, despite the analogies in CD spectra and the presence of similar chromophores, different molecules do not necessarily have the same type of structure and not even the same type of helicity. The correct assignment of the handedness to the antiparallel PNA–PNA duplexes took a more definitive turn with the appearance of PNAs incorporating chiral amino acidderived monomers [30], which can be easily synthesized with amino acidic side chain in position 2 or 5 or both (Figure 18.3f) [31, 32]. PNA–PNA duplexes of PNA with an aminoacidic side chain in position 2 show CD spectra analogous to those observed for PNAs bearing a lysine residue at the C-terminus. PNAs incorporating monomers based on l-amino acids give spectra analogous to those with a l-lysine at the C-terminus whereas PNAs incorporating monomers based on d-amino acids give spectra analogous to those with a d-lysine at the C-terminus [33]. Chiral monomers inserted in the middle of the strand induce a stronger CD signal, as compared to PNA bearing a chiral monomer at the N- or C-terminus (Figure 18.6a). Thus a stronger preference in the helix handedness of an antiparallel PNA–PNA duplex is linked to a restricted conformational mobility of the chiral residue. In all cases, as in the case of amino acids linked to the C-terminus, the CD spectra of the PNA single strand are much weaker or nearly absent [33]. Since PNAs containing substituted chiral monomers with C2-modification, obtained from d-amino acids (2d), were found to bind complementary DNA with higher affinity than their enantiomers, as determined by melting temperature measurements [30, 31], it was assumed that they are in a more favorable preorganization to bind right-handed DNA and thus it was proposed that these tend to form right-handed, rather than lefthanded, helices. This assignment was independently confirmed by a method in which the handedness of antiparallel PNA–PNA duplexes can be inferred by CD of aggregates templated by the PNA–PNA duplex (see Section 18.5) [34]. The preferential handedness of PNA–PNA duplexes, both antiparallel and parallel, were further investigated by using the so-called “chiral box” PNAs—that is, PNAs bearing three consecutive chiral monomers with amino acidic side chain (derived from lysine) in position 2 in the middle of the sequence [35, 36]. Quite interestingly, in the antiparallel PNA–PNA duplexes, l-Lys PNAs form left-handed helices, while d-Lys PNAs form right-handed helices, as previously shown. The helicity assignments can be inferred directly observing the signal of the CD spectra around 260 nm according to the model previously proposed; and the induction, as deduced from the intensity of

593

594

Mol.Ellp. (Deg/M cm) 30000 (a) 25000 20000 15000 10000 5000 0 –5000 –10000 220 230 240 250 260 270 280 290 300

θ (mdeg)

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40 30 20 10 0 –10 –20 –30 –40 220

par. L-PNA-PNA

par. D-PNA-PNA

240

260

280

300

320

30

(c)

(d) 20 10 CD (mdeg)

θ (mdeg)

antipar. L-PNA-PNA

Wavelength (nm)

Wavelength (nm) 50 40 30 20 10 0 –10 –20 –30 –40 –50

(b)

antipar. D-PNA-PNA

0 15

–10 225

250

275

300

Wavelength (nm)

325

5

350

0 120 240 360 480 600 –5

–20

–15

–30 200

220

240

260

280

300

320

Wavelength (nm)

Figure 18.6. (a) CD spectra (molar ellipticity calculated as a function of base concentration) of PNA–PNA duplexes formed by an achiral PNA (H-AGTGATCTAC-NH2 ) and complementary chiral PNAs incorporating D-Lys monomers in different positions: H-GTAGAT(2D−Lys) CACT-NH2 (dotted line, one chiral monomer in the middle), H-G T(2D−Lys) AGA T(2D−Lys) CAC T(2D−Lys) -NH2 (thin line, three scattered chiral monomers), H-GTAG A(2D−Lys) T(2D−Lys) C(2D−Lys) ACT-NH2 (thick line, chiral box PNA). (From reference 35 copyright 2000, Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.) (b) Comparison between the CD spectra of antiparallel and parallel duplexes obtained with the chiral-box PNA containing either D-or L-Lys-derived C2-modified monomers H-GTAG A(2D−Lys) T(2D−Lys) C(2D−Lys) ACT-NH2 and H-GTAG A(2L−Lys) T(2L−Lys) C(2L−Lys) ACT-NH2 . (From reference 36, copyright 2005, John Wiley & Sons, reproduced with permission.) (c) CD spectra of PNA–PNA duplexes formed by PNA containing a single chiral monomer with substitution both at C2 and at C5 (H-GTAGAT(2,5−Lys) CACT-NH2 ) and complementary achiral antiparallel PNA with chiral conflict. Solid line: one monomer containing 2L,5L lysine-derived side chains. Broken line: one monomer containing 2D,5D lysine-derived side chains (data from reference 18). All measurements were carried out at 25◦ C in phosphate buffer at pH 7.0, with 5 μM concentration of each strand. (d) CD spectra of (PNA-TTTTTSS TTTTT-L-LysNH2 )2 /PNA-A10 -GlyNH2 (dotted line) and (PNATTTTTRR TTTTT-L-LysNH2 )2 /PNA-A10 -GlyNH2 (solid line) triplexes, containing a chiral monomer with either a (S, S)- or (R, R)-diaminocyclohexane in place of the aminoethyl moiety. The insert shows the CD at 255 nm as a function of time (in seconds) upon mixing the PNAs in a 2:1 base ratio (20◦ C). (Reprinted with permission from reference 38, copyright 1997, American Chemical Society.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

