2,212 245 7MB
Pages 832 Page size 595 x 842 pts (A4) Year 2011
title author publisher isbn10 | asin print isbn13 ebook isbn13 language subject publication date lcc ddc subject
cover
next page >
cover
next page >
: : : : : : : : : : :
< previous page
page_i
next page > Page i
Metal Ions in Biological Systems: Volume 33 Probing of Nucleic Acids by Metal Ion Complexes of Small Molecules Edited by Astrid Sigel and Helmut Sigel Institute of Inorganic Chemistry University of Basel CH-4056 Basel, Switzerland
< previous page
page_i
next page >
< previous page
page_ii
next page > Page ii
ISBN: 0-8247-9688-8 This book is printed on acid-free paper. COPYRIGHT © 1996 by MARCEL DEKKER, INC. ALL RIGHTS RESERVED Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. MARCEL DEKKER, INC. 270 Madison Avenue, New York, New York 10016 Current printing (last digit): 10 9 8 7 6 5 4 3 2 1 PRINTED IN THE UNITED STATES OF AMERICA
< previous page
page_ii
next page >
< previous page
page_iii
next page > Page iii
Preface to the Series Recently, the importance of metal ions to the vital functions of living organisms, hence their health and well-being, has become increasingly apparent. As a result, the long-neglected field of ''bioinorganic chemistry" is now developing at a rapid pace. The research centers on the synthesis, stability, formation, structure, and reactivity of biological metal ion-containing compounds of low and high molecular weight. The metabolism and transport of metal ions and their complexes is being studied, and new models for complicated natural structures and processes are being devised and tested. The focal point of our attention is the connection between the chemistry of metal ions and their role for life. No doubt, we are only at the brink of this process. Thus, it is with the intention of linking coordination chemistry and biochemistry in their widest sense that the Metal Ions in Biological Systems series reflects the growing field of "bioinorganic chemistry". We hope, also, that this series will help to break down the barriers between the historically separate spheres of chemistry, biochemistry, biology, medicine, and physics, with the expectation that a good deal of future outstanding discoveries will be made in the interdisciplinary areas of science. Should this series prove a stimulus for new activities in this fascinating "field", it would well serve its purpose and would be a satisfactory result for the efforts spent by the authors. Fall 1973 HELMUT SIGEL INSTITUTE OF INORGANIC CHEMISTRY UNIVERSITY OF BASEL CH-4056 BASEL, SWITZERLAND
< previous page
page_iii
next page >
< previous page
page_v
next page > Page v
Preface to Volume 33 The role of metal ions in the structure, stability, and reactivity of double-helical DNA is well documented, and it has been swidely recognized that nucleic acids in general could not exert their biological functions without the participation of metal ions that are involved in replication, transcription, and translation processes. These facts have stimulated extensive research on the interaction between metal ions and their complexes with nucleic acids. A significant further stimulus arose from the successful use of cisplatin, i.e., cis-diamminedichloroplatinum(II), as an antitumor drug. Consequently, important research now centers around the development of transition metal complexes for use in medicine or other biological applications, i.e., as spectroscopic probes, artificial nucleases, chemotherapeutic agents, and so forth. Much of the corresponding efforts is summarized in this volume. A contribution devoted to molecular modeling of transition metal adducts with nucleic acids sets the scene for this book. It is followed by an evaluation of zinc complexes as targeting agents and considerations of the fundamental chemical properties of metallocene complexes relevant to their cytostatic activity. The general agreement that cellular DNA is the target of cisplatin has led to a contribution dealing with reactions between DNA, platinum(II), and intercalators, as well as to a comparison of the coordination chemistry of trans-diamineplatinum(II) with cisplatin, and also to considerations of interactions of metal ions in multiple-stranded DNA. Generally, metal ion complexes can interact with DNA by intercalation, groove binding, or external electrostatic binding (i.e., ''outersphere" interactions) as well as by direct coordination to nucleic acids; a comprehensive chapter focuses on such DNA interactions with substi-
< previous page
page_v
next page >
< previous page
page_vi
next page > Page vi
tution-inert transition metal ion complexes. After these evaluations the effect of metal ions on the fluorescence of dyes bound to DNA and the photolytic covalent binding of metal complexes to DNA are considered, as is electrochemically activated nucleic acid oxidation. Long-range transport of electrons between metal centers in biological matrices is fundamental to essential life processes, such as photosynthesis and oxidative phosphorylation. Now it is shown that DNA with its double-helical πstacked structure may also mediate long-range electron transport; whether the DNA π stack may actually function physiologically as a ''wire" remains to be explored. Naturally, several contributions are devoted to artificial nucleases (including hydrolytic and redox reactions), which involve metalloporphyrins and synthetic metallopeptides as well as complexes of manganese, iron, nickel, copper, lanthanides, and so forth. Closely related to the present book is Volume 32, Interactions of Metal Ions with Nucleotides, Nucleic Acids, and Their Constituents. ASTRID SIGEL HELMUT SIGEL
< previous page
page_vi
next page >
< previous page
page_vii
next page > Page vii
Contents
Preface to the Series
iii
Preface to Volume 33
v
Contributors
xv
Contents of Previous Volumes
xxi
Handbook on Toxicity of Inorganic Compounds
xli
Handbook on Metals in Clinical and Analytical Chemistry
xli
Chapter 1 Molecular Modeling of Transition Metal Complexes with Nucleic Acids and Their Constituents Jiri * Kozelka 1. Introduction
2. Methodological Considerations
3. Results Overview: Modeling of Platinum-Oligonucleotide Complexes
4. Conclusion
Abbreviations
References Chapter 2 Zinc Complexes As Targeting Agents for Nucleic Acids Eiichi Kimura and Mitsuhiko Shionoya 1. Introduction
2. Zinc Complexes for Recognition of Nucleobases
3. Zinc Complexes for Control of Genetic Processes
1
2
3
16
23
24
24
29
30
37
46
49
4. Conclusions and Prospects
< previous page
page_vii
next page >
< previous page
page_viii
next page > Page viii
Abbreviations
References Chapter 3 Metallocene Interactions with DNA and DNA-Processing Enzymes Louis Y. Kuo, Andrew H. Liu, and Tobin J. Marks 1. Introduction
2. Metallocene Carcinostatic Activity
3. Metallocene-DNA Coordination Chemistry
4. Metallocene Interactions with DNA-Processing Enzymes
5. Conclusions and Outlook
Abbreviations
References Chapter 4 Evidences for a Catalytic Activity of the DNA Double Helix in the Reaction between DNA, Platinum(II), and Intercalators Marc Boudvillain, Rozenn Dalbiès, and Marc Leng 1. Introduction
2. Reactivity of cis- and trans-DDP with DNA
3. Rearrangement of trans-{Pt(NH3)2[d(GXG)-N7-G,N7-G]} Intrastrand Crosslinks
4. Reaction between cis-DDP, DNA, and Intercalators
5. Conclusion
Abbreviations
49
50
53
54
56
59
75
78
79
79
87
88
89
92
96
99
102
102
References Chapter 5 Trans-Diammineplatinum(II): What Makes It Different from cis-DDP? Coordination Chemistry of a Neglected Relative of Cisplatin and Its Interaction with Nucleic Acids Bernhard Lippert
106
1. Introduction
107
2. Basic Properties of trans-a2PtCl2
113
3. Biological Effects
114
4. Reactions with Nucleic Acids
116
5. Reactions with Defined Oligonucleotides
119
6. Model Studies
< previous page
105
page_viii
next page >
< previous page
page_ix
next page > Page ix
7. Summary
Abbreviations
References Chapter 6 Metal Ions in Multiple-Stranded DNA Michal Sabat and Bernhard Lippert 1. Introduction
2. Structure and Stability of DNA Triplexes
3. Interactions between Metal Ions and DNA Triplexes
4. Quadruplexes
5. Four-Way DNA Junctions
6. Strand Crosslinking by Metal Ions
7. Metal Complexes As Probes in Multiple-Stranded DNA
8. Summary and Future Outlook
Abbreviations
References Chapter 7 DNA Interactions with Substitution-Inert Transition Metal Ion Complexes Bengt Nordén, Per Lincoln, Björn Åkerman, and Eimer Tuite 1. Introduction
2. Tools to Characterize Transition Metal Interactions with DNA
130
132
133
143
144
145
150
154
158
160
163
168
169
169
177
179
181
204
3. Interactions of Transition Metal Complexes with DNA
232
4. Diastereomeric Binding Geometries Studied with Polarized Spectroscopy
240
5. Concluding Remarks
241
Abbreviations
242
References Chapter 8 Effect of Metal Ions on the Fluorescence of Dyes Bound to DNA Vasil G. Bregadze, Jemal G. Chkhaberidze, and Irine G. Khutsishvili
253
254
1. Introduction
2. Fluorescence Excitation Difference Spectra of Dyes in Complexes with DNA. Estimation of the Amount of Free and Bound Dyes
< previous page
page_ix
258
next page >
< previous page
page_x
next page > Page x
3. Fluorescence Quenching by Transition Metal Ions of Ethidium Bromide, Acridine Orange, and Proflavine Intercalated in DNA
4. Conclusions
Abbreviations
References Chapter 9 Photolytic Covalent Binding of Metal Complexes to DNA Mark A. Billadeau and Harry Morrison 1. Introduction
2. d6 Metal Complexes
260
265
266
266
269
270
273
284 3. Longer Wavelength Reagents: The d3 Metal Complex,
4. Future Directions
Abbreviations
References Chapter 10 Electrochemically Activated Nucleic Acid Oxidation Dean H. Johnston, Thomas W. Welch, and H. Holden Thorp 1. Introduction
2. Electrochemistry of One-Electron Couples Bound to DNA
3. Electrochemically Activated DNA Cleavage
4. Oxidation Kinetics from Voltammetry
290
291
292
297
298
299
309
314
318
5. Conclusions
319
Abbreviations
320
References Chapter 11 Electron Transfer between Metal Complexes Bound to DNA: Is DNA a Wire? Eric D. A. Stemp and Jacqueline K. Barton
326
1. Introduction
2. Early Studies with Organic and Transition Metal Donors and Acceptors
329
332
3. Metallointercalators As Donors and Acceptors
4. Electron Transfer Quenching of DNA-Bound Metallointercalcators
5. Fast Spectroscopy of DNA-Mediated Electron Transfer
< previous page
325
page_x
338
350
next page >
< previous page
page_xi
next page > Page xi
6. Comparisons of Metallointercalating Systems with Others
7. Theory
8. Future Studies
Abbreviations
References Chapter 12 Porphyrin and Metalloporphyrin Interactions with Nucleic Acids Robert F. Pasternack and Esther J. Gibbs 1. Background
2. Porphyrin Interactions with Duplex DNA
3. Binding of Porphyrins and Metalloporphyrins to Other Nucleic Acid Structures
4. Porphyrin Assemblies on Nucleic Acids
5. New Directions and Closing Remarks
Abbreviations
References Chapter 13 Selective DNA Cleavage by Metalloporphyrin Derivatives Geneviève Pratviel, Jean Bernadou, and Bernard Meunier 1. Introduction
2. DNA Cleavage by Nonvectorized Metalloporphyrin Complexes
3. Selective DNA Cleavage by Vectorized Metalloporphyrins
356
358
360
361
362
367
368
371
382
386
390
391
391
399
400
400
413
4. Concluding Remarks: Is It Possible to Go from Oxidative DNA Cleavage to Drugs Based on Metalloporphyrins?
421
Abbreviations
421
References Chapter 14 Synthetic Metallopeptides As Probes of Protein-DNA Interactions Eric C. Long, Paula Denney Eason, and Qi Liang
427
428
1. Introduction
2. Oligopeptides Containing a Pendant Metal-Binding Domain
429
443
3. Low Molecular Weight Metallopeptides
447
4. Conclusions and Future Prospects
< previous page
420
page_xi
next page >
< previous page
page_xii
next page > Page xii
Abbreviations
References Chapter 15 Targeting of Nucleic Acids by Iron Complexes Alexandra Draganescu and Thomas D. Tullius 1. Introduction
2. The Fenton Reaction Generates the Hydroxyl Radical
3. The Hydroxyl Radical As a Probe of Nucleic Acid Structure
4. Protein-Tethered Iron Complexes
5. Conclusions
Abbreviations
References Chapter 16 Nucleic Acid Chemistry of the Cuprous Complexes of 1,10-Phenanthroline and Derivatives David S. Sigman, Ralf Landgraf, David M. Perrin, and Lori Pearson 1. Introduction
2. DNase Activity of 1,10-Phenanthroline Copper
3. Comparison of the Cleavage Chemistry of the Untargeted 2:1 Complex and Targeted 1:1 Complex
4. Binding Specificity of Tetrahedral (OP)2Cu+ and Derivatives for Free DNA
5. Conclusion
Abbreviations
448
449
453
454
456
460
479
479
479
480
485
486
487
491
498
509
509
510
References Chapter 17 Specific DNA Cleavage by Manganese(III) Complexes Dennis J. Gravert and John H. Griffin
515
516
1. Introduction
2. DNA Cleavage by [SalenMn(III)]+ and Derivatives
3. Comparative Cleavage Specificity of Minor Groove Agents
532
534
4. Conclusions
534
Abbreviations
535
References
< previous page
519
page_xii
next page >
< previous page
page_xiii
next page > Page xiii
Chapter 18 Nickel Complexes As Probes of Guanine Sites in Nucleic Acid Folding Cynthia J. Burrows and Steven E. Rokita 1. Introduction
2. DNA Secondary Structure
3. RNA Tertiary Structure
4. Experimental Methods
5. Mechanistic Considerations
6. Conclusions
Abbreviations
References Chapter 19 Hydrolytic Cleavage of RNA Catalyzed by Metal Ion Complexes Janet R. Morrow 1. Introduction
2. RNA Cleavage by Metal Ion Complexes
3. Catalyst Specificity
4. Summary and Outlook
Abbreviations
References Chapter 20 RNA Recognition and Cleavage by Iron(II)-Bleomycin Jean-Marc Battigello, Mei Cui, and Barbara J. Carter
537
538
540
546
553
555
557
557
558
561
562
566
580
585
585
586
593
594
1. Introduction
598
2. Cleavage of RNA by Fe(II)-BLM
609
3. Therapeutic Relevance
610
4. Conclusions and Outlook
611
Abbreviations
612
References
< previous page
page_xiii
next page >
< previous page
page_xiv
next page > Page xiv
Chapter 21 Metallobleomycin-DNA Ineractions: Structures and Reactions Related to Bleomycin-Induced DNA Damage David H. Petering, Qunkai Mao, Wenbao Li, Eugene DeRose, and William E. Antholine
619
620
1. Introduction
2. Cellular DNA Strand Scission by Bleomycin
3. Cellular Metal Ion Requirement for Bleomycin
4. Reactions of Bleomycin in Nuclear and Plasmid Systems
5. In Vitro Mechanisms of DNA Damage
6. Comparative Chemistry of Cobalt and Iron Bleomycin in the Absence of DNA
7. Self-Inactivation Chemistry of Activated Iron Bleomycin
8. Reactions of DNA-Bound Cobalt Bleomycin
9. Reactions of DNA-Bound Iron Bleomycin
10. Interactions between FeBlm, Other Ligands, and DNA
622
624
624
625
629
633
634
635
637
638 11. Binding of
A2 and Co(III)Blm A2 to DNA Oligomers and Site Selectivity
12. Interaction of ZnBlm with DNA
13. Site Selectivity of DNA Binding and Damage
14. Mechanism of Double-Strand Cleavage
Abbreviations
References
640
641
641
643
643
Subject Index
< previous page
649
page_xiv
next page >
< previous page
page_xv
next page > Page xv
Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin. Björn Åkerman Department of Physical Chemistry, Chalmers University of Technology, S-41296 Göteborg, Sweden (FAX: 46-31-772-3858) (177) William E. Antholine National Biomedical ESR Center, Medical College of Wisconsin, Milwaukee, WI 53226, USA (619) Jacqueline K. Barton Division of Chemistry and Chemical Engineering and the Beckman Institute, 127-72, California Institute of Technology, Pasadena, CA 91125, USA (FAX: 1-818-577-4976) (325) Jean-Marc Battigello Department of Chemistry, University of Toledo, Toledo, OH 43606-3390, USA (FAX: 1-419537-4033) (593) Jean Bernadou Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, F-31077 Toulouse Cedex, France (FAX: 33-61-553003) (399) Mark A. Billadeau Department of Chemistry, Purdue University, 1393 Brown Building, West Lafayette, IN 479071393, USA (FAX: 1-317-494-0239) (269) Marc Boudvillain Centre de Biophysique Moléculaire, CNRS, Rue Charles Sadron, F-45071 Orléans Cedex 02, France (FAX: 33-38-631517) (87) Vasil G. Bregadze Institute of Physics, Academy of Sciences, 6 Tamarashvili street, Tbilisi 380077, Georgia (FAX: 995-8832-998823) (253) Cynthia J. Burrows Department of Chemistry, Henry Eyring Building, University of Utah, Salt Lake City, UT 84112, USA (FAX: 1-801-581-8433) (537)
< previous page
page_xv
next page >
< previous page
page_xvi
next page > Page xvi
Barbara J. Carter Department of Chemistry, University of Toledo, Toledo, OH 43606-3390, USA (FAX: 1-419-5374033) (593) Jemal G. Chkhaberidze Institute of Physics, Academy of Sciences, 6 Tamarashvili street, Tbilisi 380077, Georgia (FAX: 995-8832-998823) (253) Mei Cui Department of Chemistry, University of Toledo, Toledo, OH 43606-3390, USA (FAX: 1-419-537-4033) (593) Rozenn Dalbiès Centre de Biophysique Moléculaire, CNRS, Rue Charles Sadron, F-45071 Orléans Cedex 02, France (FAX: 33-38-631517) (87) Paula Denney Eason Department of Chemistry, Indiana University-Purdue University Indianapolis, 402 N. Blackford St., Indianapolis, IN 46202-3274, USA (FAX 1-317-274-4701) (427) Eugene DeRose Department of Chemistry, University of Wisconsin-Milwaukee, P. O. Box 413, Milwaukee, WI 53201-0413, USA (FAX: 1-414-229-5530) (619) Alexandra Draganescu Department of Chemistry, The Johns Hopkins University, Baltimore, MD 21218, USA (FAX: 1-410-516-8468) (453) Esther J. Gibbs Department of Chemistry, Goucher College, Towson, MD 21204, USA (367) Dennis J. Gravert Department of Chemistry, Stanford University, Stanford, CA 94305, USA (FAX: 1-415-725-0259) (515) John H. Griffin Department of Chemistry, Stanford University, Stanford, CA 94305, USA (FAX: 1-415-725-0259) (515) Dean H. Johnston Department of Chemistry, University of North Carolina, Chapel Hill, NC 27599-3290, USA (FAX: 1-919-962-2388) (297) Irine G. Khutsishvili Institute of Physics, Academy of Sciences, 6 Tamarashvili street, Tbilisi 380077, Georgia (FAX: 995-8832-998823) (253) Eiichi Kimura Department of Medicinal Chemistry, Hiroshima University School of Medicine, 1-2-3 Kasumi, Minami-ku, Hiroshima 734, Japan (FAX: 81-82-257-5324) (29)
< previous page
page_xvi
next page >
< previous page
page_xvii
next page > Page xvii
Jiri * Kozelka Laboratoire de Chimie et Biochimie, Pharmacologie et Toxicologie, URA 400 CNRS, Université René Descartes, 45 rue des Saints-Pères, F-75270 Paris 06, France (FAX: 33-1-42868387) (1) Louis Y. Kuo Department of Chemistry, Lewis and Clark College, 0615 S.W. Palatine Hill, Portland, Oregon 97219, USA (53) Ralf Landgraf Department of Biological Chemistry, Department of Chemistry and Biochemistry, and Molecular Biology Institute, University of CaliforniaLos Angeles, Los Angeles, CA 90025-1570, USA (FAX: 1-310-206-7286) (485) Marc Leng Centre de Biophysique Moléculaire, CNRS, Rue Charles Sadron, F-45071 Orléans Cedex 02, France (FAX: 33-38-631517) (87) Wenbao Li Department of Chemistry, University of Wisconsin-Milwaukee, P. O. Box 413, Milwaukee, WI 532010413, USA (FAX: 1-414-229-5530) (619) Qi Liang Department of Chemistry, Indiana University-Purdue University Indianapolis, 402 N. Blackford St., Indianapolis, IN 46202-3274, USA (FAX: 1-317-274-4701) (427) Per Lincoln Department of Physical Chemistry, Chalmers University of Technology, S-41296 Göteborg, Sweden (FAX: 46-31-772-3858) (177) Bernhard Lippert Fachbereich Chemie, Universität Dortmund, Otto-Hahn-Strasse 6, D-44227 Dortmund, Germany (FAX: 49-231-755-3771) (105/143) Andrew H. Liu DuPont Chemicals, Jackson Lab. Chambers Works, Deepwater, NJ 08023, USA (53) Eric C. Long Department of Chemistry, Indiana University-Purdue University Indianapolis, 402 N. Blackford St., Indianapolis, IN 46202-3274, USA (FAX: 1-317-274-4701) (427) Qunkai Mao Department of Chemistry, University of Wisconsin-Milwaukee, P. O. Box 413, Milwaukee, WI 532010413, USA (FAX: 1-414-229-5530) (619) Tobin J. Marks Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208-3113, USA (FAX: 1-708-491-2990) (53)
< previous page
page_xvii
next page >
< previous page
page_xviii
next page > Page xviii
Bernard Meunier Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, F-31077 Toulouse Cedex, France (FAX: 33-61-553003) (399) Harry Morrison Department of Chemistry, Purdue University, 1393 Brown Building, West Lafayette, IN 479071393, USA (FAX: 1-317-494-0239) (269) Janet R. Morrow Department of Chemistry, Natural Sciences and Mathematics Complex, State University of New York at Buffalo, Box 603000, Buffalo, NY 14260, USA (FAX: 1-716-645-6963) (561) Bengt Nordén Department of Physical Chemistry, Chalmers University of Technology, S-41296 Göteborg, Sweden (FAX: 46-31-772-3858) (177) Robert F. Pasternack Department of Chemistry, Swarthmore College, 500 College Avenue, Swarthmore, PA 19081, USA (FAX: 1-215-328-7355) (367) Lori Pearson Department of Biological Chemistry, Department of Chemistry and Biochemistry, and Molecular Biology Institute, University of CaliforniaLos Angeles, Los Angeles, CA 90025-1570, USA (FAX: 1-310-206-7286) (485) David M. Perrin Department of Biological Chemistry, Department of Chemistry and Biochemistry, and Molecular Biology Institute, University of CaliforniaLos Angeles, Los Angeles, CA 90025-1570, USA (FAX: 1-310-206-7286) (485) David H. Petering Department of Chemistry, University of Wisconsin-Milwaukee, P. O. Box 413, Milwaukee, WI 53201-0413, USA (FAX: 1-414-229-5530) (619) Geneviève Pratviel Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, F-31077 Toulouse Cedex, France (FAX: 33-61-553003) (399) Steven E. Rokita Department of Chemistry and Biochemistry, University of Maryland, College Park, MD 20742, USA (537) Michal Sabat Department of Chemistry, University of Virginia, McCormick Road, Charlottesville, VA 22901, USA (FAX: 1-804-924-3710) (143) Mitsuhiko Shionoya Institute for Molecular Science, Okazaki National Research Institutes, Myodaiji, Okazaki 444, Japan (29)
< previous page
page_xviii
next page >
< previous page
page_xix
next page > Page xix
David S. Sigman Department of Biological Chemistry, Department of Chemistry and Biochemistry, and Molecular Biology Institute, University of CaliforniaLos Angeles, Los Angeles, CA 90025-1570, USA (FAX: 1-310-206-7286) (485) Eric D. A. Stemp Division of Chemistry and Chemical Engineering and the Beckman Institute, 127-72, California Institute of Technology, Pasadena, CA 91125, USA (FAX: 1-818-577-4976) (325) H. Holden Thorp Department of Chemistry, University of North Carolina, Chapel Hill, NC 27599-3290, USA (FAX: 1-919-962-2388) (297) Eimer Tuite Department of Physical Chemistry, Chalmers University of Technology, S-41296 Göteborg, Sweden (FAX: 46-31-772-3858) (177) Thomas D. Tullius Department of Chemistry, The Johns Hopkins University, Baltimore, MD 21218, USA (FAX: 1410-516-8468) (453) Thomas W. Welch Department of Chemistry, University of North Carolina, Chapel Hill, NC 27599-3290, USA (FAX: 1-919-962-2388) (297)
< previous page
page_xix
next page >
< previous page
page_xxi
next page > Page xxi
Contents of Previous Volumes
Volume 1 Simple Complexes* Volume 2 Mixed-Ligand Complexes* Volume 3 High Molecular Complexes* Volume 4 Metal Ions As Probes* Volume 5 Reactivity of Coordination Compounds* Volume 6 Biological Action of Metal Ions* Volume 7 Iron in Model and Natural Compounds* Volume 8 Nucleotides and Derivatives: Their Ligating Ambivalency*
Volume 9 Amino Acids and Derivatives As Ambivalent Ligands 1. Complexes of α-Amino Acids with Chelatable Side Chain Donor Atoms R. Bruce Martin 2. Metal Complexes of Aspartic Acid and Glutamic Acid Christopher A. Evans, Roger Guevremont, and Dallas L. Rabenstein 3. The Coordination Chemistry of L-Cysteine and D-Penicillamine Arthur Gergely and Imre Sóvágó 4. Glutathione and Its Metal Complexes Dallas L. Rabenstein, Roger Guevremont, and Christopher A. Evans 5. Coordination Chemistry of L-Dopa and Related Ligands Arthur Gergely and Tamás Kiss
*Out of print.