the CD signal, is higher than that observed with the same type and number of chiral monomers scattered (Figure 18.6a,b). The CD signals of parallel PNA–PNA duplexes were found to be very weak and with a maximum in the 260- to 270-nm region with a reversed preferential helicity (l-Lys, right-handed; d-Lys, left-handed, Figure 18.6b) [36]. Accordingly, the preferred mode of binding and the best mismatch recognition of the d-Lys containing PNA with (right-handed) DNA is in the antiparallel orientation, while that of l-Lys PNA is in the parallel mode, confirming the empirical assumption that chiral PNAs give a better performance in DNA recognition when having an intrinsic preference for right-handed helicity [36]. Finally, the helical preference of PNAs including monomers bearing chiral amino acid side chain in position 5, or position 5 and 2 simultaneously, has also been studied by using CD spectra in order to infer the helical preference of antiparallel PNA–PNA duplexes [18, 32]. In contrast with what has been observed for PNAs with chiral monomers substituted at C2, monomers with d-amino acids at C5 induce a preference for left-handed helices, and monomers with l-amino acids show a preference for right-handed helices. When both substitution are present simultaneously (2 and 5), a “chiral accord” (induction of the same handedness from both stereogenic centers) or a “chiral conflict” (induction of opposite helices) can arise. In the case of chiral accord, CD spectra confirm that the helical preference of the antiparallel PNA–PNA duplexes is exactly what can be expected (right-handed in the case of 2d,5l and left-handed in the case of 2l,5d). In the case of “chiral conflict”, CD spectra clearly indicate that the induction exerted by the stereogenic center in position 5 is prevalent (right-handed in the case of 2l,5l and left-handed in the case of 2d,5d, Figure 18.6c). Again, also in these cases the PNA–DNA duplex stability is related to the strength of the preference for the right-handed helical conformation [18]. Two stereogenic centers were also present in the modified PNAs synthesized by Lagriffoule et al. [37], including monomers obtained by replacing the aminoethyl portion of the backbone by a 1,2-diaminocyclohexyl moiety, either in the (S , S ) or the (R, R) configuration, thus introducing stereogenic centers at C5 and C4 (PNA structure as in Figure 18.3e). CD spectra of PNA single strands and of PNA–PNA antiparallel duplexes were very similar to those already seen for the previous chiral PNAs based on standard amino acids, both in terms of wavelength maxima and minima and in terms of CD intensities. Quite interestingly, the (S , S )-monomer, whose configuration at C5 corresponded to that obtained by the insertion of an l-amino acid side chain, also induced a right-handed helix in the PNA–PNA duplexes, further confirming that only the configuration of the stereogenic centers determine the PNA helix handedness, irrespectively of the groups inserted. Consistently (R, R)-monomers gave, as C5 d-amino acid-based monomers, left-handed PNA–PNA helices. This type of PNA were also reported to form PNA–PNA–PNA triplexes; the corresponding spectrum is reported in Figure 18.6d [38], which has unique features, showing a series of bands of the same sign in the 280- to 290-nm, 250- to 260-nm, and 220- to 230–nm regions.

18.3. DUPLEXES AND TRIPLEXES OF PNA WITH DNA AND RNA Normally, in ECD measurements of PNA with DNA and RNA the concentration of both components is calculated from the absorbance using the Lambert–Beer law and with the approximation of considering the single base extinction coefficients (ε) as additive. For PNA, the following values of ε are used, according to Nielsen’s work [2]: 13,700 for

595

596

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

A, 6600 for C, 11,700 for G, and 8600 for T, whereas for DNA and RNA the molar absorbtivity is either provided by the supplier or can be calculated with available free software [39]. For some PNAs, especially those without terminal or internal modifications, aggregation or adsorption on the walls of the cuvette is possible. Thus, the stock solution concentration is best measured at higher temperature (50–60◦ C). After mixing the various components in the measuring buffer, incubation at high temperature (normally 90◦ C for 5 min) and slow cooling at the desired temperature is performed, in order to disrupt eventual secondary structures and to favor hybridization; this step is necessary when the PNA is hybridized to long DNA or RNA fragments, or in the presence of strong self-aggregation of PNA.

18.3.1. PNA–DNA and PNA–RNA Duplexes of Achiral PNA The CD spectrum of achiral PNA–DNA antiparallel duplex was systematically described by Nord´en and co-workers in an early work on PNA-based structures [40]. The PNA–DNA antiparallel duplex formed by an achiral PNA shows a complex spectrum, containing several CD bands (Figure 18.7). Common features of PNA–DNA spectra are an intense Cotton effect in the range 260–265 nm, with typical [θ ] per base in the range of 10,000–20,000 deg M−1 cm−1 , and a second Cotton effect at 240–245 nm, which is much more variable both in sign and intensity; both bands are slightly shifted toward shorter wavelengths if compared with DNA–DNA duplexes of the same sequence. Since the PNA–DNA duplex is different from the ssDNA, in particular at 260–265 nm, where the ssDNA CD intensity is nearly zero, the process of formation of a PNA–DNA duplex can be followed by (a) titration of the DNA with increasing amounts of PNA and (b) monitoring the formation of the duplex at 260–265 nm. Job-plot can provide information on the stoichiometry of the complex, which is very useful when both PNA–DNA duplexes and PNA–DNA–PNA triplexes are likely to be formed, in particular with pyrimidine-rich PNAs. Sometimes, a longer-wavelength transition with lower intensity is present in the range 280–290 nm. The intensity and sign of this band is strongly dependent on the sequence and on the length of the duplex. Sugimoto et al. [41] reported a systematic study on the formation of PNA–DNA duplexes as a function of the length and sequence at 5◦ C, with the aim of providing evidences of the validity of the nearest-neighbor model for structure and stability. Some of their results are reported in Figure 18.7. Several general features are evident from these data: (a) The longer-wavelength band is either positive or negative, depending mainly on the sequence and not on the length; (b) the CD spectrum is very similar when the same nearest neighbors are present, but only for short oligomers; (c) for longer oligomers, the CD spectrum is entirely depending on the sequence used, regardless of the nearest-neighbor components; this effect is evident from the 10-mer duplexes on. Unfortunately, no theoretical prediction of electronic CD of these duplexes has so far succeeded to explain this rather complex behavior. The shorter-wavelength (200–230 nm) spectrum is even more complicated due to superposition of nucleobase and peptide chromophores; thus common features of different duplexes cannot be found. However, since the PNA–DNA and DNA–DNA duplexes are often similar in the 230–300 nm region, formation of a PNA–DNA duplex from hairpin DNA could be followed by Armitage, Nielsen, and co-workers using the 225 nm CD signal [42]. The parallel DNA–PNA duplex (with the N-terminus of PNA facing the 5 terminus of DNA) is formed with slow kinetics, and its spectrum is completely different from that

1.0 0 –1.0 –2.0 –3.0 200

10–5 [θ] (deg cm2 dmol–1)

10–5 [θ] (deg cm2 dmol–1)

(a) 2.0

(c) 2.0 1.0 0 –1.0 –2.0 –3.0 200

240 280 320 Wavelength (nm)

(b) 2.0 1.0 0 –1.0 –2.0 –3.0 200

240 280 320 Wavelength (nm) 10–5 [θ] (deg cm2 dmol–1)

10–5 [θ] (deg cm2 dmol–1)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

240 280 320 Wavelength (nm)

(d) 2.0 1.0 0 –1.0 –2.0 –3.0 200

240 280 320 Wavelength (nm)

Figure 18.7. CD spectra of antiparallel PNA:DNA duplexes of various sequences and length: (a) 6-mers PNA(CCGACG)/d(CGTCGG) (thick line) and PNA(CGACCG)/d(CGGTCG) (thin line); (b) 8-mers PNA(CTCACGGC)/d(GCCGTGAG) (thick line) and PNA(CACGGCTC)/d(GAGCCGTG) (thin line); (c) 10mers PNA(GCTAACAGCG)/d(CGCTGTTAGC) (thick line) and PNA(GCGCTACAAG)/d(CTTGTAGCGC) (thin line); (d) pna(ATAAATTGGATACAAA)/d(TTTGTATCCAATTTAT) (thick line) and pna(CAAATGGATTAAATAC)/d(GTATTTAATCCATTTG) (thin line). Sample concentration was 70 μM, and measurements were done phosphate buffer (pH 7.0) containing 1 M NaCl, at

5.0◦ C. (Reprinted with permission from reference 41, copyright 2001, American Chemical Society.)

of the antiparallel one (Figure 18.8b), with a negative band at 285 nm and a positive one at 260 nm [40]. The increase in ionic strength causes a decrease in the stability of the parallel duplex (Figure 18.8b), unlike the antiparallel (Figure 18.8a, where the spectra are superimposable). This suggests that the conformation of the DNA is highly distorted by this complexation process since a severe rearrangement of the conformation is required, in line with the lower stability and the slow kinetics of formation of these duplexes. The structural features of the parallel duplexes are at present unclear, although some molecular modeling calculations have been performed to predict their features, which were found to resemble more to the canonical B-form of DNA [43]. Although the spectra of both antiparallel and parallel PNA–RNA duplex were reported in early studies, the CD spectra of PNA–RNA have been less extensively studied. However, RNA complexation is one of the most interesting processes in the prospective of the use of PNA as antisense drugs (i.e., targeting mRNA and blocking translation) and even more in recent years for the use of PNA for blocking microRNA (miR) expression. Measurements involving RNA strands require special care in order to avoid contamination by ubiquitous RNases; therefore, RNase free water should be

597

598

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(b) 120 100 80 60 40 20 0 –20 –40 –60 –80 –100 –120 200 320 Δe (M–1 cm–1)

Δe (M–1 cm–1)

(a) 120 100 80 60 40 20 0 –20 –40 –60 –80 –100 –120 200

220

240 260 280 Wavelength (nm)

300

220

240 260 280 Wavelength (nm)

300

320

Figure 18.8. Circular dichroism spectra of (a) antiparallel PNA–DNA duplex (sequence: PNA N-GTAGATCACT-C/DNA 5 -AGTGATCTAC-3 ) at low (0 M NaCl, solid line) and high (500 mM NaCl, broken line) ionic strength; (b) parallel PNA:DNA duplex (sequence: PNA N-GTAGATCACT-C/DNA 5 -CATCTAGTGA-3 ) at low (0M NaCl, solid line) and high (500 mM NaCl, broken line); in both cases a 10 mM phosphate buffer at pH 7.0 was used with 5 mM strand concentration (Reprinted with permission from ref. [40] Copyright 1996 American Chemical Society.)

used in the preparation of samples and stock solutions of RNA should be frozen and checked for degradation before being reused after the first preparation. Two examples are reported in Figure 18.9 (unpublished data from our laboratory) in which PNA–DNA and PNA–RNA duplexes of the same sequence and length are compared. The overall shape is conserved, but slightly shifted for the PNA–RNA duplex. Both the 260–265 nm maximum and the 240–245 nm minimum are shifted to shorter wavelength, whereas the 280–290 nm band has the same sign and similar wavelength in the two complexes. Since the RNA–PNA complexes have higher melting temperatures than the PNA–DNA duplexes, it is reasonable to infer that the conformation corresponding to the PNA–RNA spectrum is more compatible with the PNA structure.

(a) 20

(b) 50 40 30

10 5 0 220 230 240 250 260 270 280 290 300 l (nm) –5

θ mdeg

θ (mdeg)

15

20 10 0 –10

220 230 240 250 260 270 280 290 300 λ (nm)

–20

Figure 18.9.

Comparison between PNA–DNA (thin lines) and PNA–RNA (thick lines) duplexes of the same sequence and length: (a) PNA 10-mer: H-GTAGATCACT-NH2 ; DNA (or RNA) 5 -AGTATCTAC-3 ; (b) PNA 18-mer: H-CCGCTGTCACACGCACAG-NH2 DNA (RNA) 5 CTGTGCGTGTGACAGCGG-3 . Measurements were done in a 10 mM phosphate buffer at pH 7 containing 0.1 M NaCl (PBS-buffer), concentration of each strand was 5 μM.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

PNA bearing methyl groups on amide nitrogens were described; upon introduction of 30% (three out of 10), N -methyl units in the PNA strand(s) no major changes were detected in PNA–DNA and PNA–RNA CD spectra, although full methylation gave distinctly altered CD spectra. However, no major changes were detected by CD spectroscopy in the formation of PNA–PNA duplexes for the fully methylated PNA [44].