< previous page
page_xxi
next page >
< previous page
page_xxii
next page > Page xxii
6. Stereoselectivity in the Metal Complexes of Amino Acids and Dipeptides Leslie D. Pettit and Robert J. W. Hefford 7. Protonation and Complexation of Macromolecular Polypeptides: Corticotropin Fragments and Basic Trypsin Inhibitor (Kunitz Base) Kálmán Burger Author Index-Subject Index
Volume 10 Carcinogenicity and Metal Ions 1. The Function of Metal Ions in Genetic Regulation Gunther L. Eichhorn 2. A Comparison of Carcinogenic Metals C. Peter Flessel, Arthur Furst, and Shirley B. Radding 3. The Role of Metals in Tumor Development and Inhibition Haleem J. Issaq 4. Paramagnetic Metal Ions in Tissue during Malignant Development Nicholas J. F. Dodd 5. Ceruloplasmin and Iron Transferrin in Human Malignant Disease Margaret A. Foster, Trevor Pocklington, and Audrey A. Dawson 6. Human Leukemia and Trace Elements E. L. Andronikashvili and L. M. Mosulishvili 7. Zinc and Tumor Growth Andre M. van Rij and Walter J. Pories 8. Cyanocobalamin and Tumor Growth Sofija Kanopkaite * and Gediminas Brazenas* 9. The Role of Selenium As a Cancer-Protecting Trace Element Birger Jansson 10. Tumor Diagnosis Using Radioactive Metal Ions and Their Complexes Akira Yokoyama and Hideo Saji Author Index-Subject Index
Volume 11 Metal Complexes As Anticancer Agents*
Volume 12 Properties of Copper 1. The Coordination Chemistry of Copper with Regard to Biological Systems R. F. Jameson
*Out of print.
< previous page
page_xxii
next page >
< previous page
page_xxiii
next page > Page xxiii
2. Copper(II) As Probe in Substituted Metalloproteins Ivano Bertini and A. Scozzafava 3. Copper(III) Complexes and Their Reactions Dale W. Margerum and Grover D. Owens 4. Copper Catalyzed Oxidation and Oxygenation Harold Gampp and Andreas D. Zuberbühler 5. Copper and the Oxidation of Hemoglobins Joseph M. Rifkind 6. Transport of Copper Bibudhendra Sarkar 7. The Role of Low-Molecular-Weight Copper Complexes in the Control of Rheumatoid Arthritis Peter M. May and David R. Williams Author Index-Subject Index
Volume 13 Copper Proteins*
Volume 14 Inorganic Drugs in Deficiency and Disease 1. Drug-Metal Ion Interaction in the Gut P. F. D'Arcy and J. C. McElnay 2. Zinc Deficiency and Its Therapy Ananda S. Prasad 3. The Pharmacological Use of Zinc George J. Brewer 4. The Anti-Inflammatory Activities of Copper Complexes John R. J. Sorenson 5. Iron-Containing Drugs David A. Brown and M. V. Chidambaram 6. Gold Complexes As Metallo-Drugs Kailash C. Dash and Hubert Schmidbaur
7. Metal Ions and Chelating Agents in Antiviral Chemotherapy D. D. Perrin and Hans Stünzi 8. Complexes of Hallucinogenic Drugs Wolfram Hänsel 9. Lithium in Psychiatry Nicholas J. Birch Author Index-Subject Index
*Out of print.
< previous page
page_xxiii
next page >
< previous page
page_xxiv
next page > Page xxiv
Volume 15 Zinc and Its Role in Biology and Nutrition 1. Categories of Zinc Metalloenzymes Alphonse Galdes and Bert L. Vallee 2. Models for Zn(II) Binding Sites in Enzymes Robert S. Brown, Joan Huguet, and Neville J. Curtis 3. An Insight on the Active Site of Zinc Enzymes through Metal Substitution Ivano Bertini and Claudio Luchinat 4. The Role of Zinc in DNA and RNA Polymerases Felicia Ying-Hsiueh Wu and Cheng-Wen Wu 5. The Role of Zinc in Snake Toxins Anthony T. Tu 6. Spectroscopic Properties of Metallothionein Milan Vasak * and Jeremias H. R. Kägi 7. Interaction of Zinc with Erythrocytes Joseph M. Rifkind 8. Zinc Absorption and Excretion in Relation to Nutrition Manfred Kirchgessner and Edgar Weigand 9. Nutritional Influence of Zinc on the Activity of Enzymes and Hormones Manfred Kirchgessner and Hans-Peter Roth 10. Zinc Deficiency Syndrome during Parenteral Nutrition Karin Ladefoged and Stig Jarnum Author Index-Subject Index
Volume 16 Methods Involving Metal Ions and Complexes in Clinical Chemistry 1. Some Aspects of Nutritional Trace Element Research Clare E. Casey and Marion F. Robinson 2. Metals and Immunity Lucy Treagan 3. Therapeutic Chelating Agents Mark M. Jones 4. Computer-Directed Chelate Therapy of Renal Stone Disease Martin Rubin and Arthur E. Martell 5. Determination of Trace Metals in Biological Materials by Stable Isotope Dilution Claude Veillon and Robert Alvarez 6. Trace Elements in Clinical Chemistry Determined by Neutron Activation Analysis Kaj Heydorn
< previous page
page_xxiv
next page >
< previous page
page_xxv
next page > Page xxv
7. Determination of Lithium, Sodium, and Potassium in Clinical Chemistry Adam Uldall and Arne Jensen 8. Determination of Magnesium and Calcium in Serum Arne Jensen and Erik Riber 9. Determination of Manganese, Iron, Cobalt, Nickel, Copper, and Zinc in Clinical Chemistry Arne Jensen, Erik Riber, Poul Persson, and Kaj Heydorn 10. Determination of Lead, Cadmium, and Mercury in Clinical Chemistry Arne Jensen, Jytte Molin Christensen, and Poul Persson 11. Determination of Chromium in Urine and Blood Ole Jøns, Arne Jensen, and Poul Persson 12. Determination of Aluminum in Clinical Chemistry Arne Jensen, Erik Riber, and Poul Persson 13. Determination of Gold in Clinical Chemistry Arne Jensen, Erik Riber, Poul Persson, and Kaj Heydorn 14. Determination of Phosphates in Clinical Chemistry Arne Jensen and Adam Uldall 15. Identification and Quantification of Some Drugs in Body Fluids by Metal Chelate Formation R. Bourdon, M. Galliot, and J. Hoffelt 16. Metal Complexes of Sulfanilamides in Pharmaceutical Analysis and Therapy Auke Bult 17. Basis for the Clinical Use of Gallium and Indium Radionuclides Raymond L. Hayes and Karl F. Hübner 18. Aspects of Technetium Chemistry As Related to Nuclear Medicine Hans G. Seiler Author Index-Subject Index
Volume 17 Calcium and Its Role in Biology 1. Bioinorganic Chemistry of Calcium R. Bruce Martin 2. Crystal Structure Studies of Calcium Complexes and Implications for Biological Systems H. Einspahr and C. E. Bugg 3. Intestinal and Renal Absorption of Calcium Piotr Gmaj and Heini Murer 4. Calcium Transport across Biological Membranes Ernesto Carafoli, Giuseppe Inesi, and Barry Rosen 5. Physiological Aspects of Mitochondrial Calcium Transport Gary Fiskum
< previous page
page_xxv
next page >
< previous page
page_xxvi
next page > Page xxvi
6. Mode of Action of the Regulatory Protein Calmodulin Jos A. Cox, Michelle Comte, Armand Malnoë, Danielle Burger, and Eric A. Stein 7. Calcium and Brain Proteins S. Alamà 8. The Roles of Ca2+ in the Regulation and Mechanism of Exocytosis Carl E. Creutz 9. Calcium Function in Blood Coagulation Gary L. Nelsestuen 10. The Role of Calcium in the Regulation of the Skeletal Muscle Contraction-Relaxation Cycle Henry G. Zot and James D. Potter 11. Calcification of Vertebrate Hard Tissues Roy E. Wuthier Author Index-Subject Index
Volume 18 Circulation of Metals in the Environment 1. Introduction to ''Circulation of Metals in the Environment" Peter Baccini 2. Analytical Chemistry Applied to Metal Ions in the Environment Arne Jensen and Sven Erik Jørgensen 3. Processes of Metal Ions in the Environment Sven Erik Jørgensen and Arne Jensen 4. Surface Complexation Paul W. Schindler 5. Relationships between Biological Availability and Chemical Measurements David R. Turner 6. Natural Organic Matter and Metal-Organic Interactions in Aquatic Systems Jacques Buffle 7. Evolutionary Aspects of Metal Ion Transport through Cell Membranes John M. Wood 8. Regulation of Trace Metal Concentrations in Fresh Water Systems Peter Baccini
9. Cycling of Metal Ions in the Soil Environment Garrison Sposito and Albert L. Page 10. Microbiological Strategies in Resistance to Metal Ion Toxicity John M. Wood
< previous page
page_xxvi
next page >
< previous page
page_xxvii
next page > Page xxvii
11. Conclusions and Outlook Peter Baccini Author Index-Subject Index
Volume 19 Antibiotics and Their Complexes 1. The Discovery of Ionophores: An Historical Account Berton C. Pressman 2. Tetracyclines and Daunorubicin R. Bruce Martin 3. Interaction of Metal Ions with Streptonigrin and Biological Properties of the Complexes Joseph Hajdu 4. Bleomycin Antibiotics: Metal Complexes and Their Biological Action Yukio Sugiura, Tomohisa Takita, and Hamao Umezawa 5. Interaction between Valinomycin and Metal Ions K. R. K. Easwaran 6. Beauvericin and the Other Enniatins Larry K. Steinrauf 7. Complexing Properties of Gramicidins James F. Hinton and Roger E. Koeppe II 8. Nactins: Their Complexes and Biological Properties Yoshiharu Nawata, Kunio Ando, and Yoichi Iitaka 9. Cation Complexes of the Monovalent and Polyvalent Carboxylic Ionophores: Lasalocid (X-537A), Monensin, A23187 (Calcimycin) and Related Antibiotics George R. Painter and Berton C. Pressman 10. Complexes of D-Cycloserine and Related Amino Acids with Antibiotic Properties Paul O'Brien 11. Iron-Containing Antibiotics J. B. Neilands and J. R. Valenta 12. Cation-Ionophore Interactions: Quantification of the Factors Underlying Selective Complexation by Means of Theoretical Computations Nohad Gresh and Alberte Pullman Author Index-Subject Index
Volume 20 Concepts on Metal Ion Toxicity 1. Distribution of Potentially Hazardous Trace Metals Garrison Sposito 2. Bioinorganic Chemistry of Metal Ion Toxicity R. Bruce Martin
< previous page
page_xxvii
next page >
< previous page
page_xxviii
next page > Page xxviii
3. The Interrelation between Essentiality and Toxicity of Metals in the Aquatic Ecosystem Elie Eichenberger 4. Metal Ion Speciation and Toxicity in Aquatic Systems Gordon K. Pagenkopf 5. Metal Toxicity to Agricultural Crops Frank T. Bingham, Frank J. Peryea, and Wesley M. Jarrell 6. Metal Ion Toxicity in Man and Animals Paul B. Hammond and Ernest C. Foulkes 7. Human Nutrition and Metal Ion Toxicity M. R. Spivey Fox and Richard M. Jacobs 8. Chromosome Damage in Individuals Exposed to Heavy Metals Alain Léonard 9. Metal Ion Carcinogenesis: Mechanistic Aspects Max Costa and J. Daniel Heck 10. Methods for the in Vitro Assessment of Metal Ion Toxicity J. Daniel Heck and Max Costa 11. Some Problems Encountered in the Analysis of Biological Materials for Toxic Trace Elements Hans G. Seiler Author Index-Subject Index
Volume 21 Applications of Nuclear Magnetic Resonance to Paramagnetic Species 1. Nuclear Relaxation Times As a Source of Structural Information Gil Navon and Gianni Valensin 2. Nuclear Relaxation in NMR of Paramagnetic Systems Ivano Bertini, Claudio Luchinat, and Luigi Messori 3. NMR Studies of Magnetically Coupled Metalloproteins Lawrence Que, Jr., and Michael J. Maroney 4. Proton NMR Studies of Biological Problems Involving Paramagnetic Heme Proteins James D. Satterlee 5. Metal-Porphyrin Induced NMR Dipolar Shifts and Their Use in Conformational Analysis Nigel J. Clayden, Geoffrey R. Moore, and Glyn Williams
6. Relaxometry of Paramagnetic Ions in Tissue Seymour H. Koenig and Rodney D. Brown III Author Index-Subject Index
< previous page
page_xxviii
next page >
< previous page
page_xxix
next page > Page xxix
Volume 22 ENDOR, EPR, and Electron Spin Echo for Probing Coordination Spheres 1. ENDOR: Probing the Coordination Environment in Metalloproteins Jürgen Hüttermann and Reinhard Kappl 2. Identification of Oxygen Ligands in Metal-Nucleotide-Protein Complexes by Observation of the Mn(II)-17O Superhyperfine Coupling Hans R. Kalbitzer 3. Use of EPR Spectroscopy for Studying Solution Equilibria Harold Gampp 4. Application of EPR Saturation Methods to Paramagnetic Metal Ions in Proteins Marvin W. Makinen and Gregg B. Wells 5. Electron Spin Echo: Applications to Biological Systems Yuri D. Tsvetkov and Sergei A. Dikanov Author Index-Subject Index
Volume 23 Nickel and Its Role in Biology 1. Nickel in the Natural Environment Robert W. Boyle and Heather W. Robinson 2. Nickel in Aquatic Systems Pamela Stokes 3. Nickel and Plants Margaret E. Farago and Monica M. Cole 4. Nickel Metabolism in Man and Animals Evert Nieboer, Rickey T. Tom, and W. (Bill) E. Sanford 5. Nickel Ion Binding to Amino Acids and Peptides R. Bruce Martin 6. Nickel in Proteins and Enzymes Robert K. Andrews, Robert L. Blakeley, and Burt Zerner 7. Nickel-Containing Hydrogenases José J. G. Moura, Isabel Moura, Miguel Teixeira, António V. Xavier, Guy D. Fauque, and Jean LeGall 8. Nickel Ion Binding to Nucleosides and Nucleotides R. Bruce Martin
9. Interactions between Nickel and DNA: Considerations about the Role of Nickel in Carcinogenesis E. L. Andronikashvili, V. G. Bregadze, and J. R. Monaselidze 10. Toxicology of Nickel Compounds Evert Nieboer, Franco E. Rossetto, and C. Rajeshwari Menon
< previous page
page_xxix
next page >
< previous page
page_xxx
next page > Page xxx
11. Analysis of Nickel in Biological Materials Hans G. Seiler Author Index-Subject Index
Volume 24 Aluminum and Its Role in Biology 1. Bioinorganic Chemistry of Aluminum R. Bruce Martin 2. Aluminum in the Environment Charles T. Driscoll and William D. Schecher 3. The Physiology of Aluminum Phytotoxicity Gregory J. Taylor 4. The Physiology of Aluminum Tolerance Gregory J. Taylor 5. Aluminum in the Diet and Mineral Metabolism Janet L. Greger 6. Aluminum Intoxication: History of Its Clinical Recognition and Management David N. S. Kerr and M. K. Ward 7. Aluminum and Alzheimer's Disease, Methodologic Approaches Daniel P. Perl 8. Mechanisms of Aluminum NeurotoxicityRelevance to Human Disease Theo P. A. Kruck and Donald R. McLachlan 9. Aluminum Toxicity and Chronic Renal Failure Michael R. Wills and John Savory 10. Analysis of Aluminum in Biological Materials John Savory and Michael R. Wills Author Index-Subject Index
Volume 25 Interrelations among Metal Ions, Enzymes, and Gene Expression 1. Metal Ion-Induced Mutagenesis in Vitro: Molecular Mechanisms Kathleen M. Downey and Antero G. So 2. Metallonucleases: Real and Artificial Lena A. Basile and Jacqueline K. Barton 3. Metalloregulatory Proteins: Metal-Responsive Molecular Switches Governing Gene Expression Thomas V. O'Halloran 4. Yeast Metallothionein: Gene Function and Regulation by Metal Ions David J. Ecker, Tauseef R. Butt, and Stanley T. Crooke
< previous page
page_xxx
next page >
< previous page
page_xxxi
next page > Page xxxi
5. Zinc-Binding Proteins Involved in Nucleic Acid Replication Joseph E. Coleman and David P. Giedroc 6. ''Zinc Fingers": The Role of Zinc(II) in Transcription Factor IIIA and Related Proteins Jeremy M. Berg 7. Site-Directed Mutagenesis and Structure-Function Relations in EF-Hand Ca2+-Binding Proteins Sture Forsén, Torbjörn Drakenberg, Sara Linse, Peter Brodin, Peter Sellers, Charlotta Johansson, Eva Thulin, and Thomas Grundström 8. Genetic Alteration of Active Site Residues of Staphylococcal Nuclease: Insights into the Enzyme Mechanism Albert S. Mildvan and Engin H. Serpersu 9. Alcohol Dehydrogenase: Structure, Catalysis, and Site-Directed Mutagenesis Y. Pocker 10. Probing the Mechanism of Action of Carboxypeptidase a by Inorganic, Organic, and Mutagenic Modifications David S. Auld, James F. Riordan, and Bert L. Vallee 11. Site-Directed Mutagenesis of E. coli Alkaline Phosphatase: Probing the Active-Site Mechanism and the Signal Sequence-Mediated Transport of the Enzyme John E. Butler-Ransohoff, Debra A. Kendall, and Emil Thomas Kaiser 12. Site-Directed Mutagenesis of Heme Proteins Patrick R. Stayton, William M. Atkins, Barry A. Springer, and Stephen G. Sligar 13. Exploring Structure-Function Relationships in Yeast Cytochrome c Peroxidase Using Mutagenesis and Crystallography J. Matthew Mauro, Mark A. Miller, Stephen L. Edwards, Jimin Wang, Laurence A. Fishel, and Joseph Kraut Author Index-Subject Index