18.3.2. PNA–DNA and PNA–RNA Complexes of Modified Chiral PNA As described in Section 18.2, the use of modified chiral PNA can induce pre-organization of the single-stranded PNA. This, in turn, affects the ability of the PNA to bind to DNA with the proper conformation. A duplex formed by an achiral PNA and DNA of the same length and sequence can be used as a reference for a not-distorted conformation. An example of this approach comes from the comparison of the spectra obtained for the series of PNA containing a single chiral (either C2- or C5-modified) monomer (depicted in Figure 18.3f), with complementary antiparallel DNA (Figure 18.10). Since the sequence presented is the same as in Figure 18.9a, direct comparison can be done with the achiral PNA spectrum. It turns out that the C2-modified PNA synthesized from d-Lys (2d) and C5-modified PNA derived from l-Lys (5l) have a spectrum with a (b) 20

(a) 20 PNA (2D)/apDNA

10 5

10 5 0

0 –5

PNA (5L)/apDNA

15 θ (mdeg)

θ (mdeg)

15

225

PNA (2L)/apDNA 250 275 300 325 Wavelength (nm)

–5

PNA (5D)/apDNA 225

250

275

300

325

350

Wavelength (nm)

θ (mdeg)

(c) 10 8 6 4 2 0 –2 –4 –6 –8 –10 220 230

350

240 250 260 270 280 290 300 Wavelength (nm)

Figure 18.10. CD spectra in solution of PNA containing one C2- (a) or C5-modified chiral monomer (b) derived from D- or L-Lys (H-GTAGATLys CACT-NH2 ) with antiparallel DNA (apDNA, 5 -AGTGATCTAC-3 ) in phosphate buffer at pH 7.0 (concentration: 5 μM of each strand) (from reference 18, copyright 2007, Wiley-VCH Verlag GmbH & Co. KGaA, reproduced with permission.) (c) CD spectra of PNA:DNA duplexes formed by achiral PNA (H-GTAGATCACT-NH2, thin lines with squares) or PNA (H-GTAGAD(L)-Lys TD(L)-Lys CD(L)-Lys ACT-NH2 ) containing a ‘‘chiral box’’ of C2-modified monomers derived from D- (thick solid line) or L-Lys (thin, broken line) with parallel DNA (5 -CATCTAGTGT-3 ) in phosphate buffer at pH 7.0 (concentration: 5 μM of each strand) (from reference 36, copyright 2005, John Wiley & Sons, reproduced with permission.)

599

600

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

pattern similar to the achiral PNA–DNA duplex, though with more intense bands, whereas the corresponding enantiomers (2l or 5d) have a different spectrum with a less intense band in the 260–265 nm region. This behavior parallels the preference of the corresponding PNA to form right-handed PNA–PNA helices (see above), which is an index of the preference for right-handed helical structures; this preference is also reflected in the melting temperature of these PNA–DNA duplexes, which were found to be higher for the 5l and 2d PNA (56◦ C and 52◦ C, respectively) than that of achiral PNA (50◦ C) and those of 5d and 5l (32◦ C and 47◦ C, respectively). A similar effect, but with a reversed stereoselectivity, was observed for the parallel PNA–DNA duplexes formed by PNA containing a stretch of three C2-modified chiral residues in the middle (“chiral box”); as shown in Figure 18.10, the l-Lys-containing PNA gave a spectrum of higher intensity than the achiral one, whereas the corresponding d-Lys PNA gave a different spectrum that was the superposition to that of the single strands. Thus the 2D “chiral box” PNA was demonstrated to bind DNA only in an antiparallel orientation, giving rise to a complete direction control [35]. Chiral cyclic PNA monomers have been largely used as alternative structures to increase PNA performances; several reviews are available which describe the synthesis and DNA binding properties of these derivatives [22, 45]. However, chiroptical properties of the PNA and of their complexes have not been described in all cases. The PNA–DNA spectra for chiral cyclic PNA derivatives can be very different from those of acyclic PNA–DNA and PNA–RNA duplexes with modified backbones. As an example, CD spectra of antiparallel PNA–DNA duplexes formed by a PNA containing a d- or l-trans-4-aminoprolyl monomer (tT*) at the N-terminus ((tT*)-AT2 AT2 AT2 ) were described by Ganesh and co-workers (Figure 18.11b,c). In this case, both the intensity and the maxima observed were different from those of unmodified PNA–DNA and strongly dependent on stereochemistry of the monomer. In spite of the drastic changes in the overall conformation, both PNAs were reported to bind to DNA better than unmodified aminoethylglycine-based PNA (aeg-PNA). CD spectra were also utilized in order to probe complexation in the case of fluorinated olefinic PNA (FOPA) [46], (1S,2R/1R,2S )-ciscyclopentyl-[47], aminoethylprolyl-[48], piperidine-[49], or pyrrolidinyl-PNA [50]. Since prediction of CD spectra from modeling studies is still not available, CD is sometimes used for assessing a possible similarity of the conformation of the duplexes formed with those of natural oligonucleotides or with those formed by unmodified PNA. A 1:1 complex was detected by CD for pyrrolidinyl PNA which showed preferential binding for DNA over RNA [50]. Similarly, Vilaivan and co-workers used the CD spectrum to demonstrate the occurrence of 1:1 complex of a T10 PNA based on a (S , S )-prolyl-2-aminocyclopentanecarboxylic acid (ACPC) backbone with A10 DNA with a conformation resembling that of dA10 :dT10 , but with higher stability [51]. Ganesh and co-workers used the CD intensity of PNA–RNA and PNA–DNA duplexes to give a rationale for the preference of cyclohexanyl-PNA for RNA over DNA, since the former gave a well-structured spectrum similar to that of unmodified PNA, whereas the latter showed a highly distorted pattern [52, 53]. Kumar and Ganesh described the properties of aminoethylpipecolyl (aepip)-aegPNA chimerae using CD. The insertion of one (2S , 5R)-monomer into an unmodified PNA sequence gave a spectrum closely resembling the PNA–DNA duplex, with slightly higher intensity (Figure 18.11d). This was in line with the observed higher stability of the antiparallel duplex formed by the modified PNA [54].

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

Similarity between the CD spectra of polyA RNA complexes with the backboneextended pyrrolidine-PNA (bepPNA) and that of aegPNA was used by Kumar and coworkers to explain the preference of these molecules for binding RNA over DNA [55]. Finally, many PNA-related molecules such as pyrrolidine-based oxy-peptide nucleic acids [56] and amino acid-modified oligonucleotides [57, 58] have been studied using approaches similar to those described above.

18.3.3. PNA–DNA–PNA Triplexes PNA–DNA–PNA triplexes are formed by Watson–Crick and Hoogsteen base pairing in TAT and CGC+ (C+ is protonated cytosine) triplets and are the only complex formed when the PNA is entirely composed of pyrimidine and the DNA entirely of purine nucleobases [1, 2]. A typical example of the CD spectrum of a PNA–DNA–PNA triplex in the case of a T8 PNA–LysNH2 with poly dA in a 2:1 ratio (on a nucleobase ratio) is Base

(a)

(b) 2 O

Aminoprolyl (ap)

D-trans-ap

1

N

CD m deg

H N O Base

0

240

–1

260 l nm

280

300

–2 –3

Aepip

–4

N

(c) +4

–5

O

L-trans-ap

CD m deg

+3 +2 +1 0 –1

240

260 l nm

–2

280

300

(d) Ellipticity (mdeg)

N H

3 2

(ii) (i)

1 0 –1 220

240 260 280 Wavelength (nm)

300

320

Figure 18.11. Examples of CD spectra of duplexes of modified PNA. (a) Structure of trans4-aminoprolyl-PNA (ap-PNA) and of aminoethylpipecolyl (aepip-PNA) monomer, (b) D-trans-4aminoprolyl-PNA, and (c) L-trans-4-aminoprolyl-containing PNA with antiparallel DNA (sequence: PNA H-(Lt.T* or Dt-T*)-ATTATTATT-CONHCH2 CH2 COOH, DNA: 5 dAAT AAT AAT A 3 ) in buffer, 10 mM phosphate at pH = 7.3, c = 6 μM of each strand (b,c: reprinted from reference 24, copyright 1999, with permission from Elsevier); (d) aepip-PNA monomer inserted in an achiral PNA strand (sequence: A T G T* T C T C T T T-(b-Ala)-OH, where T∗ = aminoethylpipecolyl monomer aepip) hybridized with antiparallel DNA (i) and compared to the unmodified PNA (ii), thus showing little distortion from PNA–DNA structure. Buffer: 10 mM sodium phosphate, pH 7.30, concentration = 2 μM of each strand. (Reprinted from reference 24, copyright 1999, with permission from Elsevier.)

601

602

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

25 20 15 10 5 0

(b) 8 35°C 0 0.5 1 1.5 2 2.5 [PNA]/[poly(dA)]

15

0

CD at 255 nm

CD (m deg)

(a) 30

6 21°C 4 12°C 2 0

–15 230

250 270 290 Wavelength (nm)

310

0

200

400 600 800 1000 1200 Time (seconds)

Figure 18.12. (a) Evolution of the CD signal of DNA poly(dA) (25 μM in bases) with increasing concentration of PNA T8 from bottom to top at 260 nm: 0.00, 0.33,0.67,1.00, 1.33,1.67,1.83,2.00, and 2.33. Inset: signal at 260 nm as a function of PNA–DNA base ratio. (Reprinted with permission from reference 14, copyright 1993, American Chemical Society.) (b) Kinetics followed by CD at 255 nm upon mixing 80 μM bases of PNA-T8 with 20 μM base pairs of poly(dA)–poly(dT) performed at 12◦ C, 21◦ C, and 31◦ C (in a 5 mM phosphate buffer, pH 7.0, containing 50 mM NaCl). (Reprinted with permission from reference 59, copyright 1996, American Chemical Society.)

reported in Figure 18.12a. This combination has been shown to give rise to a very stable triplex structure, which is formed also in long dsDNA segments by a strand displacement process [59]. The triplex has a typical spectrum with maxima at 285 nm, 275 nm, and 255 nm and a minimum at 240–245 nm. The 255 nm signal was used by Nielsen, Nord´en, and co-workers to establish the stoichiometry of the complex; it was possible, following the CD signal at this wavelength, to measure the kinetics of formation of triplex structures, which was found to be in the timescale of minutes (Figure 18.12b). The spectrum of a transient species could also be inferred from kinetic data at different wavelengths [59]. A case of PNA–DNA2 triplex formed by cytosine-rich homopyrimidine PNA was also reported, and it was documented following the variation of the CD signal as a function of PNA concentration [60]. Similarly, oligopyrimidine PNA containing one or two modified aepip monomers were shown to form PNA2 –DNA triplexes with polypurine DNA, and this was accompanied by a CD signal typical of PNA2 –DNA triplexes (Figure 18.13a) [54]. One interesting case is that of ornithine-based PNA, depicted in Figure 18.13b [61]. This system was shown to be not very efficient in binding DNA and to give slightly more stable triplexes with polyA RNA. A careful analysis of its optical purity revealed that severe epimerization had taken place during the solid-phase synthesis of these compounds. The optically pure enantiomers (TD-Orn )10 and (TL-orn )10 were synthesized using special reaction procedures [61, 62], and were found to bind to complementary RNA forming triplexes with low stability and very peculiar spectral features. In fact, upon complexation with both PNAs, a reversal of the CD spectrum was observed, indicating that the RNA conformation was completely modified by triplex formation (Figure 18.13). The spectrum was also very different from that of a typical PNA–DNA–PNA triplex. Analysis at different temperatures revealed two major transitions: One led to a CD spectrum similar to that of RNA and was more visible in the CD spectrum, and the second was more evident in the UV melting and was attributed to dissociation of base pairs [61]. Analogous results were obtained for systems based on Lys and 2-aminobutyric acids as backbones [63].