Volume 26 Compendium on Magnesium and Its Role in Biology, Nutrition, and Physiology 1. Bioinorganic Chemistry of Magnesium R. Bruce Martin 2. Magnesium in the Environment Raili Jokinen 3. Magnesium in Plants: Uptake, Distribution, Function, and Utilization by Man and Animals Stanley R. Wilkinson, Ross M. Welch, Henry F. Mayland, and David L. Grunes
< previous page
page_xxxi
next page >
< previous page
page_xxxii
next page > Page xxxii
4. Magnesium in Animal Nutrition H. Meyer and J. Zentak 5. Dietary Magnesium and Drinking Water: Effects on Human Health Status John R. Marier 6. Magnesium in Biology and Medicine: An Overview Nicholas J. Birch 7. Role of Magnesium in Enzyme Systems Frank W. Heaton 8. Magnesium: A Regulated and Regulatory Cation Michael E. Maguire 9. Magnesium Transport in Prokaryotic Cells Marshall D. Snavely 10. Hormonal Regulation of Magnesium Homeostasis in Cultured Mammalian Cells Robert D. Grubbs 11. Functional Compartmentation of Intracellular Magnesium T. Günther 12. Membrane Transport of Magnesium T. Günther and H. Ebel 13. Intestinal Magnesium Absorption H. Ebel 14. The Renal Handling of Magnesium Michael P. Ryan 15. Magnesium and Liver Cirrhosis Leon Cohen 16. Hypomagnesemia and Hypermagnesemia Nachman Brautbar, Atul T. Roy, Philip Hom, and David B. N. Lee 17. Systemic Stress and the Role of Magnesium H. G. Classen 18. Magnesium and Lipid Metabolism Yves Rayssiguier 19. Magnesium and the Cardiovascular System: Experimental and Clinical Aspects Updated Burton M. Altura and Bella T. Altura
20. Magnesium and the Peripheral (Extradural) Nervous System: Metabolism, Neurophysiological Functions, and Clinical Disorders Jerry G. Chutkow 21. Magnesium and the Central (Intradural) Nervous System: Metabolism, Neurophysiological Functions, and Clinical Disorders Jerry G. Chutkow
< previous page
page_xxxii
next page >
< previous page
page_xxxiii
next page > Page xxxiii
22. The Role of Magnesium in the Regulation of Muscle Function Christopher H. Fry and Sarah K. Hall 23. Magnesium in Bone and Tooth Colin Robinson and John A. Weatherell 24. Magnesium and Osteoporosis Leon Cohen 25. The Role of Magnesium in Pregnancy, for the Newborn, and in Children's Diseases Ludwig Spätling 26. Magnesium and Its Role in Allergy Nicole Hunziker 27. Magnesium and Its Relationship to Oncology Jean Durlach, Michel Bara, and Andrée Guiet-Bara 28. The Assessment of Magnesium Status in Humans Ronald J. Elin 29. Magnesium and Placebo Effects in Human Medicine H. G. Classen 30. Determination of Magnesium in Biological Materials Hans G. Seiler Author Index-Subject Index
Volume 27 Electron Transfer Reactions in Metalloproteins 1. Mediation of Electron Transfer by Peptides and Proteins: Current Status Stephan S. Isied 2. Electron and Atom Group Transfer Properties of Protein Systems Hans E. M. Christensen, Lars S. Conrad, Jens Ulstrup, and Kurt V. Mikkelsen 3. Electron Tunneling Pathways in Proteins David N. Beratan, José Nelson Onuchic, and Harry B. Gray 4. Diprotein Complexes and Their Electron Transfer Reactions Nenad M. Kostic * 5. Electron Transfer between Bound Proteins George McLendon
6. Long-Range Electron Transfer in Cytochrome c Derivatives with Covalently Attached Probe Complexes Robert A. Scott, David W. Conrad, Marly K. Eidsness, Antonius C. F. Gorren, and Sten A. Wallin 7. Photoinduced Electron Transfer Reactions in Metalloprotein Complexes Labeled with Ruthenium Polypyridine Francis Millett and Bill Durham
< previous page
page_xxxiii
next page >
< previous page
page_xxxiv
next page > Page xxxiv
8. Stereoselective Effects in Electron Transfer Reactions Involving Synthetic Metal Complexes and Metalloproteins Klaus Bernauer 9. Properties and Electron Transfer Reactivity of [2Fe-2S] Ferredoxins A. Geoffrey Sykes 10. Electron Transfer in Photosynthetic Reaction Centers Sethulakshmi Kartha, Ranjan Das, and James A. Norris 11. Modeling the Primary Electron Transfer Events of Photosynthesis Michael R. Wasielewski 12. Electrochemical Investigation of Metalloproteins and Enzymes Using a Model Incorporating Microscopic Aspects of the Electrode-Solution Interface Alan M. Bond and H. Allen O. Hill Author Index-Subject Index
Volume 28 Degradation of Environmental Pollutants by Microorganisms and Their Metalloenzymes 1. General Strategies in the Biodegradation of Pollutants Thomas W. Egli 2. Oxidation of Aromatic Pollutants by Lignin-Degrading Fungi and Their Extracellular Peroxidases Kenneth E. Hammel 3. Biodegradation of Tannins James A. Field and G. Lettinga 4. Aerobic Biodegradation of Aromatic Hydrocarbons by Bacteria Shigeaki Harayama and Kenneth N. Timmis 5. Degradation of Halogenated Aromatics by Actinomycetes Bruno Winter and Wolfgang Zimmermann 6. Enzymes Catalyzing Oxidative Coupling Reactions of Pollutants Jean-Marc Bollag 7. Mechanism of Action of Peroxidases Helen Anni and Takashi Yonetani 8. Mechanistic Aspects of Dihydroxybenzoate Dioxygenases John D. Lipscomb and Allen M. Orville 9. Aerobic and Anaerobic Degradation of Halogenated Aliphatics Dick B. Janssen and Bernard Witholt
10. Mechanisms of Reductive Dehalogenation by Transition Metal Cofactors Found in Anaerobic Bacteria Lawrence P. Wackett and Craig A. Schanke
< previous page
page_xxxiv
next page >
< previous page
page_xxxv
next page > Page xxxv
11. Bacterial Degradation of Hemicelluloses Wolfgang Zimmermann 12. Degradation of Cellulose and Effects of Metal Ions on Cellulases Anil Goyal and Douglas E. Eveleigh 13. Metalloproteases and Their Role in Biotechnology Guido Grandi and Giuliano Galli 14. Metal-Dependent Conversion of Inorganic Nitrogen and Sulfur Compounds Peter M. H. Kroneck, Joachim Beuerle, and Wolfram Schumacher Author Index-Subject Index
Volume 29 Biological Properties of Metal Alkyl Derivatives 1. Global Bioalkylation of the Heavy Elements John S. Thayer 2. Analysis of Organometallic Compounds in the Environment Darren Mennie and Peter J. Craig 3. Biogeochemistry of Methylgermanium Species in Natural Waters Brent L. Lewis and H. Peter Mayer 4. Biological Properties of Alkyltin Compounds Yasuaki Arakawa and Osamu Wada 5. Biological Properties of Alkyl Derivatives of Lead Yukio Yamamura and Fumio Arai 6. Metabolism of Alkyl Arsenic and Antimony Compounds Marie Vahter and Erminio Marafante 7. Biological Alkylation of Selenium and Tellurium Ulrich Karlson and William T. Frankenberger, Jr. 8. Making and Breaking the Co-Alkyl Bond in B12 Derivatives John M. Pratt 9. Methane Formation by Methanogenic Bacteria: Redox Chemistry of Coenzyme F430 Bernhard Jaun 10. Synthesis and Degradation of Organomercurials by BacteriaA Comment by the Editors Helmut Sigel and Astrid Sigel
11. Biogenesis and Metabolic Role of Halomethanes in Fungi David B. Harper Author Index-Subject Index
< previous page
page_xxxv
next page >
< previous page
page_xxxvi
next page > Page xxxvi
Volume 30 Metalloenzymes Involving Amino Acid-Residue and Related Radicals 1. Free Radicals and Metalloenzymes: General Considerations Ei-Ichiro Ochiai 2. Peroxidases: Structure, Function, and Engineering Thomas L. Poulos and Roger E. Fenna 3. Photosystem II Curtis W. Hoganson and Gerald T. Babcock 4. Ribonucleotide Reductase in Mammalian Systems Lars Thelander and Astrid Gräslund 5. Manganese-Dependent Ribonucleotide Reduction and Overproduction of Nucleotides in Coryneform Bacteria George Auling and Hartmut Follmann 6. Prostaglandin Endoperoxide Synthases William L. Smith and Lawrence J. Marnett 7. Diol Dehydrase from Clostridium glycolicum: The Non-B12-Dependent Enzyme Maris G. N. Hartmanis 8. Diol Dehydrase and Glycerol Dehydrase, Coenzyme B12-Dependent Isozymes Tetsua Toraya 9. Adenosylcobalamin (Vitamin B12 Coenzyme)-Dependent Enzymes Ei-Ichiro Ochiai 10. S-Adenosylmethionine-Dependent Radical Formation in Anaerobic Systems Kenny K. Wong and John W. Kozarich 11. The Free Radical-Coupled Copper Active Site of Galactose Oxidase James W. Whittaker 12. Amine Oxidases Peter F. Knowles and David M. Dooley 13. Bacterial Transport of and Resistance to Copper Nigel L. Brown, Barry T. O. Lee, and Simon Silver Author Index-Subject Index
Volume 31 Vanadium and Its Role for Life 1. Inorganic Considerations on the Function of Vanadium in Biological Systems Dieter Rehder 2. Solution Properties of Vanadium(III) with Regard to Biological Systems Roland Meier, Martin Boddin, Steffi Mitzenheim, and Kan Kanamori
< previous page
page_xxxvi
next page >
< previous page
page_xxxvii
next page > Page xxxvii
3. The Vanadyl Ion: Molecular Structure of Coordinating Ligands by Electron Paramagnetic Resonance and Electron Nuclear Double Resonance Spectroscopy Marvin W. Makinen and Devkumar Mustafi 4. Vanadyl(IV) Complexes of Nucleotides Enrique J. Baran 5. Interactions of Vanadates with Biogenic Ligands Debbie C. Crans 6. Use of Vanadate-Induced Photocleavage for Detecting Phosphate Binding Sites in Proteins Andras Muhlrad and Israel Ringel 7. Vanadium-Protein Interactions N. Dennis Chasteen 8. Stimulation of Enzyme Activity by Oxovanadium Complexes Paul J. Stankiewicz and Alan S. Tracy 9. Inhibition of Phosphate-Metabolizing Enzymes by Oxovanadium(V) Complexes Paul J. Stankiewicz, Alan S. Tracy, and Debbie C. Crans 10. Vanadium-Dependent Haloperoxidases Hans Vilter 11. Vanadium Nitrogenases of Azotobacter Robert R. Eady 12. Amavadin, the Vanadium Compound of Amanitae Ernst Bayer 13. Vanadium in Ascidians and the Chemistry of Tunichromes Mitchell J. Smith, Daniel E. Ryan, Koji Nakanishi, Patrick Frank, and Keith O. Hodgson 14. Biochemical Significance of Vanadium in a Polychaete Worm Toshiaki Ishii, Izumi Nakai, and Kenji Okoshi 15. Vanadium Transport in Animal Systems Kenneth Kustin and William E. Robinson 16. Vanadium in Mammalian Physiology and Nutrition Forrest H. Nielsen 17. Vanadium Compounds As Insulin Mimics Chris Orvig, Katherine H. Thompson, Mary Battell, and John H. McNeill 18. Antitumor Activity of Vanadium Compounds Cirila Djordjevic
19. Methods for the Spectroscopic Characterization of Vanadium Centers in Biological and Related Chemical Systems C. David Garner, David Collison, and Frank E. Mabbs
< previous page
page_xxxvii
next page >
< previous page
page_xxxviii
next page > Page xxxviii
20. Analytical Procedures for the Determination of Vanadium in Biological Materials Hans G. Seiler Author Index-Subject Index
Volume 32 Interactions of Metal Ions with Nucleotides, Nucleic Acids, and Their Constituents 1. Phosphate-Metal Ion Interactions of Nucleotides and Polynucleotides Cindy Klevickis and Charles M. Grisham 2. Sugar-Metal Ion Interactions Shigenobu Yano and Masami Otsuka 3. Dichotomy of Metal Ion Binding to N1 and N7 of Purines R. Bruce Martin 4. General Conclusions from Solid State Studies of Nucleotide-Metal Ion Complexes Katsuyuki Aoki 5. Solution Structures of Nucleotide Metal Ion Complexes. Isomeric Equilibria Helmut Sigel and Bin Song 6. Stacking Interactions Involving Nucleotides and Metal Ion Complexes Osamu Yamauchi, Akira Odani, Hideki Masuda, and Helmut Sigel 7. Effect of Metal Ions on the Hydrolytic Reactions of Nucleosides and Their Phosphoesters Satu Kuusela and Harri Lönnberg 8. Metal Complexes of Sulfur-Containing Purine Derivatives Erich Dubler 9. Mechanistic Insight from Kinetic Studies on the Interaction of Model Palladium(II) Complexes with Nucleic Acid Components Tobias Rau and Rudi van Eldik 10. Platinum(II)-Nucleobase Interactions. A Kinetic Approach Jorma Arpalahti 11. NMR Studies of Oligonucleotide-Metal Ion Interactions Einar Sletten and Nils Åge Frøystein 12. Metal Ion Interactions with DNA: Considerations on Structure, Stability, and Effects from Metal Ion Binding Vasil G. Bregadze 13. Electron Transfer Reactions through the DNA Double Helix Thomas J. Meade
14. The Role of Metal Ions in Ribozymes Anna Marie Pyle
< previous page
page_xxxviii
next page >
< previous page
page_xxxix
next page > Page xxxix
15. Ternary Metal Ion-Nucleic Acid Base-Protein Complexes Michal Sabat 16. Metal-Responsive Gene Expression and the Zinc-Metalloregulatory Model David A. Suhy and Thomas V. O'Halloran 17. The Role of Iron-Sulfur Proteins Involved in Gene Regulation M. Claire Kennedy 18. Current Status of Structure-Activity Relationships of Platinum Anticancer Drugs: Activation of the transGeometry Nicholas Farrell 19. Cisplatin and Derived Anticancer Drugs: Mechanism and Current Status of DNA Binding Marieke J. Bloemink and Jan Reedijk 20. Proteins That Bind to and Mediate the Biologial Activity of Platinum Anticancer Drug-DNA Adducts Joyce P. Whitehead and Stephen J. Lippard 21. Interactions of Metallopharmaceuticals with DNA Michael J. Clarke and Michael Stubbs Subject Index
Volume 34 Mercury and Its Effects on Environment and Biology (tentative) 1. Analytical Methods for the Determination of Mercury(II) and Methylmercury Compounds. The Problem of Speciation James H. Weber 2. Mercury in Lakes and Rivers Markus Meili 3. Biogeochemical Cycling of Mercury in the Marine Environment William F. Fitzgerald and Robert P. Mason 4. Soils As a Source of Mercury and Methylmercury in the Aquatic Ecosystem Kevin Bishop and Y. H. Lee 5. Environmental Mercury Pollution from Mining of Gold and Silver Jerome O. Nriagu and Henry K. T. Wong 6. Accumulation of Mercury in Soil and Effects on the Soil Biota Lage Bringmark
7. Biogeochemistry of Mercury in the Air-Soil-Plant System Ki-Hyun Kim, Mark Barnett, and Steve Lindberg 8. Atmospheric Mercury Concentrations and Fluxes Åke Iverfeldt and John Munthe
< previous page
page_xxxix
next page >
< previous page
page_xl
next page > Page xl
9. Microbial Transformation of Mercury Species and Their Importance in the Biogeochemical Cycle of Mercury Franco Baldi 10. Bioaccumulation of Mercury in the Aquatic Food Chain in Newly Flooded Areas R. A. Bodaly, R. J. P. Fudge, B. D. Hall, M. J. Paterson, D. M. Rosenberg, J. W. M. Rudd, and V. L. St. Louis 11. Mercury in the Food Web: Accumulation and Transfer Mechanisms Alain Boudou and Francis Ribeyre 12. Mercury Speciation in Blood and Brain Tissue Birger Lind 13. Physiology and Toxicology of Mercury László Magos 14. Metabolism of Methylmercury in the Brain and Its Toxicological Significance N. Karle Mottet, Marie Vahter, Jay Charleston, and Lars Friberg 15. Maternal-Fetal Mercury Transport and Fetal Methylmercury Poisoning Rikuzo Hamada and Mitsuhiro Osame 16. Effects of Mercury on the Immune System Michael K. Pollard and Per Hultman 17. The Impact of Mercury Released from Dental 'Silver' Fillings on Plasmid-Encoded Resistances in the Primate Oral and Intestinal Bacterial Flora Cynthia A. Liebert, Joy Wireman, Tracy Smith, and Anne O. Summers 18. Inhibition of Brain Tubulin-GTP Interaction by Mercury: Similarity to Alzheimers Diseased Brain Observations Boyd W. Haley 19. Interaction of Mercury with Nucleic Acids and Their Components Einar Sletten and Willy Nerdal 20. The Metalloregulatory Protein, MerR, and Mercury-199 As a Probe of Protein Structure David L. Huffman and Thomas V. O'Halloran 21. Bacterial Mercury Resistance Genes Jon L. Hobman and Nigel L. Brown Subject Index
Other volumes are in preparation. Comments and suggestions with regard to contents, topics, and the like for future volumes of the series are welcome.
< previous page
page_xl
next page >
< previous page
page_xli
next page > Page xli
The following Marcel Dekker, Inc. books are also of interest for any reader dealing with metals or other inorganic compounds:
Handbook on Toxicity of Inorganic Compounds edited by Hans G. Seiler and Helmut Sigel, with Astrid Sigel In 74 chapters, written by 84 international authorities, this book covers the physiology, toxicity, and levels of tolerance, including prescriptions for detoxification, for all elements of the Periodic Table (up to atomic number 103). The book also contains short summary sections for each element, dealing with the distribution of the elements, their chemistry, technological uses, and ecotoxicity as well as their analytical chemistry.
Handbook on Metals in Clinical and Analytical Chemistry edited by Hans G. Seiler, Astrid Sigel, and Helmut Sigel This book is written by 80 international authorities and covers over 3500 references. The first part (15 chapters) focuses on sample treatment, quality control, etc., and on the detailed description of the analytical procedures relevant for clinical chemistry. The second part (43 chapters) is devoted to a total of 61 metals and metalloids; all these contributions are identically organized covering the clinical relevance and analytical determination of each element as well as, in short summary sections, its chemistry, distribution, and technical uses.
< previous page
page_xli
next page >
< previous page
page_1
next page > Page 1
1 Molecular Modeling of Transition Metal Complexes with Nucleic Acids and Their Constituents Jiri * Kozelka Laboratoire de Chimie et Biochimie Pharmacologie et Toxicologie, URA 400 CNRS, Université René Descartes, 45 rue des Saints-Pères, F-75270 Paris 06, France
2
1. Introduction
3
2. Methodological Considerations
3
2.1. Basic Principles
3
2.1.1. History
4
2.1.2. The Energy Function
2.1.3. Energy Minimization and Molecular Dynamics Simulations
2.2. Modeling of Oligonucleotides and Nucleic Acids
7
7
2.2.1. The Multiple Minima Problem
2.2.2. Electrostatic Interactions: Truncation Schemes
9
11
2.2.3. Representation of the Solvent
12
2.3. Modeling of Transition Metal Complexes
2.3.1. Variability of the Coordination Geometry
2.3.2. Ab Initio Calculations on Transition Metal Complexes
< previous page
5
page_1
12
12
next page >
< previous page
next page >
page_2
Page 2 13
2.3.3. The ESFF Force Field
16
3. Results Overview: Modeling of Platinum-Oligonucleotide Complexes
17
3.1. GG Intrastrand Crosslink
19
3.2. Models of the Minor Cisplation-DNA Adducts and Transplatin-DNA Adducts
20 3.3 Molecular Dynamics Simulations of
Adducts with Dinucleotides 23
4. Conclusion
23
Acknowledgment
24
Abbreviations
24
References
1 Introduction It has been widely recognized that nucleic acids could not exert their biological functions without the participation of metal ions. Metal ions are involved in replication, transcription, and translation processes [1]. Catalytic RNAs (ribozymes) are in fact true metalloenzymes, with the catalytic species being a metal complex [2]. The metal ions involved are usually the ubiquitous dications Mg2+ or Ca2+. However, transition metal ions can frequently substitute for these; such substitutions can be useful for studying the reaction mechanisms and for special effects. For instance, the error-prone replication of DNA in the presence of Mn2+ ions was recently utilized in the polymerase chain reaction (PCR) technique [3]. Because of their unique electronic properties, transition metal complexes serve as probes for nucleic acids structure and reactivity. The multiple use of transition metal compounds in the chemistry and molecular biology of nucleic acids has been reviewed [4,5]. Research on interactions between transition metals and nucleic acids has received considerable impetus by the discovery of the antitumor properties of cis-[PtCl2(NH3)2] [6,7]. With increasingly powerful spectroscopic methods, structural work on metal adducts with nucleic acids or their constituents has become more detailed and the size of systems that can be studied has
< previous page
page_2
next page >
< previous page
page_3
next page > Page 3
been steadily expanding. The transformation of spectroscopic data into three-dimensional molecular structures requires, in turn, powerful modeling tools. Computer modeling is a relatively recent methodology designed to represent and to refine structural models for molecules. It is generally based on the definition of a limited number of atom types; the forces between them are described by the so-called force field. Although the ultimate goal is to derive a force field that is transferable to a variety of systems, current force fields are usually optimized for a class of compounds. Thus, for the modeling of oligonucleotides, a different force field is generally used than that for the modeling of transition metal complexes. As discussed in detail below, modeling of nucleic acids is by no means a trivial matter and neither is modeling of transition metal complexes. The modeler of transition metal complexes with nucleic acids has the ''privilege" of being confronted with difficulties arising from the representation of both classes of compounds. Solutions have begun to emerge only recently and accordingly the number of applications has been relatively limited so far. The main part of this chapter (Sec. 2) tries therefore to expose the methodological problems that have to be addressed and to delineate possible solutions to these problems. The general principles of molecular modeling will be summarized only briefly; for a comprehensive description, the reader is referred to several reviews and textbooks, in which the general principles [810], modeling of nucleic acids [11], and modeling of transition metal complexes [12,13] are covered. In Sec. 3, applications of molecular modeling to platinum-oligonucleotide complexes are reviewed. 2 Methodological Considerations 2.1 Basic Principles 2.1.1 History In the catalogue of Baird and Tatlock's Laboratory Equipment Company of 1901 [14], we find "Dr. Eiloart's Atomic Models for use in teaching organic chemistry," consisting of wooden tetrahedra representing carbon atoms with four substituents. They are based on Van't Hoff's [15] and Le Bel's [16] concept of stereochemistry and represent precursors of
< previous page
page_3
next page >
< previous page
page_4
next page > Page 4
Corey-Pauling-Kolton (CPK) and Dreiding models. All these model systems use a limited number of atom types characterized by specific bond lengths and valence angles. Rotations about single bonds permit passing from one conformation to another; strain arising from the limited flexibility of the valence angles and, in the case of CPK models, repulsion between nonbonded atoms which are represented by van der Waals spheres permit a gross selection of conceivable conformations. The advent of computers has allowed the display of three-dimensional models on a screen and the comparison of the stabilities of different conformations on a quantitative basis. 2.1.2 The Energy Function The strain energy of a molecule can be quantified as the sum of energy terms representing the destabilization due to local deviations from ideal geometry and to contributions of nonbonded interactions. A typical energy function, as used by the program AMBER [17], is shown in Eq. (1). This energy function defines a force field in which the atoms are considered to move. Calculations using a force field built on the principles of classical mechanics are known as molecular mechanics or force field calculations.
In Eq. (1), the first three terms account for deviations of bond lengths (r), valence angles (θ), and dihedrals (φ) from their ideal values (r0, θ0 and [(γ + π/n] ± 2πk/n (k = 0,1, . . . , n), respectively), and the last two terms for nonbonded interactions between atoms separated by at least three bonds. Van der Waals energy is expressed by two different Lennard-Jones terms: a general one in which the attractive (negative) contribution is proportional to the 6th power of the interatomic distance rij and an additional term for atom pairs which can be hydrogen-bonded, in which the attractive contribution is proportional to the 10th power. The stabilizing term permits a closer mutual approach of the hydrogen-bonded pairs, in accordance with experimental evidence.
< previous page
page_4
next page >
< previous page
page_5
next page > Page 5
The Coulomb term in Eq. (1) involves a dielectric coefficient, ε. The latter should account, in principle, for two effects: (1) the electronic polarizability; (2) the screening by solvent molecules if these are not explicitly included in the calculations. It is clear that the two effects are completely different and that our formula is an oversimplification. A common approach consists of utilizing a ''distance-dependent ε," i.e., replacing the dielectric constant by a function ε(r), usually in the form ε(r) = K·r. Lavery's nucleic acids modeling program JUMNA [18] employs the ε(r) function (2), based on Hingerty's model for the dielectric
damping of the electrostatic interaction between two charges in a polar solvent [19]. Both ε(r) functions are isotropic and do not distinguish, for example, between the interaction of two nucleic acid phosphate residues belonging to two nucleotides of a base pair (i.e., interacting across the double helix) and that of two phosphates at the same distance, interacting across a groove (i.e., through water). Nevertheless, the distance-dependent ε has been widely used in modeling nucleic acids, this practice being justified by the fact that the energy-minimized structures appear reasonable, the lengths and energies of hydrogen bonds are in the correct range, and the function is simple and differentiable (see, however, the critical discussion in [20]). The description of the electrostatic interactions by the Coulomb term using a constant or distance-dependent dielectric coefficient remains clearly imperfect and is probably the largest source of error in molecular mechanics calculations. 2.1.3 Energy Minimization and Molecular Dynamics Simulations There are two basic approaches utilizing molecular mechanics calculations. In the first, static approach referred to as energy minimization, a starting conformation of a molecule (or an arrangement of a system of molecules) is refined so as to minimize the potential energy. The different optimization methods used in energy minimization are discussed in [8] and [11]. The fastest and most robust energy minimizer that has been tested in our laboratory is the quasi-Newtonian algorithm using preconditioned gradients by Gilbert and Lemaréchal [21] implemented in the AMBER-based packages ORAL [22] and MORMIN [23].
< previous page
page_5
next page >
< previous page
page_6
next page > Page 6
In the second approach, known as molecular dynamics (MD), the thermic motion of a molecule (or of a system of molecules) is simulated, based on Newton's equations of motion. At each step of the simulation, the forces acting on all independent particles (atoms or rigid residues) are calculated from the first derivatives of the potential energy, such as that defined in Eq. (1), with respect to the particle coordinates ri:
The solution of Newton's equations yields the acceleration matrix, which is then used to modify the velocity vectors. The algorithms used for the calculation of the new coordinates after a small time interval are reviewed in [8] and [11]. Perhaps the most common is the Verlet method [24]. It assumes that the mean velocity of a particle υi during the time interval ∆t (typically 1 fsec) is equal to the instantaneous velocity at the midpoint of the interval:
The velocity at time (t + ∆t/2), on the other hand, can be calculated from the velocity at time (t ∆t/2):
where ai is the average acceleration during the interval between (t ∆t/2) and (t + ∆t/2). The Verlet method assumes that this average acceleration is equal to the exact acceleration at time t, calculated from Newton's equations of motion:
where mi is the mass of the particle. The new coordinates are then calculated as
Thus the velocity is calculated at odd half-integral multiples of ∆t, whereas the coordinates are updated at integral multiples of ∆t. From the new coordinates, the new acceleration matrix is calculated, and so on. An energy minimization tries to bring bond lengths to equilibrium distances, i.e., to create a hypothetical state at temperatures of 0 K, where even the zero point vibrations are absent in an energy-minimized model. One has therefore to be careful not to overestimate the value of a
< previous page
page_6
next page >
< previous page
page_7
next page > Page 7
static, energy-minimized model. An MD simulation, on the other hand, reflects thermic motion and entropy effects. It allows for the calculation of statistically averaged quantities such as the free energy or mean geometrical parameters, which can be compared with observables from physical measurements. MD simulations permit crossing energy barriers between conformational domains. Whereas simple energy minimization will generally lead to the closest energy minimum, which is not necessarily the global minimum, in MD simulations thermic energy is added to the system and energy barriers can therefore be surmounted. MD simulations are thus used in searching for other conformational domains. In order to encompass a larger conformational space, simulations at a high temperature, such as 1000 K, are frequently used. This is possible thanks to the harmonic potential used for the calculation of bond energies, which prevents the molecules from breaking apart, as they would in reality at such a high temperature. Such a simulation has, of course, no physical meaning per se, but enables a large-scale conformational sampling. 2.2 Modeling of Oligonucleotides and Nucleic Acids Three features make the modeling of nucleic acids a delicate matter: (1) the large flexibility of the backbone composed solely of single bonds, giving rise to multiple minima of conformational energy; (2) the large overall negative charge, making the treatment of electrostatic interactions particularly important; (3) the nonglobular shape, causing the structure of nucleic acids to be strongly solvent-dependent. These three points are addressed in the following paragraphs. 2.2.1 The Multiple Minima Problem The main chain of an oligonucleotide consists of six single bonds per monomeric unit. Although the individual torsion angles are restricted to certain domains and are not completely independent of one another [25], the flexibility of the backbone remains considerable and in solution oligonucleotides exist generally as equilibrium mixtures of interconverting conformations. Conformational equilibria are also sometimes detected in oligonucleotide crystals, manifest by atoms disordered over two or more positions. This flexibility renders oligonucleotides partic-
< previous page
page_7
next page >
< previous page
page_8
next page > Page 8
ularly difficult to model. A conformational search will usually lead to a number of energy-minimized structures of similar energies and the selection between the calculated models will therefore rely on experimental observations. The latter will, however, generally be time-averaged over the interconverting conformations. 2.2.1.1 Structure Refinement Using Interatomic Distances from NMR A typical example of time-averaged experimental data are interatomic distances determined from nuclear Overhauser effect (NOE) buildup rates. The dipolar relaxation rate between two nuclei falls off with the sixth power of the distance. If an interatomic distance changes rapidly on the millisecond time scale of the NOE measurement between a short and a long value, the extracted experimental value, r(exp) = 1/6, will be somewhere between the two values. The common practice of using the experimental interatomic distances rexp as constraints for the simulated distances rmodel in energy minimization calculations or in MD simulations, in the form of a penalty function (3) added to the energy function [26], will in this case favor a nonexistent high-energy conformation with rmodel = rexp.
As a possible solution of the problem, MD simulations may use distance constraints based not on the instantaneous distance rmodel (t) but on a value time-averaged with the sixth power weighting [27]:
In the routine SANDER of the program AMBER [17], the sixth-power weighting of rmodel (t) has been combined with an exponential ''memory function" [27,28] [Eq. (5)], weighting recent values of rmodel (t) more heavily than older values. Equation (5) ensures that the current value of rmodel (t) still has some influence on the dynamics at long times t'. The time constant τ is used as an adjustable damping factor.
< previous page
page_8
next page >
< previous page
page_9
next page > Page 9
An even more challenging approach has been taken by van Gunsteren et al. [29]. They simulated a duplex octanucleotide with 1245 water molecules and 14 Na+ ions without constraints, and compared subsequently the averaged interproton distances with those determined by nuclear Overhauser spectroscopy (NOESY). The fact that only 80% of the NOE distances were satisfied within experimental error, with maximum deviations up to 2.9 Å, suggests that some parts of the oligonucleotide were poorly simulated. This can be due either to the fact that the simulation was too short (60 psec) and the conformational sampling therefore incomplete, or that the energy ''valley" in which the molecule moved during the simulation was not the correct one. One possible check of the latter hypothesis would consist of carrying out a restrained MD simulation first, leading the conformation toward the correct valley and then removing the constraints and comparing the final simulation with the experiment. 2.2.2 Electrostatic Interactions: Truncation Schemes Since nucleic acids are polyanions, the description of electrostatic interactions is of crucial importance for their modeling. In large molecules, the number of interacting atom pairs becomes excessive and the nonbonded energy terms have therefore to be truncated above a threshold distance. The problem is that whereas the dispersion energy of two particles i and j falls off with the sixth power of the distance, , the electrostatic interaction falls off only with rij. Thus, even at cutoff distances as long as 15 Å, electrostatic interactions are still significant and the truncation has a nonnegligible effect on the energy. Three functional forms have been widely used for truncation of long-range forces [30]: (1) simple truncation beyond a cutoff distance; (2) switching function which multiplies the electrostatic term; the function is unity for distances less than a switch-on distance and zero for distances larger than the cutoff distance; (3) reducing the electrostatic energy to zero at the cutoff distance by shifting the function in such a way that the first derivative is also zero at that distance (for illustration, see figure 1 in [30]). The simple truncation presents a discontinuity at the cutoff distance. This can cause serious problems during energy minimizations or MD simulations, as certain interatomic distances cross the cutoff limit.
< previous page
page_9
next page >
< previous page
page_10
next page > Page 10
A more sophisticated variant is the so-called twin-range cutoff, with a short-range cutoff within which the interactions are calculated at each step, and a long-range cutoff delimiting a zone in which the electrostatic energy is recalculated every, say, 20 steps [31,32]. This procedure is justified by the fact that long-range interactions change more slowly than short-range interactions and enables an extension of the cutoff limit with a minimal increase of computer time. The switching method avoids discontinuities in the energy function but large spurious derivatives (i.e., large forces) can arise in the switching region, especially if the latter is short. There are a number of switching functions in use. The program CHARMM [33], for instance, uses the expression (6) together with a distance-dependent dielectric.
The shifted potential avoids discontinuities in the energy function and the first derivatives, but modifies the energy function for short interatomic distances. In the program CHARMM, for instance, the electrostatic energy term is multiplied by the function (7).
Recently, reports on problems arising from the use of these classical truncation methods in MD simulations of solvated systems have accumulated. Artefactual ion-ion radial distributions [3436] and orientational correlation functions [34,37], reduced peptide flexibility [38], and deviations from experimentally derived protein [3942] or nucleic acid [42] structures were observed even with cutoff distances of 1415 Å [36,37,39,40]. Schreiber and Steinhauser [40] pointed out that with increasing cutoff distance the electrostatic energy does not monotonously converge to its final value but approaches it in an oscillatory way. The Ewald summation [43] has been proposed as an alternative method which eliminates the above problems [37,38,4042]. The main drawback of the Ewald method is its high computational cost. The latter can be alleviated by so-called particle-mesh techniques [41,44]. The
< previous page
page_10
next page >
< previous page
page_11
next page > Page 11
''particle-mesh-Ewald" method was recently successfully applied to MD simulations of fully solvated ubiquitin, and DNA and RNA oligonucleotides [42]. 2.2.3 Representation of the Solvent Applications of MD approaches to DNA can be classified into three groups [45]: (1) simulations with implicit inclusion of environment effects, e.g., by reducing the phosphate charge [46] or adding "hydrated" counterions [47]; (2) simulations using NOE constraints which reflect all effects, including those of the environment [26,27]; (3) simulations taking into account solvent and counterions explicitly. The most extended MD simulations of the latter type have been carried out by the Beveridge group, and the reader is referred to their recent review [48]. We will limit ourselves to emphasizing that the methodology is far from being well established, and so far no convincing accord between a detailed NMR study and an MD simulation has been achieved. The fact that in all studies (except one; see below) with explicit water representation artefactual base-pair dissociation occurred, unless the WatsonCrick hydrogen bonds were reinforced by (weak) harmonic constraints, suggests that there remain some fundamental problems of the force field to be solved. The recent communication by Cheatham et al., according to which the particle-mesh-Ewald technique yielded stable MD trajectories for solvated DNA and RNA oligonucleotides [42], hints that the truncation of long-range electrostatic interactions could have been a principal cause of the artefactual helix denaturation. One possible source of error is the water model itself. Most of the studies have employed the rigid body models SPC or TIP3P, which both fail to reproduce the experimental diffusion constant of water. Daggett and Levitt, who compared different water models [49], therefore raised the question of how reasonable the simulated motions of a solute can be when the water motion is 6075% too fast. Two features of the standard water models have been made responsible for the inappropriately high diffusion constant: the rigidity of the O-H bonds [49,50] and the neglect of atomic polarization [51,52]. Allowing for flexibility of the O-H bonds [49,50] or inclusion of atomic polarization in the electrostatic energy [51,52] decreased the diffusion constant toward the experimental value.