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(a)

(b) 30 (i)

1

(ii) (iii)

(iv)

0

θ (mdeg)

Ellipticity (mdeg)

2

Base

−1 N

−2

N H

220

240

260

(a)

10

(b)

0

(c)

−10

(d)

300

320

−30 250

Wavelength (nm)

O

Base H N

−20 O

280

20

(e) 270

HN O

290

310

λ (nm)

Figure 18.13. (a) Triplex formation by homopyrimidine PNA containing aepip-PNA monomer (T*, inset) with DNA 5 -GCAAAAAAAACG-3 showing similarities of the triplex formed; PNA: (i) H-TTTTTTTT*-(β-Ala)-OH, (ii) H-T T T T*TTTT*-(β-Ala)-OH, (iii) H-TTTTTTTT-(β-Ala)-OH, (iv) HT*TTTTTTT-(β-Ala)-OH; measurements were made in 10 mM sodium phosphate, pH 7.30, strand concentration: 1.5 μM. (Reprinted from reference 54, copyright 2004, with permission from Elsevier.) (b) Circular dichroism of: (a) RNA polyA (50 μM in bases); (b) D-ornithine T10 (structure shown in inset, 100 μM in bases); (c) L-ornithine T10 (100 μM in bases); (d) RNA polyA (50 μM in bases) L-ornithine T10 (100 μM in bases); (e) RNA polyA (50 μM in bases) D-ornithine T10 (100 μM in bases). Measurents were made in PBS buffer containing 2.5% of DMF. (From reference 61,2002, copyright John Wiley & Sons, reproduced with permission.)

Similarly, Lowe and co-workers reported the formation of triplexes of PNA poly-T made entirely of N -aminoethyl-d-proline monomers; the low thermal stability of these triplexes were accompanied by a drastic change in the CD spectrum, with negative maxima observed in the 260–280 nm region [64].

18.3.4. PNA in Quadruplexes and i-Motifs Several different noncanonical DNA structures have found application in the formation of nanostructures and nanomotors [65]. G-quadruplex and i-motifs are two main examples of this class. The G-quadruplex motif is formed by several inter- or intrastrand interactions based on stacked guanine tetrads formed by guanine-rich DNA or RNA strands. Guanine-rich sequences with quadruplex forming potential are located in the promoter regions of regulatory genes [66]. I-motifs are structures occurring with C-rich oligonucleotides, consisting of two parallel duplexes, formed by interaction of cytosine and protonated cytosine at appropriate pH, that are intercalated in an antiparallel orientation [67]. It has been shown that carefully designed G-rich PNA oligomers were capable of forming PNA4 quadruplexes [68], hybrid PNA2 –DNA2 [69], PNA2 –RNA [70, 71], or (PNA–DNA chimeras)4 [72] tetramolecular and trimolecular [73]. C5-modified chiral PNA could be used in order to induce selectivity in quadruplex versus duplex formation [74]. CD spectroscopy is highly informative in these type of studies. For example, quadruplex-forming oligonucleotides with one PNA at either 3 - or 5 -end were analyzed by CD measurements, showing that the overall conformation was conserved, as inferred from the persistence of a positive band at 264 nm and a negative band at 243 nm, typical of parallel DNA quadruplexes [75]. Thermal denaturation of these structures can be easily followed by CD measurements at different temperatures; the kinetics of quadruplex

603

604

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

formation was studied by monitoring the ellipticity at 264 nm as a function of time after heating. A G3 PNA oligomer bearing an acrydone unit at the N-terminus was mixed with a three-repeat fragment of human telomeric DNA d(GGGTTAGGGTTAGGG). The measured UV melting temperature was only one degree higher than that of the DNA quadruplex. However, the CD spectrum changed significantly, showing different intensity of the 265 nm and 295 nm bands, characteristic of parallel and antiparallel quadruplex structures, respectively [76]. The interaction of guanine-rich PNA with a portion of c-Myc RNA was studied, comparing the formation of PNA–DNA duplex via Watson–Crick base pairing and mixed quadruplex complexes based on potassium-promoted guanine tetrad formation. A 2:1 PNA–RNA complex with structural features resembling those of parallel quadruplex was demonstrated using CD signals [74]. DNA4 i-motifs show a characteristic CD profile with a positive maximum near 285 nm, which is followed by a negative trough with a minimum centered near 265 nm [67]. PNA were shown to be able to form tetrameric PNA4 [77, 78] and mixed PNA2 –DNA2 [79] and PNA2 –RNA2 [80] i-motifs. Only for the mixed PNA2 –DNA2 structure, CD data are available: A 1:1 mixture of DNA and PNA showed a CD signal that is amplified nearly twofold compared to the DNA i-motif, with the same shape, suggesting that the chirality of the DNA duplex is transferred to the nucleobases forming the PNA duplex [80].

18.4. THERMAL DENATURATION STUDIES The thermal stability of duplexes and other structures formed by PNA are normally carried out by UV measurements at variable temperatures and are based on the hyperchromic effect that is observed as a consequence of strand separation [81]. Similarly, since the CD spectra of either PNA–DNA or PNA–RNA are different from the sum of the single components (especially in the case of achiral PNA or PNA conjugated with amino acids and peptides which show no CD absorption), measurement of CD as a function of temperature can allow to detect the melting transition and, by fitting into two-state models, can give the same thermodynamic information as the UV melting. Melting of PNA–DNA duplexes is reversible, whereas the occurrence of hysteresis in the cooling curve compared to the heating can be considered as significant evidence for triplex formation, since the annealing process is slow. There are several advantages in using CD spectroscopy in these type of measurements. First of all, the intensity of a CD band chosen at an appropriate wavelength at which the two single strands give weak or no signal allows to follow the melting in a more specific way. The study of the interaction of an antitumor PNA 18-mer conjugated with a cationic peptide, able to act as anti-gene inhibitor of the MYCN oncogene in cells [82], with its target DNA provided an interesting example. A very unusual melting curve was observed in the UV, which showed an increase in absorbance near the temperature of 90◦ C (Figure 18.14a). Since a typical sigmoidal curve could not be detected in this case, in order to rule out that this increase was due to an aspecific effect, the CD melting curve at 264 nm, where both ssDNA and ssPNA CD was zero, was studied. Since the PNA–DNA mixture showed a maximum, it was possible to clearly prove the occurrence of a melting transition with mid-point near 90◦ C (Figure 18.14b), thus confirming the exceptional stability of this adduct. Careful analysis of ternary mixtures of dsDNA and PNA at variable temperature allowed us to prove a duplex invasion