< previous page
page_11
next page >
< previous page
page_12
next page > Page 12
2.3 Modeling of Transition Metal Complexes Two features characteristic of transition metal complexes impede force field calculations on these compounds: (1) The variable coordination geometry for a given transition metal ion and the relatively large amplitude of bond angle variations for the same coordination complex complicate the classification into atom types and make the use of a harmonic term in the expression of bond angle energy doubtful. (2) The complex electronic structure of transition metal elements limits the applicability of conventional quantum chemical calculations used for the evaluation of force field parameters. These two issues are discussed in the following sections. 2.3.1 Variability of the Coordination Geometry Most transition metal ions form complexes with various coordination geometries and coordination numbers. A wellknown case involves hemoglobin in whose deoxy form the Fe(II) ion is pentacoordinate, whereas after oxygen binding it becomes hexacoordinate. For models of oxyand deoxyhemoglobin, two different iron atom types are therefore needed, with all the necessary force field parameters. Thus programs for modeling of coordination compounds require in general large parameter databases. A second point concerning the structural variability of coordination complexes is the relative flexibility of bond angles. A harmonic potential, valid only for small displacements, is rather inappropriate. The problem can be coped with by the use of a so-called Urey-Bradley potential [53], which takes into account nonbonded interactions between geminal atoms. Hambley et al. [54] used such a type of potential, replacing valence-angle bending terms by 1,3 nonbonded interactions, for the modeling of cobalt complexes. The methodology was subsequently applied in a series of studies by Hambley et al. and Comba et al. (reviewed in [12]). 2.3.2 Ab Initio Calculations on Transition Metal Complexes Quantum chemical ab initio calculations are frequently used to derive force field parameters. For instance, a common practice to determine
< previous page
page_12
next page >
< previous page
page_13
next page > Page 13
atomic charges employs an iterative routine which optimizes the fit between the potential resulting from the atomic monopoles and the exact ab initio potential [55]. The soaring development of computers has enabled ab initio calculations on polyatomic molecules that were unimaginable some years ago. However, since the computer time is proportional to the fourth to fifth power of the number of basis orbitals [56], all-electron calculations are feasible only for atoms up to a certain size. For elements of the fourth row and larger, generally only the valence orbitals are taken explicitly into account (as so-called pseudoorbitals), the influence of the core electrons being represented by an ''effective core potential" (ECP) [57]. We were interested in how the calculated bond lengths and the dipole moment in a heavy metal complex depend on the basis set and eventual post-self-consistent field (SCF) treatment and compared the results of different ab initio calculations on the test complex cis-[PtCl2-(PMe3)2] with experimental parameters. Table 1 shows that only calculations including the Møller-Plesset second-order correction (MP2) [56] yielded an acceptablealthough still not very goodaccord with experimental data. This comparison indicates that it is absolutely indispensable to take into account electron correlation effects in calculations involving heavy metals and casts doubt on atomic charges determined previously for platinum complexes on the basis of simple Hartree-Fock (HF)-SCF calculations [61,62]. There is increasing evidence that molecular orbital calculations based on the density functional theory yield results comparable with those of HF-SCF-configuration interaction (CI) calculations, while using significantly less computer time [63]. These methods could in future replace the HF methods in molecular orbital calculations used for the parametrization of force fields. 2.3.3 The ESFF Force Field Pursuing the idea of an all-element force field [63a], Biosym Technologies have developed a "rule-based" force field, i.e., a force field in which the parameters are not individually determined but are derived using empirical rules from a limited set of fundamental atomic parameters. The "extensible and systematic force field" (ESFF) supports 879 atom
< previous page
page_13
next page >
< previous page
next page >
page_14
Page 14 TABLE 1 Calculated and Experimental Bond Lengths and Dipole Momentsa, and Calculated Atomic Charges for cis-[PtCl2(PMe3)2]b HF (LANL1MB) PtCl [Å] PtP [Å] µ [D]
HF (LANL1DZ)
MP2 (LANL1DZ)
2.47
2.466
2.485
2.64
2.442
2.391
16.55
14.60
13.25
0.165
0.203
0.508
0.446(3)
0.458(3)
0.347(2)
0.48(2)
0.83(4)
0.88(3)
exp. 2.372(7)c 2.243(9)c 10.9a
Atomic charges [e]d Pt Cl P C
0.38(4)
H Computer time (min) per optimization cyclee
0.13(1) 15
0.50(7) 0.14(2) 65
0.49(7) 0.13(2) 580
aMeasured on the triethylphosphine analog, cis-[PtCl2(PEt3)2] [60]. bThe basis sets LANL1MB and LANL1DZ of Gaussian92 [56] combine ECP and pseudoorbital parameters of Hay and Wadt [57] for the platinum atom with a minimal basis set or with an extended double-zeta set for the remaining atoms, respectively. cAverage values with esd's from two structures [58,59]. dAverage values with esd's. eOn an IBM RS/6000 computer. types covering all elements of the periodic table up to Rn. A conventional force field supporting such a number of atom types would be practically intractable. Of particular interest to the modeling of transition metal complexes are the novel functional forms used in the bond angle, torsion, and electrostatic energy terms, as briefly discussed below [64]. 2.3.3.1 Bond Angle Energy The angle term harmonic in θi has been replaced by four different cosine forms for general, linear, perpen dicular, and equatorial angle types:
< previous page
page_14
next page >
< previous page
next page >
page_15
Page 15
These expressions are matched such that the force constant, i.e., the second derivative of Eai with respect to θi at the reference value
, is 2Ki in all cases. In the energy expression for equatorial angles, r13,i is the distance between
the terminal atoms 1 and 3, the angle atomic parameter for the center atom 2, and n is characteristic of a given symmetry, e.g., n = 4 for octahedral complexes. The exponential term represents a repulsion barrier preventing 2-1 and 2-3 bonds from overlapping, since the first term yields 0 for θi = 0.
The parameters Ki,
, and α0 are calculated from fundamental atomic parameters of the three atoms involved,
i.e., the ionization potential and the angle atomic parameters
.
2.3.3.2 Torsion Energy The torsion energy is calculated using expression (8). It is a function of the torsion angle τ as well as of the two and . Equation (8) has two advantages, valence angles θ1 and θ2 involved, and their reference values compared to the classical cosine term [Eq. (1)]. First, the energy goes smoothly to 0 and is free of derivative singularity, as either valence angle approaches 0 or π, since sin and sin in the denominators are nonzero and the numerators go smoothly to zero when θ1 or θ2 approaches 0 or π. Second, the τ-dependent numerator of the second term, sinn(i)θ1 sinn(i)θ2 cos (n(i)τ), can be expressed in terms of cos2θ1, cos2θ2, and a function F(θ,τ) = sin θ2 sin θ2 cos (τ), which in turn can be written as a product of unit vectors along the bonds 1-2, 2-3, and 3-4: F(θ,τ) = (r12 × r23)·(r23 × r34) = (r12·r23)(r23·r34)(r12·r34). For instance, for n(i) = 3, sin3θ1 sin3θ2 cos (3τ) = sin3θ1 sin3θ2 (4 cos3τ 3 cos τ) = 4F3(θ,τ) 3F(θ,τ) sin2θ1 sin2θ2. Therefore, the torsion energy can be expressed in terms of dot products of the unit vectors r12, r23, and r34, which simplifies considerably the calculation of the energy and its first and second derivatives.
2.3.3.3 Electrostatic Energy The electrostatic energy is calculated using the classical Coulomb potential. However, in contrast to conven-
< previous page
page_15
next page >
< previous page
page_16
next page > Page 16
tional force fields, the atomic charges are not explicitly supplied in a parameter file but derived from two fundamental atomic parameters, the electronegativity χi and the hardness ηi, corresponding to the first and second derivatives of the electrostatic energy with respect to the atomic charge qi, respectively. The charges are determined at the initial geometry by minimizing the electrostatic energy (9) with the constraint that the total charge is equal to the charge of the system.
A disputable point is the fact that the charges are calculated only once, at the initial geometry. During the energy minimization or MD simulation, the geometry can change and therefore the initially determined charges may no longer be adequate. As pointed out by Shi [65], recalculation of charges during the calculation is, on the one hand, desirable, but on the other hand, could be dangerous, since the structure may become temporarily distorted during the energy minimization or MD simulation, which could lead to physically unreasonable charges. 2.3.3.4 Validation The ESFF force field has been validated by comparing energy-minimized structures with X-ray data of 579 compounds covering the first six rows of the periodic table [64]. However, the really stringent test on organic macromolecules, including their adducts with metal complexes, which will prove whether ESFF allows the correct identification of low-energy conformations, remains to be carried out. The novel energy functional terms will be implemented in the Discover 95.0 version of the Biosym software. 3 Results Overview: Modeling of Platinum-Oligonucleotide Complexes As we have seen in the preceding section, modeling of both nucleic acids and transition metal complexes are domains in development. It is therefore not astonishing that efforts to model transition metal complexes with nucleic acids or their constituents have been scarce so far. The only area where extensive force field calculations were applied to a
< previous page
page_16
next page >
< previous page
page_17
next page > Page 17
system involving a transition metal atom bound to a nucleic acid residue have been platinum(II) complexes with oligonucleotides. These studies were motivated by the antitumor properties of certain platinum complexes [6,7] and facilitated by the fact that Pt(II) is a ''well-behaved" metal ion, i.e., its d8 electronic configuration and the high ligandfield effect on the d-orbital energies are reflected in a strong preference for a square planar structure from which Pt(II) complexes deviate appreciably only in cases of extreme ligand-ligand repulsion [66]. In the following, we give a short (not exhaustive) review of the main results obtained from modeling studies on platinum-(oligo)nucleotide complexes. Complementary information can be found in a review article by Hambley [67]. 3.1 GG Intrastrand Crosslink The major adduct formed by the antitumor drug cisplatin (cis-[PtCl2NH3)2]) with DNA is the intrastrand GG N7,N7 crosslink (see [68] for review). Most of the structural studies have therefore dealt with this bifunctional adduct. The first three papers [6971] focused on the question of which structural types are geometrically feasible. Miller et al. [69] started from the hypothesis that the Pt-GG crosslink induces a kink in the double helix [72,73]. For the tetramer d(CG*G*G)-d(CCCG), they have found that with rigid coplanar base pairs and C2'-endo sugars, and with fixed χ angles around the glycosidic bonds, G*-Pt-G*crosslinked models can be constructed without a gap in the backbone if the helix axis is kinked between the central two GC pairs by 4070° toward the major groove. In a different approach, Kozelka et al. compared fully relaxed models of double-stranded platinated oligonucleotides with [71] and without [70] a kink in the helix axis. Both structure types appeared feasible, i.e., did not present clashes or unusual backbone torsion angles, and their energies, calculated with a modified version of AMBER [74], were similar, yielding an example of the multiple minima problem (Sec. 2.2.1). Based on structural differences between the unkinked and kinked model, Herman et al. subsequently correlated data from an extended nuclear magnetic resonance (NMR) study of the decanucleotide d(GCCG*G*ATCGC)-d(GCGATCCGGC), crosslinked at the G* guanines with , with molecular models [75]. This correlation has allowed the unkinked structure to be
< previous page
page_17
next page >
< previous page
page_18
next page > Page 18
ruled out and led to a refined model with a mobile cytosine complementary to the 5'-platinated G*. Figure 1 shows the two positions of the mobile cytosine, one stacked on the 5' side and one stacked on the 3' side of the unplatinated strand. The energy barrier between the two positions was determined in an adiabatic mapping experiment [11] to be ~12 kJ/mol, which indicated that interconversions between the two positions would be rapid on the NMR time scale. Evidence for a kinked structure of intrastrand GG platinum crosslinks has emerged from elegant gel electrophoresis experiments by Lippard et al. [76,77]. The kink angel (~32°), determined by comparison with the bend angle of an A6 tract measured with the same method, was
Fig. 1. Stereo views of the central part of the decanucleotide d(GCCG*G* ATCGC)-d(GCGATCCGGC) crosslinked at the G* guanines with in two model structures. Only the d(CG*G*)-d(CCG) trinucleotide is shown, with the platinated strand (right-hand side) oriented with its 5' end toward the top. The C cytosine stacks on the 3' side (top model) or on the 5' side (bottom model). (Coordinates from [75].)
< previous page
page_18
next page >
< previous page
page_19
next page > Page 19
significantly lower than that of the molecular mechanics models (5560°) [75]. Cumulation of inaccuracies of both methods is probably at the origin of this discrepancy. The above modeling work has demonstrated the necessity of backing up molecular mechanics calculations with experimental results, the energy calculations being insufficiently accurate to permit the identification of the true structure among the different minimum energy models. An example of how dangerous molecular modeling without the experimental feedback can be is the ''intermediate conformation" (IC) model for the pentamer d(TCG*G*T)d(ACCGA) crosslinked at the G* guanines with by McCarthy et al. [78]: the backbone torsion angles of this model feature an exotic sequence of values which contradicts the experimental evidence from related NMR work [73,75]. Unfortunately, their comparative study of GG adducts with primary amine derivatives of an interesting project per sehas been carried out using this IC model [79,80]. 3.2 Models of the Minor Cisplatin-DNA Adducts and Transplatin-DNA Adducts The structures of the minor adducts formed between cisplatin and DNA, i.e., the A*G* [81,82] and G*XG* [83,84] intrastrand crosslinks, and the G*C-G*C interstrand crosslink [85] were investigated by different groups. Two studies focused on intrastrand G*AG* [86] and interstrand G*-C* [87] crosslinks formed by the inactive isomer, trans- [PtCl2(NH3)2] ("transplatin"). The modeling of the interstrand crosslinks was backed up by gel electrophoresis experiments which indicated a large kink (>34°) in the double helix of the cisplatin adduct and a moderate kink (~26°) in that of the transplatin adduct. As in the case of the intrastrand GG crosslinks, the energy differences between the various models obtained for one adduct were too small to permit an energy-based selection. We see once more that unambiguous structural predictions cannot be based on energy calculations alone with the present force fields. A nice example of the complementarity between NMR spectroscopy and molecular mechanics calculations is the proof for the existence of macrochelate complexes of purine 5' -nucleoside mono- and triphosphates featuring an N7Pt-OP crosslink presented by Reily et al. [88].
< previous page
page_19
next page >
< previous page
page_20
next page > Page 20
3.3 Molecular Dynamics Simulations of
Adducts with Dinucleotides
In a cautious effort to exploit the MD technique, we recently embarked on simulations of the dinucleotide complexes cis-[Pt(NH3)2{d(GpG)}] + (1) and cis-[Pt(NH3)2{r(GpG)}] + (2) [89]. These simple models of the Pt-GG crosslink were chosen not only because of their small size (so that simulations of several hundreds of picoseconds could be carried out) but because they represent two interesting helicoidal structures which are mutually inverted, as indicated by the inversion of the two H8 resonances in their NMR spectra [90]. A correlation between NMR and circular dichroism (CD) spectra, on the one hand, and the geometrical features of the conformations occurring during the simulations, on the other hand, allowed us to conclude that 1 forms a lefthanded and 2 a right-handed helical structure in solution [89]. Figure 2 shows the evolution of three geometrical parameters of 2 with time during a 630-psec simulation at 350 K. These parameters include the helicity, the nonbonded distance separating the hydrogen of the 2'-OH group of the 5'-nucleotide from one terminal oxygen of the phosphodiester group, OA, and the torsion angle χ2 about the glycosidic bond of the 5'-nucleotide. These three parameters are clearly correlated and delimit two distinct conformational domains: {1} in which the helicity is predominantly left-handed, the OH . . . OP separation oscillates around 2.7 Å, and the χ2 value is in the syn range; and {2} with preponderantly right-handed helicity, a mean OH . . . OP distance of 4.2 Å, and the χ2 angle in the anti range. Although the number of interconversions between the two domains (7 during the 630-psec run) is not sufficient to permit a statistical analysis, the simulation suggests that both domains are roughly equally populated. A similar figure resulted for a simulation of 1. Obviously, our force field is not accurate enough to allow the identification of the (apparently small) energy differences between the conformations {1} and {2}, which cause 1 to prefer {1} and 2 to prefer {2}. Representative energy-minimized models of both conformations of 2 are shown in Fig. 3. For conformation {1}, we observe that the 2'-OH group of the 5'-nucleotide forms a three-center hydrogen bond with two oxygen atoms of the phosphodiester group; both H . . . O separations being
< previous page
page_20
next page >
< previous page
page_21
next page > Page 21
Fig. 2. Evolution of three geometrical parameters of cis-[Pt(NH3)2{r(GpG)}]+ with time, as observed in an MD simulation at 350 K. The helicity is measured as the sum of the dihedral angles between the base planes and the platinum coordination plane; positive values indicate a right-handed helicoidal arrangement. (From [89b].)
< previous page
page_21
next page >
< previous page
page_22
next page > Page 22
Fig. 3. Molecular models representing energy-minimized conformations of cis-[Pt(NH3)2{r(GpG)}]+ belonging to the conformational domains {1} (left) and {2} (right). The water molecule displayed in the conformation {2} was not present in the MD simulation but was added before energy minimization. (From [89b].) ~2.7 Å. For conformation {2}, the shortest H . . . O separation is ~4.2 Å, as expected from Fig. 2; this separation precludes a direct hydrogen bond but allows for water-mediated hydrogen bonding with nearly ideal geometry. Such hydrogen bonding was not possible in the MD simulations which were run "in vacuo" [the water molecule displayed in Fig. 3 (right) was added subsequently in order to visualize this possibility of indirect hydrogen bonding]. We believe that inclusion of the solvent in the MD simulations will favor the right-handed conformation {2}, since the latter will be stabilized by the water bridge. It is this watermediated hydrogen bonding, specific to the complex 2, to which we attribute the reversal of helicity from left-handed in 1 to right-handed in 2. All of the quoted work on platinum-oligonucleotide complexes demonstrates that neither energy minimizations nor MD simulations permit unequivocal structural predictions; nevertheless, they do allow spectroscopic data to be interpreted and visualized in terms of three-dimensional models. A compilation of force field parameters for the modeling of platinum complexes of guanine derivatives has recently been published [91].
< previous page
page_22
next page >
< previous page
page_23
next page > Page 23
4 Conclusion An established force field for molecular mechanics calculations on transition metal complexes with nucleic acids is presently not available. Two main obstacles have to be surmounted on the way to such a force field: First, an operational procedure for nucleic acids simulations in water has to be worked out; an improved treatment of longrange electrostatic interactions and the development of more realistic water models are probably the most urgent tasks. The recent brief account on MD simulations of protein, DNA, and RNA using the particle-mesh-Ewald method [42] possibly announces a breakthrough in this domain. Second, an energy functional allowing for the various coordination geometries occurring in transition metal complexes has to be implemented in an all-element force field. The ESFF force field may be a step in this direction. Although the present state of the art of molecular mechanics calculations on transition metal-nucleic acid complexes is less than satisfactory, it can be expected that the increasing need for structural models for this class of bioinorganic compounds will provide sufficient impetus for accelerated developments in the field. Acknowledgment The author is deeply indebted to Dr. Richard Lavery for numerous discussions and helpful comments. He expresses thanks to Drs. Shenghua Shi and Tom Thacher (Biosym Technologies) for providing a manuscript of the ESFF Project Report and for a very constructive discussion, to Drs. Joël Pothier and Jean-Claude Chottard for careful proofreading and critical comments on the manuscript, and to Pascale Augé, Frédéric Allain, and all the other students and colleagues for their contributions to the work quoted in Sec. 3 of this chapter. Computer time and facilities provided by the IDRIS center of the Centre National de la Recherche Scientifique and the excellent technical assistance of Dr. Jean-Marie Teuler are gratefully acknowledged.
< previous page
page_23
next page >
< previous page
page_24
next page > Page 24
Abbreviations
CD
circular dichroism
CI
configuration interaction
CPK
Corey-Pauling-Koltun
ECP
effective core potential
ESFF
extensible and systematic force field
Et
ethyl group
HF
Hartree-Fock
MD
molecular dynamics
Me
methyl group
MP2
Møller-Plesset second-order perturbation theory
NMR
nuclear magnetic resonance
NOE
nuclear Overhauser effect
NOESY
nuclear Overhauser spectroscopy
PCR
polymerase chain reaction
SCF
self-consistent field
References 1. G. L. Eichhorn (ed.), Inorganic Biochemistry, Vol. 2, Elsevier, New York, 1973, Chaps. 33 and 34. 2. M. Yarus, FASEB J., 7, 31 (1993). 3. R. C. Cadwell and G. F. Joyce, PCR Meth. Appl., 2, 28 (1992). 4. J. K. Barton and S. J. Lippard in Metal Ions in Biology, Vol. 1 (T. G. Spiro, ed.), Wiley-Interscience, New York (1980), p. 31 ff. 5. T. D. Tullius in Metal-DNA Chemistry (T. D. Tullius, ed.), ACS Symposium Series No. 402, American Chemical Society, Washington, DC, 1989, p. 1 ff.
6. B. Rosenberg, L. Van Camp, J. E. Trosko, and V. H. Mansour, Nature, 222, 385 (1969). 7. M. J. Bloemink and J. Reedijk in Metal Ions in Biological Systems, Vol. 32 (A. Sigel and H. Sigel, eds.), Marcel Dekker, New York, 1995, Chap. 19. 8. M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids, Clarendon Press, Oxford, 1989.
< previous page
page_24
next page >
< previous page
page_25
next page > Page 25
9. W. F. van Gunsteren and H. J. C. Berendsen, Angew. Chem. Int. Ed. Engl., 29, 992 (1990). 10. U. Dinur and A. T. Hagler, Rev. Comput. Chem., 2, 99 (1991). 11. J. A. McCammon and S. C. Harvey, Dynamics of Proteins and Nucleic Acids, Cambridge University Press, Cambridge, UK, 1989. 12. P. Comba, Coord. Chem. Rev., 123, 11 (1993). 13. B. P. Hay, Coord. Chem. Rev., 126, 177 (1993). 14. Baird and Tatlock, Price List of Chemical and Scientific Apparatus and Pure Chemicals, London, 1901. 15. M. J. H. Van't Hoff, Bull. Soc. Chim. Fr., 23, 295 (1875). 16. M. Le Bel, Bull. Soc. Chim. Fr., 22, 237 (1874). 17. D. A. Pearlman, D. A. Case, J. C. Caldwell, G. L. Seibel, U. C. Singh, P. Weiner, and P. A. Kollman, AMBER 4.0, University of California, San Francisco, 1991. 18. R. Lavery in Structure and Expression, Vol. 3, DNA Bending and Curvature (W. K. Olson, M. H. Sarma, M. S. Sarma, and M. Sundaralingam, eds.), Adenine Press, 1988, p. 191211. 19. B. Hingerty, R. H. Richie, T. L. Ferrel, and J. E. Turner, Biopolymers, 24, 427 (1985). 20. V. Fritsch and E. Westhof, J. Chim. Phys., 88, 2543 (1991). 21. J. C. Gilbert and C. Lemaréchal, Math. Program., 45, 407 (1989). 22. K. Zimmermann, J. Comput. Chem., 12, 310 (1990). 23. J. Pothier, J. Gabarro-Arpa, and M. Le Bret, J. Comput. Chem., 14, 326 (1993). 24. L. Verlet, Phys. Rev., 159, 98 (1967). 25. M. Sundaralingam, Biopolymers, 7, 821 (1969). 26. M. Nilges, G. M. Clore, A. M. Gronenborn, A. T. Brunger, M. Karplus, and L. Nilsson, Biochemistry, 26, 3718 (1987). 27. D. A. Pearlman and P. A. Kollman, J. Mol. Biol., 220, 457 (1991). 28. A. E. Torda, R. M. Scheek, and W. F. van Gunsteren, J. Mol. Biol., 214, 223 (1990). 29. W. F. van Gunsteren, H. J. C. Berendsen, R. G. Geurtsen, and H. R. J. Zwinderman, Ann. N.Y. Acad. Sci., 482, 287 (1986). 30. R. J. Loncharich and B. R. Brooks, Proteins: Struct. Funct. Genet., 6, 32 (1989). 31. W. B. Streett, D. J. Tildesley, and G. Saville, Mol. Phys., 35, 639 (1978).
< previous page
page_25
next page >
< previous page
page_26
next page > Page 26
32. F. Avbelj, J. Moult, D. H. Kitson, M. N. G. James, and A. T. Hagler, Biochemistry, 29, 8658 (1990). 33. B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, and M. Karplus, J. Comput. Chem., 4, 187 (1983). 34. K. Heinzinger and P. C. Vogel, Z. Naturforsch., 29a, 1164 (1974). 35. J. D. Madura and B. M. Pettitt, Chem. Phys. Lett., 150, 105 (1988). 36. P. Auffinger and D. Beveridge, Chem. Phys. Lett., 234, 413 (1995). 37. H. Schreiber and O. Steinhauser, Chem. Phys., 168, 75 (1992). 38. P. E. Smith and B. M. Pettitt, J. Chem. Phys., 95, 8430 (1991). 39. D. H. Kitson, F. Avbelj, J. Moult, D. T. Nguyen, J. E. Mertz, D. Hadzi, and A. T. Hagler, Proc. Natl. Acad. Sci. USA, 90, 8920 (1993). 40. H. Schreiber and O. Steinhauser, Biochemistry, 31, 5856 (1992). 41. D. M. York, T. A. Darden, and L. G. Pedersen, J. Chem. Phys., 99, 8345 (1993). 42. T. E. Cheatham III, J. L. Miller, T. Fox, T. A. Darden, and P. A. Kollman, J. Am. Chem. Soc., 117, 4193 (1995). 43. P. Ewald, Ann. Phys., 64, 253 (1921). 44. T. Darden, D. York, and L. Pedersen, J. Chem. Phys., 98, 10089 (1993). 45. S. N. Rao and P. Kollman, Biopolymers, 29, 517 (1990). 46. B. Tidor, K. K. Irikura, B. Brooks, and M. Karplus, J. Biomol. Struct. Dyn., 1, 231 (1983). 47. U. C. Singh, S. C. Weiner, and P. A. Kollman, Proc. Natl. Acad. Sci. USA, 82, 755 (1985). 48. D. A. Beveridge, S. Swaminathan, G. Ravishanker, J. Whitka, J. Srinivasan, C. Prévost, S. Louise-May, F. M. DiCapua, and P. H. Bolton in Advances in Biomolecular Simulations (R. Lavery, J. L. Rivail, and J. Smith, eds.), AIP Conference Proceedings No. 239, American Institute of Physics, 1991, p. 311 ff. 49. V. Daggett and M. Levitt, Annu. Rev. Biophys. Biomol. Struct., 22, 353 (1993). 50. M. Levitt, Chem. Scr., 29A, 197 (1989). 51. H. J. C. Berendsen, J. R. Grigera, and T. P. Straatsma, J. Phys. Chem., 91, 6269 (1987). 52. P. Ahlström, A. Wallqvist, S. Engström, and B. Jönsson, Mol. Phys., 68, 563 (1989). 53. E. J. Jacob, H. B. Thompson, and L. S. Bartell, J. Chem. Phys., 47, 3736 (1967).
< previous page
page_26
next page >
< previous page
page_27
next page > Page 27
54. T. W. Hambley, C. J. Hawkins, J. A. Palmer, and M. R. Snow, Aust. J. Chem., 34, 45 (1981). 55. B. H. Besler, K. M. Merz, and P. A. Kollman, J. Comp. Chem., 11, 431 (1990). 56. J. B. Foresman and A. E. Frisch, Exploring Chemistry with Electronic Structure Methods, Gaussian, Inc., Pittsburgh, 1993. 57. P. J. Hay and W. R. Wadt, J. Chem. Phys., 82, 270 (1985). 58. G. G. Messmer, E. L. Amma, and J. A. Ibers, Inorg. Chem., 6, 725 (1967). 59. A. Del Pra and G. Zanotti, Cryst. Struct. Commun., 8, 737 (1979). 60. J. Chatt and R. G. Wilkins, J. Chem. Soc., 4300 (1952). 61. H. Basch, M. Krauss, W. J. Stevens, and D. Cohen, J. Am. Chem. Soc., 25, 684 (1986). 62. J. Kozelka, R. Savinelli, G. Berthier, J. P. Flament, and R. Lavery, J. Comp. Chem., 13, 45 (1992). 63. E. Wimmer in Density Functional Methods in Chemistry (J. K. Labanowski and J. W. Andzelm, eds.), SpringerVerlag, New York, 1991, Chap. 2. 63a. A. K. Rappé, C. J. Casewit, K. S. Colwell, W. A. Goddard III, and W. M. Skiff, J. Am. Chem. Soc., 114, 10024 (1992). 64. S. Shi, L. Yan, Y. Yang, and J. Shaulsky, ESFF Forcefield Project Report, Biosym Technologies, San Diego, November 1994. 65. S. Shi, personal communication. 66. J. Kozelka, H.-P. Lüthi, E. Dubler, and R. W. Kunz, Inorg. Chim. Acta, 86, 155 (1984). 67. T. W. Hambley, Comm. Inorg. Chem., 14, 1 (1992). 68. S. E. Sherman and S. J. Lippard, Chem. Rev., 87, 1153 (1987). 69. K. J. Miller, E. R. Taylor, H. Basch, M. Krauss, and W. J. Stevens, J. Biomol. Struct. Dyn., 2, 1157 (1985). 70. J. Kozelka, G. A. Petsko, G. J. Quigley, and S. J. Lippard, J. Am. Chem. Soc., 107, 4079 (1985). 71. J. Kozelka, G. A. Petsko, G. J. Quigley, and S. J. Lippard, Inorg. Chem., 25, 1075 (1986). 72. L. L. Munchausen and R. O. Rahn, Biochim. Biophys. Acta, 414, 242 (1975). 73. J. H. J. den Hartog, C. Altona, J. H. van Boom, G. A. van der Marel,
< previous page
page_27
next page >
< previous page
page_28
next page > Page 28
C. A. G. Haasnoot, and J. Reedijk, J. Biomol. Struct. Dyn., 2, 1137 (1985). 74. J. Kozelka, S. Archer, G. A. Petsko, S. J. Lippard, and G. J. Quigley, Biopolymers, 26, 1245 (1987). 75. F. Herman, J. Kozelka, V. Stoven, E. Guittet, J. P. Girault, T. Huynh-Dinh, J. Igolen, J. Y. Lallemand, and J. C. Chottard, Eur. J. Biochem., 194, 119 (1990). 76. J. A. Rice, D. M. Crothers, A. L. Pinto, and S. J. Lippard, Proc. Natl. Acad. Sci. USA, 85, 4158 (1988). 77. S. F. Bellon and S. J. Lippard, Biophys. Chem., 35, 179 (1990). 78. S. L. McCarthy, R. J. Hinde, K. J. Miller, J. S. Anderson, H. Basch, and M. Krauss, Biopolymers, 29, 823 (1990). 79. S. L. McCarthy, R. J. Hinde, K. J. Miller, J. S. Anderson, H. Basch, and M. Krauss, Biopolymers, 29, 785 (1990). 80. K. J. Miller, S. L. McCarthy, and M. Krauss, J. Med. Chem., 33, 1043 (1990). 81. T. W. Hambley, Inorg. Chem., 30, 937 (1991). 82. M. H. Fouchet, PhD thesis, Unversité Paris 6, 1992. 83. K. Mazeau, F. Vovelle, A. Rahmouin, M. Leng, and M. Ptak, Anti-Cancer Drug Design, 4, 63 (1989). 84. C. J. van Garderen and L. P. A. van Houte, Eur. J. Biochem., 225, 1169 (1994). 85. M. Sip, A. Schwartz, F. Vovelle, M. Ptak, and M. Leng, Biochemistry, 31, 2508 (1992). 86. C. A. Lepre, L. Chassot, C. E. Costello, and S. J. Lippard, Biochemistry, 29, 811 (1990). 87. V. Brabec, M.Sip, and M. Leng, Biochemistry, 32, 11676 (1993). 88. M. D. Reily, T. W. Hambley, and L. G. Marzilli, J. Am. Chem. Soc., 110, 2999 (1988). 89. (a) P. Augé, F. Allain, J. P. Girault, J. A. H. Cognet, J. C. Chottard, and J. Kozelka, manuscript in preparation; (b) J. Kozelka in Proceedings of the Seventh International Symposium on Platinum and Other Metal Coordination Compounds in Cancer Chemotherapy (H. M. Pinedo and J. H. Schornagel, eds.), 1995, in press. 90. J. Kozelka, M. H. Fouchet, and J. C. Chottard, Eur. J. Biochem., 205, 895 (1992). 91. S. Yao, J. P. Plastaras, and L. G. Marzilli, Inorg. Chem., 33, 6061 (1994).
< previous page
page_28
next page >
< previous page
page_29
next page > Page 29
2 Zinc Complexes As Targeting Agents for Nucleic Acids Eiichi Kimura1 and Mitsuhiko Shionoya2 1Department of Medicinal Chemistry, Hiroshima University School of Medicine, 1-2-3 Kasumi, Minami-ku, Hiroshima 734, Japan 2Institute for Molecular Science, Okazaki National Research Institutes, Myodaiji, Okazaki 444, Japan
30
1. Introduction
30
1.1. Zinc Macrocyclic Polyamine Complexes
34
1.2. Consequences of the Acidifying Properties of Zinc
37
2. Zinc Complexes for Recognition of Nucleobases
2.1. A New Dimension to Supramolecular Chemistry with the Recognition of Nucleobases by Zinc Macrocyclic Polyamine Complexes
38
2.2. Three-Point Recognition of the Thymine Base by a ZnII-Cyclen Complex
2.3. Multipoint Recognition of Nucleobases by a ZnII Complex of an Acridine-Pendant Cyclen
43
46
3. Zinc Complexes for Control of Genetic Processes
3.1. Inhibition of Polynucleotide Hybridization by a ZnII-Cyclen Complex
< previous page
37
page_29
46
next page >
< previous page
page_30
next page > Page 30
3.2. Inhibition of in Vitro Poly(Phe) Synthesis by a ZnII-Cyclen Complex
49
49
4. Conclusions and Prospects
49
Abbreviations
50
References
1 Introduction 1.1 Zinc Macrocyclic Polyamine Complexes The ZnII ion has a very flexible and adaptable coordination sphere in solution and therefore it is called ''colorless chameleon" by Sigel and Martin [1]. For instance, the ZnII ion has a tetrahedral or distorted tetrahedral coordination sphere with the low coordination number of 4 to stabilize the structures of proteins and nucleic acids, whereas in aqueous solution the coordination number of ZnII is usually 6. The 5-coordinate ZnII species are also important in the catalytic active center of ZnII-containing enzymes as intermediates during the course of an enzymatic turnover. Moreover, the equilibria between 6-coordinate species are rapid. To specifically keep the lower coordination numbers in aqueous solution, sterically rigid and therefore well-designed ligands are required. Zinc(II) complexes with simple amine donors (L) are generally unstable in aqueous solution. Competition between L and H2O (or the stronger donor OH) must always be taken into consideration [Eq. (1)]. It is an extremely challenging task to design appropriate ligands that (1) form stable and inert complexes at physiological pH, (2) leave catalytic (i.e., ligand exchangeable) sites open on ZnII, and (3) offer environments structurally and/or functionally similar to those seen in biological systems.
Recently, macrocyclic polyamine ligands have been developed to prepare very stable metal complexes, due to the "macrocyclic effects", with well-defined coordination numbers and geometries [27]. This is clearly demonstrated by comparison of the stabilities of a series of ZnII-tetraamine complexes (Chart 1 and Table 1). We can be assured from its
< previous page
page_30
next page >
< previous page
page_31
next page > Page 31
Chart 1 extremely great stability constant that the macrocyclic tetraamines such as cyclen ([12]aneN4, 1,4,7,10tetraazacyclododecane) can firmly hold ZnII even at neutral pH in aqueous solution. Under such a circumstance, one may anticipate a well-defined coordination sphere and examine the reactivity at the fifth, i.e., at the catalytic site. The simplest way to measure the reactivity of ZnII was to determine its acidifying properties. Two examples are given: (1) the deprotonation of a water molecule coordinating to ZnII (Scheme la) and
Scheme 1
< previous page
page_31
next page >
< previous page
page_32
next page > Page 32
TABLE 1 Comparison of Stability Constants for ZnII-Tetraamine Complexes at 25°C Complex
Stability constant (log K)a
ZnII(NH3)4 ZnII(en)2c ZnII(trien)e ZnII-cyclen
9.32b 10.6 ± 0.1d 12.0f 15.3 ± 0.1g
aAt [NH3] = [en] = [trien] = [cyclen] = 1 M. bLog K = [ZnII(NH3)4]/[ZnII][NH3]4; Ref. 8. cen = ethylenediamine. d Log K = [ZnII(en)2]/[ZnII][en]2; Ref. 9. etrien = triethylenetetramine. fLog K = [ZnII(trien)]/[ZnII] [trien]; Ref. 10. gLog K = [ZnII-cyclen]/[ZnII] [cyclen]; Ref. 11. (2) the deprotonation of a phenolic OH, the macrocycle of which was specifically tailored to allow its coordination to ZnII (Scheme 1b). It is of interest to find that these deprotonations are greatly facilitated by the ZnII ion trapped in a tetraamine macrocycle, cyclam ([14]aneN4, ([14]aneN4, 1,4,8,11-tetraazacyclotetradecane); the pKa value of H2O dropped from 15.5 to 9.8 in 1 and the pKa of phenol from 10 to 5.8 in 3. This observation is also of interest in view of the fact that ZnII may be saturated with fourfold coordination, yet ZnII in cyclam still holds an appreciable further coordination tendency. Even more revealing regarding the acidifying properties of ZnII in less saturated systems is a 12-membered macrocyclic triamine, [12]aneN3 (1,5,9-triazacyclododecane). In comparison to the N4-coordinating system, the ZnII are 9.8 for the cyclam complex ion bound to N3 is more acidifying: the pKa values for ) vs. 7.3 for the [12]aneN3 complex ( ) under the same conditions (Scheme 2a) [12]. The phenol ( attached to [12]aneN3, 7, dissociates its proton with the pKa value of 6.8 to yield a trigonal bipyramidal complex 8, rather than an anticipated tetrahedral complex 10 with three nitrogen donors being equivalent at the basal position (Scheme 2b) [1214].
< previous page
page_32
next page >
< previous page
page_33
next page > Page 33
Scheme 2 This fact is a good experimental demonstration of the properties of ZnII on the basis of a calculation that predicts feasibility of a 5-coordinate, trigonal bipyramidal structure with shorter bonds for the three equatorial donors and longer bonds for the two axial donors. One of the three macrocyclic nitrogen donors elongates to an axial position to vacate an equatorial site for the anionic phenolate oxygen donor. This also implies that ZnII prefers an anionic donor to a neutral N donor. The H2O at the other axial position deprotonates with the higher pKa value of 10.7 to form 9. Use of macrocyclic polyamines as ligands for ZnII has the following advantages: (1) A systematic structural modification (such as ring size, the kind of donors atoms, donor numbers, etc.) is possible, so that one
< previous page
page_33
next page >
< previous page
page_34
next page > Page 34
can establish the optimum macrocyclic ligands. (2) ZnII-macrocyclic complexes in general are kinetically and thermodynamically stable, so that rigid and well-defined complex structures are ensured to which various additional functionalities can be attached. (3) Because of the extreme stabilities, ZnII-macrocyclic polyamine complexation is complete at low pH (which is not the case with linear polyamine complexes), and hence the distinctive properties of ZnII for the other donor(s); consequently, the pKa for ZnII-OH2 or the affinity of ZnII for anionic species may be estimated easily and accurately near neutral pH. 1.2 Consequences of the Acidifying Properties of Zinc The strong acidifying properties of ZnII in some of the macrocyclic polyamine complexes result in the facile generation of ZnII-OH species near neutral pH. The question then would arise as to whether these ZnII-OH species can behave as strong bases like Na+-OH or Mg2+-OH. Fortunately, ZnII-[12]aneN3 in the deprotonated form 6 was isolated from an aqueous solution of ZnII-[12]aneN3 at pH 8 in the presence of NaClO4. The X-ray crystal structure showed an extremely short ZnII-OH distance of 1.944 Å to be compared with 2.02 Å, the average of the ZnII-N bonds. From the pKa value of 7.3 for , one could calculate the affinity of OH for ZnII[12]aneN3 (5) in aqueous solution, log K = 6.4, which is the highest affinity among all the anionic donors. One may thus suspect some difficulty in the liberation of a free OH ion (naturally this would be solvated) from ZnIIOH, unlike in the cases with Na+-OH and Mg2+-OH, which have ionic bonds. Nevertheless it was quite a surprise to find that aromatic sulfonamides such as 1113 are subject to deprotonation near neutral pH in the presence of ZnII[12]aneN3, as disclosed by the UV spectral changes accompanying the generation of the conjugate bases of 1113 (Chart 2) [12,15]. Acetazolamide (11), which is the strongest acid (pKa = 7.5) among them, in fact can be isolated as a 1:1 complex (14) with ZnII-[12]aneN3 in its deprotonated form from aqueous CH3CN solution at neutral pH. The ease with which the sulfonamides 1113 were deprotonated with ZnII-OH near neutral pH would be understood if one assumes a strong interaction between ZnII and the conjugate base of the sulfonamides-SO2NH. Otherwise the relatively weak base ZnII-OH could
< previous page
page_34
next page >
< previous page
page_35
next page > Page 35
Chart 2 not sufficiently neutralize the weak acids 12 or 13 at neutral pH. Thus, the driving force for equilibrium (2) should come from the great stability
of the resulting 1:1 complexes (14) (Scheme 3). The facile deprotonation and simultaneous coordination of the aromatic sulfonamide anion in-
Scheme 3
< previous page
page_35
next page >
< previous page
page_36
next page > Page 36
deed occurs at pH
< previous page
page_37
next page > Page 37
Fig. 1. Plots of the pKa for a conjugate acid vs. logK(A). (Reproduced by permission from Ref. 15.) determine the apparent log K at appropriate pH. Accordingly, one would be convinced that at physiological or neutral pH, acetazolamide (pKa = 7.5) is the most potent CA inhibitor and
(pKa = 4.6) the most appropriate substrate for CA.