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(a)

(b)

40

1.8 35

1.75 1.7

30 θ (mdeg)

25

1.6 A260

i)

i)

1.65

1.55 1.5 1.45

20 15 10

1.4 5

1.35 15

35

55

75

15

t (°C)

35

55

75

t (°C)

(c) 1.05

(d) 1.00

1

0.80

Tm (UV) = 84°C Norm CD

Norm. Abs

ii)

0

1.3

0.95

iii)

0.9 0.85

0.60 0.40 0.20

T1/2(CD) = 75°C

0.00

0.8 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

20 25 30 35 40 45 50 55 60 65 70 75 80 85

Temperature (°C)

Temperature (°C)

Figure 18.14. Comparison between UV and CD melting curves. (a) UV melting curve of a high-melting PNA–DNA duplex (PNA: H-ATG CCG GGC ATG ATC T—PKKKRKV-NH2 , DNA 5 -A GAT CAT GCC CGG CAT-3 ); (b) CD melting curves of PNA and DNA as in (a) showing specificity of the 260-nm signal for PNA–DNA (i) and lack of interference by PNA alone (ii) or DNA alone (iii). Measurements in (a) and (b) were made phosphate buffer (10 mM phosphate, 0.1 M NaCl, 0.1 mM EDTA); C = 5 μM of each strand; scan speed 18◦ C/min (From reference 83, 2008, copyright John Wiley & Sons, reproduced with permission). (c) UV melting curve of a PNA–PNA 15-mer duplex (H–GTGACAGTGATCTAC–Lys–NH2 and H–GTAGATCACTGTCAC–NH2 ) in water at pH 6.8, showing a clear inflection point at 84◦ C (c = 5 μM of each strand); (d) normalized CD at 260 nm of the same solution as in (c), showing a mid-point for the transition at lower temperature [84].

process—that is, formation of a PNA–DNA duplex from dsDNA by displacement of one strand [83]. A second advantage in using CD is represented by the possibility to detect conformational transitions that are not or scarcely detected by UV. The conformational changes observed for the ornithine-PNA mentioned above are examples of this effect. However, this implies that the transition observed in the CD melting curve not always coincides with the UV transition, since the former is sensitive to conformational changes, whereas the latter is dictated by unpairing of the nucleobases. An example of how the two effect are correlated but different comes from our own experience on the helicity of PNA–PNA duplexes. As shown in Figure 18.14c,d, by measuring the melting transition of a PNA–PNA duplex in which one of the strands has l-Lys at its C-term, a curve with mid-transition at 84◦ C was observed in the UV [84]. By measuring the CD of the duplex

605

606

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

at 260 nm, the loss of helicity accompanying the melting process could be measured, with a final zero CD intensity as expected. However, the mid-point of the transition was in this case at 75◦ C, significantly lower than that observed in the UV measurement. This implies that a transition to a different conformation with lower or no CD signal precedes the actual base dissociation, thus suggesting a higher degree of complexity of the melting process. It is evident that the CD “melting temperature” should not be considered as the Tm of the duplex in this case. It is a common experience that CD and UV melting temperatures can be different, an effect that is probably due to some degree of conformational rearrangement that precedes the actual base dissociation. However, in most cases the same results are abtained with both techniques.

18.5. INTERACTIONS OF DYES WITH PNA–XNA DUPLEXES The binding of dyes on PNA:DNA duplexes is very useful for structural diagnostics. Though DNA intercalators do not interact with the PNA–DNA duplex, due to structural difference, several groove binders were found to interact with PNA–DNA duplexes. CD spectroscopy was used to assess the binding of several molecules with PNA containing duplexes and triplexes. The 4 ,6-diamidino-2-phenylindole (DAPI), a DNA minor groove binder, was found to exhibit a circular dichroism with a positive sign and amplitude consistent with minor groove binding. Similarly, a PNA–DNA duplex containing a central AATA motif, a typical minor groove binding site for distamycin A and DAPI, showed the interaction for both drugs, though with strongly reduced affinity compared to that of dsDNA [85]. A very important class of molecules able to interact with PNA–DNA and PNA–PNA duplexes is represented by cyanine dyes, intensely colored compounds that have found a widespread application in numerous fields [86]. The typical dye structure consists of two heteroaromatic fragments linked by a polymethine chain. The extensive conjugation in the cyanines leads to long-wavelength absorption maxima and large molar absorptivities. For example, 3,3 -diethylthiacarbocyanine (DiSC2 (5)) dye (Figure 18.15a) is constituted by two benzothiazole units connected by a pentamethine linker. The extensive conjugation yields an absorption maximum at 651 nm and an extinction coefficient of 260,000 M−1 cm−1 in methanol. In many cases, cyanine dyes exist not as isolated monomers but rather as aggregates of multiple chromophores. The photophysical and photochemical properties of these aggregates have been studied in great detail and are often quite distinct from those of the monomeric dye. Armitage and co-workers [87] demonstrated the spontaneous assembly of helical cyanine dye aggregates using double-helical DNA as a nanotemplate to precisely control the spatial dimensions of the aggregate (Figure 18.15b). The binding of the dye into the minor groove of DNA, followed by additional insertion of dimers to adjacent sites, is highly cooperative. Due to the lack of hydrogen bond donor and acceptor groups on the dye, there is minimal hydrogen-bonding interaction between the cyanine dye and the DNA. Instead, the binding appears to be driven primarily by hydrophobic effect and van der Waals interactions. Since the PNA backbone is less hydrophilic than DNA, the possibility that DNA-binding cyanine dyes would also associate with PNA-containing duplexes has been also considered [34]. Indeed the dye did not bind to these structures as an isolated monomer or dimers, but rather spontaneously assembled into a helical aggregate using the PNA as a template. Upon binding to PNA–DNA, the dyes show significant changes in optical properties such as UV–vis absorbance, circular dichroism,