The discovery that ZnII-macrocyclic polyamine complexes can deprotonate sulfonamides led us to suspect that they might also deprotonate imides having similar pKa values. This proposal was successfully attested with the nucleobase thymine. 2 Zinc Complexes for Recognition of Nucleobases 2.1 A New Dimension to Supramolecular Chemistry with the Recognition of Nucleobases by Zinc Macrocyclic Polyamine Complexes Molecular recognition of DNA, RNA, and related biomolecules is responsible for a wide range of biochemical processes such as complemen-
< previous page
page_37
next page >
< previous page
page_38
next page > Page 38
tary base pairings in genetic information storage and transfer [17], or oligonucleotide recognition by ribozymes [18] and restriction enzymes [19], etc. Recently, a number of artificial organic receptor molecules were synthesized for nucleic acid constituents to mimic such biochemical processes [4,2025]. The molecular architecture of these receptors was basically instituted as an assembly of naturally occurring noncovalent binding elements, e.g., hydrogen bonding, hydrophobic or electrostatic interactions, etc. However, these interactions are not effective individually in aqueous solution for the association of the host-guest molecules unless they are in polymeric assemblies as seen in double-stranded DNA. In comparison, metal coordination, which may be a strong binding force, can serve as an effective binding element for host-guest interactions in aqueous solution. Therefore, metal complexes with well-designed ligands have a bright prospect in developing new targeting agents for biomolecules such as nucleic acids. The studies described above demonstrated how the acidifying property of ZnII is reinforced by the complexation with [12]aneN3 or cyclen, which results in the lowering of the pKa values of the ZnII-bound H2O from ~9 (for aquated ZnII ion) to 7.3 and 7.9, respectively, at 25°C [12]. The concept of strong acidifying property and anion binding ability of ZnII-macrocyclic polyamine complexes has been instrumental in discovering a novel type of nucleobase recognition with high selectivity in aqueous solution by a ZnII-macrocyclic tetraamine complex as described below. 2.2 Three-Point Recognition of the Thymine Base by a ZnII-Cyclen Complex The fundamental knowledge derived from the model study of zinc enzymes using ZnII-macrocyclic polyamine complexes has been successfully developed to a new aspect of the molecular recognition of nucleobases containing an ''imide" functionality (Chart 3). It was originally envisioned that, as shown in Fig. 2c, when a ZnII-cyclen complex (19) or its derivatives interact with the "imide" functionalities, the ZnII would first dissociate the "imide" proton to form a ZnII-N bond, where-upon the "imide" carbonyls with developing negative charges would
< previous page
page_38
next page >
< previous page
page_39
next page > Page 39
Chart 3 become better acceptors for the acidic NH hydrogens of cyclen at the complementary positions. This binding motif (involving inorganic supercomplexes) is quite different from those seen in biological systems (Fig. 2a) or in earlier artificial types of nucleobase receptor (involving only organic supercomplexes) (Fig. 2b). Indeed, strong association was found in a 1:1 ternary complex between the ZnII-cyclen complex (19) and AZT (azidothymidine), which was isolated from an aqueous solution at pH ~8.5 [26]. The X-ray crystal
< previous page
page_39
next page >
< previous page
page_40
next page > Page 40
Fig. 2. Three types of thymine recognition: (a) the base pairing seen in biological systems (b) an example of artificial organic host molecules [23] (c) our envisioned type. (Reproduced by permission from Ref. 26.) analysis of the ternary complex (20) revealed a distorted square pyramidal N5-coordinate structure with a strong interaction between ZnII and the N(3)-deprotonated ''imide" anion and the two complementary hydrogen bonds between cyclen NH groups and the "imide" carbonyls (Fig. 3). The ZnII-N(3) bond distance of 2.053(8) Å is shorter than the average ZnII-NH (cyclen) bond distance of 2.153 Å. We now can see a new type of three-point binding motif.
Potentiometric and spectroscopic titrations of deoxyribonucleosides, dA (2'-deoxyadenosine), dG (2'deoxyguanosine), dC (2'-deoxycytidine), and dT (thymidine), and related compounds in the presence of the ZnIIcyclen complex (19) disclosed highly selective 1:1 binding of 19 to dT (log K = 5.6, K = [ZnL-S]/([ZnL][S]) M1 at 25°C) and its derivatives, AZT (log K = 5.6), U (uridine, 5.2), and Ff (ftorafur, 4.6) (Fig. 4, line b). Other nucleosides containing an amino group in place of the carbonyl oxygen of dT (i.e., dG) or containing no acidic proton (i.e., dA and dC) did not bind at all to 19, possibly due to the nonbonding steric repulsion between the amino groups or to the lack of N anion formation. Compared to dT, Ino (inosine), which lacks one carbonyl group at the C(2) position of dT, showed a weaker affinity (log K = 4.2) for 19. It is therefore concluded that the additional two complementary hydrogen bonds between the "imide" carbonyl oxygens of thymine derivatives and the cyclen NH groups significantly contribute to the stability of the ternary complexes. This complementary three-point binding mode makes the binding of 19 to the thymine compounds highly selective in aqueous solution. The ZnII complex 19 has thus set a new prototype of host
< previous page
page_40
next page >
< previous page
page_41
next page > Page 41
Fig. 3. Interaction between ZnII-cyclen and AZT and X-ray crystal structure of the resulting complex 20. (Reproduced by permission from Ref. 26.)
< previous page
page_41
next page >
< previous page
page_42
next page > Page 42
Fig. 4. Plot of the complex formation constants of 19 and 21 at 25°C for N(3)-[or N(1)- for Ino] deprotonated nucleosides, log K(ZnL-S), against pKa values for the conjugate acids: (a) ( ) for 21; (b) ( ) for 19. (Reproduced by permission from Ref . 28.) molecules for DNA/RNA nucleobase recognition in a hydrophilic environment. Various transition metal ions and their complexes [27] have long been known to interact with nucleobases. The most nucleophilic site, N(7) of the guanine residue, tends to be a major target of these metal ions. For instance, cisPtII(NH3)2Cl2 and a number of its analogs bind to
< previous page
page_42
next page >
< previous page
page_43
next page > Page 43
the N(7) site of the guanine or adenine residue. Thymine and uracil bases, in general, offer no metal binding sites aside from the weakly coordinating carbonyl groups in the absence of the deprotonated imide anion. It is to be emphasized that the ZnII-cyclen complex (19) is exceptional in recognizing only the imide functionality of the thymine base. 2.3 Multipoint Recognition of Nucleobases by a ZnII Complex of an Acridine-Pendant Cyclen To further develop this novel type of nucleobase recognition in aqueous solution at physiological pH, the macrocyclic polyamines may be modified, with additional functional groups, capable of interacting with nucleobase and/or ribose moieties. In anticipation of a step forward to a more efficient ''multipoint" recognition, an acridine pendant has been introduced onto the cyclen ring (see 21) [28]. It was hoped that the aromatic acridine ring would produce an additional "π-π stacking interaction" with the aromatic nucleobases, whereby an even stronger interaction might occur in aqueous solution at physiological pH. Moreover, such a stacking agent might initially be inserted into double-stranded DNA to allow the subsequent interaction of the cationic ZnII-cyclen subunit with other DNA nucleophilic sites such as the negatively charged phosphates which are spaced along the DNA backbone. In 21, the three NH groups of the cyclen ring and the acridine pendant are spatially directed to the incoming substrate. It seemed reasonable to us to expect that if thymine homologs coordinated to ZnII at the deprotonated imide N(3), the two carbonyl oxygens would supplement the interaction by forming two hydrogen bonds with the two NH groups of the cyclen, and additionally the acridine ring would reinforce the complex stability by means of a π-π stacking interaction with the pyrimidine ring. To study the interaction between 21 and the "imide" functionality of nucleobases in aqueous solution, potentiometric pH titrations of 21 were conducted in the presence of dT or its homologs. The complex formation constants with 21 were all found to be greater than those with the original ZnII-cyclen complex (19) to support our prediction of an additional binding force from a π-π stacking interaction. The order of the affinities, dT (log K= 7.2), AZT (7.2) > BU (5-bromouridine, 7.0) > U
< previous page
page_43
next page >
< previous page
page_44
next page > Page 44
(6.9) > Ff (6.6) > AU (6-azauridine, 6.3), is consistent with that of the basicities of the conjugate base N(3) (Fig. 4; line a). Moreover, a linear relationship exists between the log K(ZnL-S) and pKa values of the conjugate acids. This fact indicates that 21 binds to all these nucleosides in the same manner with the ZnII-N coordinate interaction acting as the controlling force. Exactly the same trend was found for 19. The enhanced stability by ∆ log K of 1.6 ~ 2.0 can be translated into ∆∆G0 = 2.2 ~ 2.7 kcal/mol for the π-π interaction in aqueous solution. In addition, the two linear plots are not parallel (Fig. 4), which implies that the stability enhancement by the acridine ring varies with each pyrimidine structure. From the shallower slope for 21, it may be considered that the nucleobases that have the more acidic ''imide" group tend to enjoy more of a stabilizing contribution from the additional stacking interactions with the acridine. The electron-withdrawing substituent on BU or Ff decreases the thymine π-electron density, which may help to increase the π-π electronic interaction. As for Ino, the affinity with 21 is smaller (log K = 5.7) than that predicted from the amide pKa value of 8.75, as found with 19. This fact again supports the notion that direct and/or indirect hydrogen bonding by the two carbonyl groups of the thymine derivatives serves to supplement the thermodynamic stability of the ternary complexes in aqueous solution. 1H nuclear magnetic resonance (NMR) spectra of dT in the presence of 21 in aqueous media exhibit upfield shifts for a set of thymine and anomeric sugar protons as well as for the acridine protons, which points to an appreciable π-π stacking interaction accompanying the coordination of thymine to ZnII. In addition, the very slow deuterium exchange of the two NH groups of cyclen supports the notion of strong hydrogen bonding between the two NH groups in 21 and the two "imide" carbonyl oxygens of dT. The X-ray crystal structure of the 1:1 ternary complex of 21 with N(3)-deprotonated 1-methylthymine (22), is consistent with the structure concluded from the above solution behavior (Fig. 5). It is evident that 1-methylthymine firmly binds to ZnII in 21. The ternary complex (22) assumes a distorted square pyramidal structure with coordination from four nitrogens of the cyclen moiety and an N(3")-deprotonated "imide" anion of the pyrimidine ring. The most remarkable feature of the coordinate structure is the very short Zn(1)N(3") bond distance of
< previous page
page_44
next page >
< previous page
page_45
next page > Page 45
Fig. 5. Interaction of 21 with 1-methylthymine and X-ray crystal structure of 22. (Reproduced by permission from Ref. 28.)
< previous page
page_45
next page >
< previous page
page_46
next page > Page 46
1.987(4) Å. All three NH groups of the cyclen ring are spatially directed to the thymine base bound to the central ZnII ion. The carbonyl oxygen O(2'') of the pyrimidine ring forms a hydrogen bond directly with a cyclen NH group, while the other carbonyl oxygen O(4") binds indirectly via a water molecule to a diagonal NH group. The acridine lies face-to-face with the plane of the thymine substrate with the interplane separation ranging from 3.285 to 3.419 Å (normally ~3.4 Å for cofacial interaction), showing a well-arranged interfacial stacking between them. As for the interaction of 21 with other nucleosides, only dG interacts, but as a weaker guest (for details, see [28]). 3 Zinc Complexes for Control of Genetic Processes 3.1 Inhibition of Polynucleotide Hybridization by a ZnII-Cyclen Complex A central goal of designing DNA-or RNA-targeted agents is to manipulate the individual gene activities by agents that recognize and/or modify specific nucleic acid sequences. Such a compound would be one of the most promising tools for researchers at the interface between chemistry and molecular biology. As described above, the ZnII-cyclen complex (19) is a highly selective host in aqueous solution at physiological pH for dT and U among the DNA and RNA nucleosides, respectively [26]. Such a selective nucleobase receptor should inhibit or control some key genetic processes that involve molecular recognition via base pairing (Scheme 5). In accordance with this notion, we have examined the biochemical ability of 19 to regulate genetic processes which involve poly(U) [29]. Synthetic homopolymers poly(A) and poly(U), which separately are in random coil conformation in aqueous solution, associate together to form well-ordered double-stranded poly(A+U). This hybridization is detectable by a decrease in the optical density at 260 nm (A260). The effect of the ZnII-cyclen complex (19) on the hybridization of poly(A) and poly(U) to poly(A+U) was checked by the time course of A260 (Fig. 6a). As the concentration of 19 increased (r [19]added/[uracil]poly(U) = 0 ~ 2), the
< previous page
page_46
next page >
< previous page
page_47
next page > Page 47
Scheme 5 rate at which the absorbance lowered decreased. At r > 1, practically no hybridization occurred at all. Subsequent melting experiments were performed to examine the degree of hybridization (Fig. 6b). Without 19 the hybridized strand showed a single break in the A260 with Tm 42°C and hyperchromicity 31%. On the other hand, with increasing amounts of 19 (r < 2), the break in the absorbance plot shifted to lower temperatures with smaller hyperchromicities, until at r = 2 there was no break at all, which implies that the hybridization was completely inhibited at this ratio. An ultraviolet titration study showed that 48% of the total uracil bases in poly(U) formed a ternary complex with 19 at r = 2. A 31P NMR study provided additional evidence for the inhibition of poly(A)poly(U) hybridization by 19 (for details, see [29]). Virtually no significant interaction was observed between 19 and other synthetic poly(A), poly(G), and poly(C).
< previous page
page_47
next page >
< previous page
page_48
next page > Page 48
Fig. 6. (a) The time dependence of A260 for a 1:1 mixture of poly(A) and poly(U) at various concentrations of ZnII-cyclen 19 in 10 mM NaCl and Tris-HCl 5 mM (pH 7.6) at 25°C. (b) Melting temperature profiles of each sample succeeding the experiments of Fig. 6a. r = [19]added/[uracil]poly(U). (Reproduced by permission from Ref. 29.)
< previous page
page_48
next page >
< previous page
page_49
next page > Page 49
3.2 Inhibition of in Vitro Poly(Phe) Synthesis by a ZnII-Cyclen Complex The biochemical inhibitory effect of 19 on in vitro poly(Phe) (polyphenylalanine) synthesis in cell-free extracts was studied [29]. Poly(U) was used as a messenger RNA (where a UUU sequence is a codon for phenylalanine) in this experiment. In the presence of 0.558 mM (r = 1) and 5.58 mM of 19 (r = 10), the poly(Phe) synthesis underwent 15% and 26% inhibition, respectively. This inhibition reaction may possibly involve the formation of the ternary complex of 19 with uracil bases in poly(U). 4 Conclusions and Prospects Simple compounds that recognize and bind strongly to a specific nucleobase or sequence to compete with natural gene control elements might make new prototypes for biochemical as well as chemical probes for nucleic acids. For this purpose, transition metal complexes are quite advantageous and are currently attracting great interest. In this chapter, we demonstrated that the ZnII-cyclen complex and its functionalized derivatives are a new type of ideal receptor molecules that in neutral aqueous solution recognize and bind selectively and reversibly to the thymine base and its homologs. Recently, we also discovered that the same ZnII-cyclen complex induces a dramatic transition from right-handed B form to putative left-handed Z form of poly(dG-dC)·poly(dG-dC) at very low concentration [30]. These findings suggest that ZnII in macrocyclic polyamine complexes may become a new tool to control some gene expressions involving base pairing processes with single-stranded nucleic acids as well as chemical probes of unusual DNA structures. Abbreviations
[12]aneN3
1,5,9-triazacyclododecane
AU
6-azauridine
< previous page
page_49
next page >
< previous page
page_50
next page > Page 50
AZT
azidothymidine
BU
5-bromouridine
CA
carbonic anhydrase
cyclam
[14]aneN4, 1,4,8,11-tetraazacyclotetradecane
cyclen
[12]aneN4, 1,4,7,10-tetraazacyclododecane
dA
2'-deoxyadenosine
dC
2'-deoxycytidine
dG
2'-deoxyguanosine
dT
thymidine
en
ethylenediamine
Ff
ftorafur
Ino
inosine
NMR
nuclear magnetic resonance
poly(A)
polyadenylic acid
poly(C)
polycytidylic acid
poly(G)
polyguanylic acid
poly(Phe)
polyphenylalanine
poly(U)
polyuridylic acid
trien
triethylenetetraamine
References 1. H. Sigel and R. B. Martin, Chem. Soc. Rev., 23, 83 (1994). 2. E. Kimura and T. Koike, Comm. Inorg. Chem., 11, 285 (1991).
3. E. Kimura, Tetrahedron, 48, 6175 (1992). 4. E. Kimura in Crown Ethers and Analogous Compounds (M. Hiraoka, ed.), Elsevier, Amsterdam, 1992, p. 381 ff. 5. E. Kimura in Crown Compounds Toward Future Applications (S. R. Cooper, ed.), VCH, New York, 1992, p. 81 ff. 6. E. Kimura in Progress in Inorganic Chemistry, Vol. 41 (K. D. Karlin, ed.), John Wiley and Sons, New York, 1994, p. 443 ff. 7. E. Kimura and M. Shionoya in Transition Metals in Supramolecular Chemistry, Vol. 448 (L. Fabbrizzi and A. Poggi, eds.), Kluwer Academic, Dordrecht, 1994, p. 245 ff. 8. In Stability Constants of Metal-Ion Complexes, Part A (E. Högfeldt, ed.), Pergamon Press, Oxford, 1982.