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(a)

S

(d)

S

+

30

N

N

20 10

(b)

0 θ (mdeg)

Cyanine Dye Aggregate Exciton Chirality

–10 –20 –30 –40

(c)

S

–50

S

+

–60

N

N

C2H5

θ (mdeg)

(e)

250 200 dye bound to antiparallel 150 L-PNA - PNA 100 50 0 –50 –100 –150 –200 –250 400 450

450

CH3

CH3

500

550

600

650

C2H5

(f) 250 dye bound to antiparallel D-PNA - PNA

500

700

Wavelength (nm)

550

600

Wavelength (nm)

650

700

200 dye bound to parallel 150 D-PNA - PNA 100 50 0 –50 –100 –150 –200 –250 400 450

dye bound to parallel L-PNA - PNA

500

550

600

650

700

Wavelength (nm)

Figure 18.15. (a) Structure of the 3,3 -diethylthiacarbocyanine dye (DiSC2 (5)) and design of aggregate interaction with PNA–DNA and PNA–PNA duplexes. (b) Scheme of the interaction of DiSC2 (5) with PNA–PNA or PNA–DNA duplexes. (c) Structure of a DiSC2 (5) analogue bearing chiral pendant arms. (d) CD titration of a PNA–DNA duplex with DiSC2 (5). [PNA–DNA] = 5.0 μM; dye is added in 0.5 μM aliquots. Spectra were recorded at 15◦ C. (Reprinted with permission from reference 34, copyright 1999, American Chemical Society.) (e) CD spectra of the cyanine dye DiSC2 (5) bound to antiparallel PNA–PNA duplexes: (solid line) D-Lys chiral box PNA–achiral antiparallel PNA; (broken line) L-Lys chiral box PNA–achiral antiparallel PNA. The PNA duplex concentrations are 5 μM, and the dye concentration is 15 μM; spectra were recorded at 20◦ C. (f) CD spectra of the cyanine dye DiSC2 (5) bound to parallel PNA–PNA duplexes: (solid line) D-Lys chiral box PNA–achiral antiparallel PNA; (broken line) L-Lys chiral box PNA–achiral antiparallel PNA, same experimental conditions as in e. (e, f: from reference 36, copyright 2005, John Wiley & Sons, reproduced with permission.)

and fluorescence. For example, the dye DiSC2 (5) exhibits upon binding an ∼120 nm shift to shorter wavelength of the main absorption band in the UV–vis. This shift is correlated to the formation of H-type aggregates, with decrease of the ground-state energy and an increase of excited-state energy [88]. This results in an instantaneous color change from blue to purple, providing a simple colorimetric indicator for PNA hybridization. This is useful for colorimetric test development, since sometimes the color change can be perceived by the naked eye [89] and easily measured by UV–vis spectroscopy [90]. Nevertheless, the use of circular dichroism is much more convenient, since only the bound dye-duplex aggregates are detected, whereas the eventual excess of free dye is CD-silent. Furthermore, the CD spectra of these aggregates can be used to establish the helix handedness of the PNA–XNA duplexes.

607

608

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

18.5.1. Cyanine Dye Aggregation Templated by PNA–DNA Duplexes Multimeric binding of DiSC2 (5) to antiparallel PNA–DNA duplex can be inferred from CD spectra, as shown in Figure 18.15c [34]. Solutions for these CD analyses are usually prepared by dissolving the PNA and DNA strands in an aqueous sodium phosphate buffer (10 mM, pH 7.0; NaCl). Stock solutions of the dye are prepared by solving the dye in methanol and measuring the concentrations spectrophotometrically using the manufacturer’s extinction coefficient (ε651 = 260,000 M−1 cm−1 in methanol). Very little amount of dye is required for most assays (c = 5–30 μM). The concentration of the target can be varied, although it is possible to detect 1–5 μM strand concentrations, with the usual dye:PNA ratio being 2:1 or 3:1. The buffer should also include at least 10% methanol in order to prevent adsorption of the dye to the walls of the sample container and aspecific aggregation. A detergent can also be used for this purpose, or other additives, such as succinyl-β-cyclodextrin [91]. DiSC2 (5) (Figure 18.15a), being a symmetrical achiral molecule, has no CD spectrum. Upon interaction with either the right-handed PNA–DNA duplex or the DNA–DNA duplex, the dye adopts the chirality of the target duplex and exhibits an induced circular exciton split CD spectrum, with crossover at 534 nm. Increasing dye concentration results in a positive CD signal at 558 nm and a negative signal at 529 nm. An isoelliptic point is observed at 544 nm, except for the first two spectra with lower dye concentration, where the crossover is around 534 nm (Figure 18.15d). As mentioned above, the bisignate bands observed are attributed to exciton coupling among multiple DiSC2 (5) molecules [15]. Exciton coupling is favored when the individual chromophores have large extinction coefficients and/or are positioned close to one another [15, 92]. Thus, the observed exciton coupling in this experiment indicates that multiple dye molecules are simultaneously bound to the PNA–DNA duplex. In addition, the positive splitting provides evidence of a right-handed helical relationship for the dye chromophores, consistent with the template of the aggregate by the PNA–DNA duplex, which is itself a right-handed helix. Since the ratio of DiSC2 (5):PNA–DNA is