< previous page
page_50
next page >
< previous page
page_51
next page > Page 51
9. In Critical Stability Constants, Vol. 6 (R. M. Smith and A. E. Martell, eds.), Plenum Press, New York, 1989. 10. G. Anderegg and P. Blauenstein, Helv. Chim. Acta, 65, 913 (1982). 11. T. Koike, S. Kajitani, I. Nakamura, E. Kimura, and M. Shiro, J. Am. Chem. Soc., 117, 1210 (1995). 12. E. Kimura, T. Shiota, T. Koike, and M. Shiro, J. Am. Chem. Soc., 112, 5805 (1990). 13. E. Kimura, T. Koike, and K. Toriumi, Inorg. Chem., 27, 3687 (1988). 14. For ZnII-[12]aneN3 appended with an imidazole pendant, see E. Kimura, Y. Kurogi, M. Shionoya, and M. Shiro, Inorg. Chem., 30, 4524 (1991). 15. T. Koike, E. Kimura, I. Nakamura, Y. Hashimoto, and M. Shiro, J. Am. Chem. Soc., 114, 7338 (1992). 16. F. Botre, G. Gros, and B. T. Strorey (eds.), Carbonic Anhydrase, VCH, New York, 1990. 17. B. Alberts, D. Bray, J. Lewis, M. Raff, K. Roberts, and J. D. Watson in The Molecular Biology of Cell, Garland, New York, 1983, p. 437 ff. 18. T. R. Cech, Science, 236, 1532 (1987). 19. J. A. McClerin, C. A. Frederick, B.-C. Wang, P. Greene, H. W. Boyer, J. Grable, and J. M. Rosenberg, Science, 234, 1526 (1986). 20. E. Kimura, M. Kodama, and T. Yatsunami, J. Am. Chem. Soc., 104, 3182 (1982). 21. E. Kimura, Top. Curr. Chem., 128, 131 and 141 (1985). 22. M. W. Hosseini, A. J. Blacker, and J.-M. Lehn, J. Am. Chem. Soc., 112, 3896 (1990) and references cited therein. 23. A. D. Hamilton, in Bioorganic Chemistry Frontiers, Vol. 2 (H. Dugas, ed.), Springer-Verlag, Berlin, 1991, p. 115 ff. 24. M. M. Conn, G. Deslonchamps, J. de Mendoza, and J. Rebek, Jr., J. Am. Chem. Soc., 115, 3548 (1993) and references cited therein. 25. U. Kral, J. Sessler, and H. Furuta, J. Am. Chem. Soc., 114, 8704 (1992). 26. M. Shionoya, E. Kimura, and M. Shiro, J. Am. Chem. Soc., 115, 6730 (1993). 27. For example: H. Sigel, Chem. Soc. Rev., 22, 255 (1993) and references cited therein.
< previous page
page_51
next page >
< previous page
page_52
next page > Page 52
28. M. Shionoya, T. Ikeda, E. Kimura, and M. Shiro, J. Am. Chem. Soc., 116, 3848 (1994). 29. M. Shionoya, M. Sugiyama, and E. Kimura, J. Chem. Soc., Chem. Commun., 1747 (1994). 30. M. Shionoya, E. Kimura, H. Hayashida, G. Petho, and L. G. Marzilli, Supramol. Chem., 2, 173 (1993).
< previous page
page_52
next page >
< previous page
page_53
next page > Page 53
3 Metallocene Interactions with DNA and DNA-Processing Enzymes Louis Y. Kuo,1 Andrew H. Liu,2 and Tobin J. Marks3 1Department of Chemistry, Lewis and Clark College, 0615 S.W. Palatine Hill Road, Portland, OR 97219, USA 2DuPont Chemicals, Jackson Lab. Chambers Works, Deepwater, NJ 08023, USA 3Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208-3113, USA
54
1. Introduction
54
1.1. Metallocene Synthesis, Descriptive Chemistry, Molecular and Electronic Structure
56
2. Metallocene Carcinostatic Activity
56
2.1. In Vivo and in Vitro Results
58
2.2. In Vivo Mechanistic Information
59
3. Metallocene-DNA Coordination Chemistry
59
3.1. Aqueous Metallocene Hydrolysis Chemistry
3.2. Metallocene Binding to Nucleobases and Nucleotides. NMR and X-ray Crystallographic Studies
3.3. Metallocene-Oligo/Polynucleotide Interactions
4. Metallocene Interactions with DNA-Processing Enzymes
4.1. Endonucleases, Polymerases, Kinases, and Topoisomerases
< previous page
page_53
62
70
75
75
next page >
< previous page
page_54
next page > Page 54 78
5. Conclusions and Outlook
79
Acknowledgments
79
Abbreviations
79
References
1 Introduction 1.1 Metallocene Synthesis, Descriptive Chemistry, Molecular and Electronic Structure The discovery of a metallocene-based organometallic carcinostatic agent in 1979 by Köpf and Köpf-Maier [1] stimulated much interest in the mechanism of the carcinostatic activity. Indeed, the remarkable potency of these compounds (vide infra) warrants a deeper mechanistic understanding that may provide the knowledge base for the rational design and implementation of other, even more potent organotransition metal pharmaceuticals. Accordingly, this chapter summarizes the fundamental chemical and physicochemical properties of metallocene complexes that are relevant to their cytostatic activity. In this regard, there is biological evidence, which we review herein, that has been interpreted in terms of parallels between the carcinostatic mechanism of metallocenes and that of the well-known antitumor agent cis-dichloro-diammineplatinum(II) [cis-(NH3)2PtCl2, ''cisplatin"] [24]. We therefore highlight basic chemical information that addresses this issue and suggest that in spite of apparent biological parallels with cisplatin, the metallocene compounds display fundamentally different coordination chemistry and likely operate via very different carcinostatic pathways than cisplatin. The two basic classes of metallocene-based compounds that have been reported to exhibit antitumor activity are ferricenium salts and metallocene diacido complexes. The ferricenium salts are of the form
where Cp = η5-
C5H5, , 2,4,6-(NO2)3C6H2O [5], or CCl3CO2·CCl3COOH [5,6]. Here the central iron(III) is coordinated to two cyclopentadienyl (Cp) ligands in the classic D5d/D5h "sandwich" configuration (Fig. 1). The coordinatively saturated ferricenium salts are water-soluble, hydrolytically stable, relatively strong oxidizing
< previous page
page_54
next page >
< previous page
page_55
next page > Page 55
Fig. 1. (A) Structures of metallocene dihalides and the ferricenium cation [Fe(η5-C5H5)2]+ with H atoms omitted. (B) Qualitative orbital representation of the frontier d orbitals in the equatorial plane of the bent M(η5-C5H5)2 moiety. agents and are reported to inhibit the growth of Ehrlich ascites and other experimental solid tumor systems [7,8]. However, the paucity of biological coordination chemical information on these complexes as well as the lack of relevant information on the mode of action leaves their carcinostatic mechanism open to speculation. Clearly, more studies are needed. In contrast, there is far more known about the aqueous coordination chemistry and details of biological action for the groups 46 metallocene diacido complexes (Cp2MX2) that provides a foundation for mechanistic discussion of the cytostatic activity. The metallocene diacido complexes are part of a large class of ''bent sandwich compounds" that have played a major role in the organometallic chemistry of early and middle transition metals [9]. The central metal is in a distorted tetrahedral coordination geometry with two π-bonded Cp ligands and two other ligands bound in a cis configura-
< previous page
page_55
next page >
< previous page
page_56
next page > Page 56
tion (Fig. 1). In a molecular orbital description, the electronic structure of bent metallocenes can be understood by Cp-M-Cp below 180° rehybridizes three of the five distorting a D5h molecular structure [9,10]. Contracting metal d orbitals such that the frontier orbitals are directed to the open side of the Cp2M ''wedge" [10] and, depending on the d electron count, the Cp2M moiety exhibits variable acid-base reactivity. For example, group 4 Cp2MCl2 complexes (M = Ti, Hf, Zr) have an empty central frontier orbital that renders the complexes Lewis acidic while the group 6 (M = Cr, Mo, W) complexes have a filled orbital and frequently display Lewis basic character. This explains the dramatic difference in the coordination chemistry of the group 4 and group 6 metallocenes wherein the former tend to bind to π-basic ligands such as OR while the latter tend to bind π-acceptor ligands such as ethylene [10,11]. The fundamental differences in the coordination chemistry of these two classes of metallocenes can also be rationalized in terms of hard-soft-acid-base (HSAB) principles, which will be addressed later in this chapter in the context of biomolecular ligands. Cp2MX2 complexes are typically synthesized from their respective metal halide salts and cyclopentadienide reagents (e.g., NaCp) [1214]. The bent-sandwich shape of the metallocene moiety can accommodate sterically demanding anionic and neutrally charged nonhalide ligands within the "equatorial plane" of the Cp2M wedge. Indeed, the large number of known unique Cp2MLn complexes [15] with various ligands L attest to the chemical versatility of these systems. That the Cp ligands can be derivatized with a variety of substituents further underscores the chemical flexibility of the metallocene diacido complexes. The synthetic variations have included alkylation, chelating bridges, and the attachment of chiral auxiliaries [15]. Taken together, the variability of acido and cyclopentadienyl ligands further adds to the diversity of complexes that could serve as potential pharmaceuticals. 2 Metallocene Carcinostatic Activity 2.1 In Vivo and in Vitro Results Cp2MX2 complexes show impressive in vivo and in vitro carcinostatic activities. At optimal doses, Cp2MX2 complexes where M = Ti(IV), V(IV),
< previous page
page_56
next page >
< previous page
next page >
page_57
Page 57 Nb(IV), and Mo(IV) and X = F, Cl, Br, I, NCS, N3, and OC(O)R [1,1642] have been shown to be highly active agents against Ehrlich ascites tumor (EAT) [1,18,23,28,29], sarcoma 180 [20,21,25], B16 melanoma [20,21,25], colon 38 adenocarcinoma [22,25], Lewis lung carcinoma [26], and human epidermoid (HEp-2) [3436] tumor cells. In addition, these complexes show impressive antitumor activity against human colon [17,24,27,30] and gastrointestinal [17,24] carcinomas and lung [17,31] malignancies xenografted in athymic mice. In comparison to cisplatin [2,3], Cp2VCl2 was shown to be 87% as cytotoxic against HEp-2 carcinoma under identical conditions [36]. Furthermore, in vivo tests show that at equitoxic dose levels Cp2VCl2 is as effective as cisplatin in prolonging the survival of strain A mice implanted with murine mammary adenocarcinoma (TA3Ha) [36]. In terms of-toxicity, strain A mice treated with Cp2VCl2 and Cp2TiCl2 show no evidence of renal or small intestinal damage whereas cisplatin exhibits significant renal damage [25,37,38]. Furthermore, bone marrow function, which is markedly reduced after cisplatin treatment, is only slightly affected by Cp2TiCl2 [20,3739], and while some symptoms of hepatotoxicity are observed, they are found to be transient and reversible [21,3639]. Ionic derivatives of Cp2MCl2 [4043] have greater water solubility and also exhibit pronounced antitumor activity against Ehrlich ascites and several experimental solid tumor systems. These agents include [Cp2MCl2]n+[X]n systems in high oxidation states [42,43] [M = Re(V), Nb(V), and Mo(VI), ], and Cp2TiCl2 derivatives of the form [Cp2TiCl(NCCH3)]+[FeCl4],
,
,
, ,
, [Cp2Ti(o-S(NACH3)C6H4)]+[I] (bipy = 2,2'-bipyridyl, phen = o-phenanthroline, o-S(NACH3)C6H4 = N-methyl-o-aminothiophenol), and [Cp2Ti(bisdithiolenyl)][N(C2H5)4]+ [41]. In contrast, the Zr and Hf metallocene dichlorides have negligible tumor-inhibiting properties in terms of prolonging the survival of mice infected with Ehrlich tumor cells [23]. Furthermore, free cyclopentadiene and dicyclopentadiene do not effect the same systemic antitumor activity as the η5-Cp metal complexes. Neither hydrocarbon significantly inhibits tumor growth in mice bearing solid EAT even at 10 times the metallocene dichloride dosages [44]. Most structure-activity studies have been carried out with Cp2TiX2 complexes where the Cp and acido ligands were sequentially modified.
< previous page
page_57
next page >
< previous page
page_58
next page > Page 58
Titanocene derivatives with F, Br, I, NCS as the acido ligands are as potent as Cp2TiCl2 against EAT cells with optimum cure rates of 100% [45]. Interestingly, an optimum cure rate of 100% is also observed with titanocene derivatives of carboxylates [21,22], aminothiophenolates [21,22], phenolates [46], thiophenolates [46], selenophenolates and dithiolenes [46]. It has also been shown that complexes of Cp2TiCl2 with adriamycin exhibit antitumor activity against P-388 leukemia [47]. However, any modification of the Cp2TiCl2 Cp ligands as shown in Fig. 2 reduces the antitumor potency against fluid EAT [48]. 2.2 In Vivo Mechanistic Information That the clinically successful cisplatin inhibits DNA replication and mitotic activity has generated considerable interest in the interaction of cis-diammineplatinum(II) with DNA and nucleic acid constituents [4,4952]. There is a considerable body of evidence suggesting that the carcinostatic activity of cisplatin is due to discrete (NH3)2Pt-DNA adducts (vide infra) that induce DNA lesions and inhibit DNA replication
Fig. 2. Structures of some Cp2TiCl2 derivatives having substituted cyclopentadienyl ligands tested for antitumor activity.
< previous page
page_58
next page >
< previous page
page_59
next page > Page 59
necessary for tumor cell division [5052]. Moreover, these particular adducts have been implicated as a structural motif targeted by the repair mechanism of tumor cells that are resistant to cisplatin chemotherapy [49]. It has also been proposed that the primary biological target of Cp2MX2 complexes is DNA [17,18]. Thus, morphological alterations of treated tumor cells suggest that the primary Cp2MX2 biological target is the cellular nucleus [53,54]. This notion is also supported by electron energy loss studies (EELS) which show an accumulation of the respective metals (Ti and V) in the nuclear heterochromatin of the tumor cells [17,55]. Furthermore, experiments with tritium-labeled precursors of DNA, RNA, and proteins showed that DNA synthesis is significantly and persistently depressed in EAT cells while RNA and protein synthesis is only slightly and reversibly inhibited [56,57]. Interestingly, UV spectroscopic studies also suggest an interaction between aqueous Cp2TiCl2 and Cp2VCl2 and nucleic acids that manifests itself as an alteration in the secondary structure of the polynucleotide [58]. 3 Metallocene-DNA Coordination Chemistry 3.1 Aqueous Metallocene Hydrolysis Chemistry While there appear to be analogies in the biological modes of action between Cp2MX2 and cisplatin, the two classes of complexes exhibit dramatically different aquation and coordination chemistries. Investigation of the aqueous properties of Cp2MCl2 complexes (M = Ti, Zr, V, Mo) focused first on understanding the hydrolytic stability of the M-Cp and M-Cl linkages [59,60], which better defines the nature of aquated Cp2MCl2 species obtained upon dissolution. The rates of appearance of free cyclopentadiene, as monitored by 1H NMR spectroscopy, show that Cp2VCl2 and Cp2MoCl2 are relatively resistant to ring protonolysis [Eq. (1)]. After 26 days, the CpH/metal ratio for both Cp2VCl2 [59] and
< previous page
page_59
next page >
< previous page
page_60
next page > Page 60
Cp2MoCl2 [60] is
< previous page
page_64
next page > Page 64
relatively nonlabile bonds with donor atomic sites of nucleotides and nucleobases [60]. As the basis for understanding Cp2Mo2+ coordination to mononucleotides, Cp2Mo2+ complexes of 9-methyladenine and 1-methylcytosine were prepared [Eq. (2)] and characterized by NMR spec-
troscopy and X-ray diffraction. The reaction of Cp2MoCl2(aq) with 9-methyladenine yields two [Cp2Mo(9methyladenyl)][PF6] isomers that represent kinetic and thermodynamic reaction products. 1H NMR indicates that both complexes have magnetically equivalent Cp ligands, and the chemical shift displacements for the nucleobase protons of [Cp2Mo(9-methyladenyl)][PF6] relative to 9-methyladenine suggest HN6/N1 and HN6/N7 chelation modes for the kinetic and thermodynamic products, respectively [60] (Scheme 2). The reaction of
Scheme 2 Cp2MoCl2(aq) with 1-methylcytosine yields a single [Cp2Mo-(1-methylcytosyl)][PF6] product with 1H NMR spectral parameters that indicate magnetically equivalent Cp ligands and suggest N4/N3 chelation (Scheme 2). X-ray diffraction studies [60] were carried out on the kinetic isomer of [Cp2Mo(9-methyladenyl)][PF6] (Fig. 4) and on [Cp2Mo(1-methylcytosyl)[PF6] (Fig. 5). In both complexes, the Mo(IV) ion adopts the familiar ''clam shell" geometry found in many Cp2MoXn complexes [6873] with each Mo(IV) ion π-bonded to two staggered η5-C5H5 ligands and σ-bonded to two nitrogen atoms of the coordinated nucleobase. In
< previous page
page_64
next page >
< previous page
page_65
next page > Page 65
Fig. 4. Perspective views (ORTEP) of the nonhydrogen atoms of the [Cp2Mo(9-methyladenyl)][PF6] kinetic product, showing the cationic portion. Thermal ellipsoids are drawn to include 30% probability. (Reproduced by permission from Ref. 60.)
< previous page
page_65
next page >
< previous page
page_66
next page > Page 66
Fig. 5. Perspective views (ORTEP) of the nonhydrogen atoms of [Cp2Mo(1-methylcytosyl)][PF6] showing the cationic portion. Thermal ellipsoids are drawn to include 30% probability. (Reproduced by permission from Ref. 60.)
< previous page
page_66
next page >
< previous page
page_67
next page > Page 67
accordance with 1H NMR results, the alkylated nucleobase is situated in the equatorial plane of the metallocene wedge and bisects the Cp-Mo-Cp angle to produce magnetically equivalent Cp ligands. Moreover, the sites of metal coordination to the nucleobases are in agreement with NMR results which confirm Mo(IV) ion binding to both N6 and N1 in [Cp2Mo(9-methyladenyl)][PF6], and to N4 and N3 for [Cp2Mo(1-methylcytosyl)][PF6] [60]. The metal coordination to both alkylated nucleobases involves a four-membered Mo(IV) chelate ring, deprotonation of one amino proton, and simultaneous coordination to both the endo- and exocyclic nitrogen atoms of the nucleobase. In light of the fact that metals typically coordinate to the endocyclic N3 and O2 atoms of pyrimidines and to the N7 atom of purines [64,7476], the crystallographically characterized Cp2Mo2+ binding mode to the alkylated nucleobases is unprecedented [7779], and the four-membered chelation appears strained. However, such four-membered chelation is seen in several nonnucleobase Cp2MoXn systems [6873], and according to theoretical studies, of the d0, d1, and d2 Cp2MX2 systems, d2 complexes should exhibit the smallest X-M-X bond angles (7682°) [10]. This prediction is in accord with the propensity of characterized Cp2Mo2+ chelation modes with alkylated nucleobases. 1H and 31P NMR studies [60] of Cp2Mo2+-nucleotide interactions (pH 7.4) show that Cp2Mo2+ coordinates in a covalent fashion to the N7 and O(phosphate) sites of purine mononucleotides to form 1:1 Cp2Mo-(nucleotide) adducts. Characteristic spectral features for this coordination mode include a 33 ppm downfield displacement of the 31P NMR signal, 0.30 and +0.10 ppm displacements of the H8 and H2 1H signals, respectively, and the appearance of nonequivalent Cp ligands. Aqueous cryoscopy and fast atom bombardment mass spectroscopy (FAB/MS) show that the Cp2Mo(5'-dAMP) complex is monomeric while X-ray diffraction studies reveal a dimeric Cp2Mo(5'-dGMP) adduct (vide infra). The analogous complexes with 5'-dCMP and 5'-dTMP likely have similar structures that involve a covalent bond to the N3 and O(phosphate) positions (Fig. 6) [60]. Interestingly, Cp2Mo2+ interactions with the methyl phosphate ester of 5'-dGMP [Me(5'-dGMP)] involve N7 coordination but no direct Cp2Mo2+ bonding to the phosphodiester group [60]. This suggests either a water bridge between Cp2Mo2+ and the phosphodiester moiety or no interaction (Fig. 6). 1H NMR studies show that Cp2MoCl2(aq) exhibits little or no coordinative
< previous page
page_67
next page >
< previous page
page_68
next page > Page 68
Fig. 6. Structures of likely Cp2Mo2+ binding modes to (A) 5'-dAMP (B) 5'-dTMP (C) 5'-dCMP (D) Me(5'-dGMP).
< previous page
page_68
next page >
< previous page
page_69
next page > Page 69
nucleotide selectivity [60] and causes, as in the case of Cp2VCl2(aq), little or no disruption in the Watson-Crick base pairing. Furthermore, the covalent bonding in the Cp2Mo2+-nucleotide complex is weak as it can be disrupted by the presence of other free nucleotides. To date, there is only one crystal structure of a Cp2M2+-nucleotide complex isolated from aqueous media (pH 7.4). The complex [Cp2Mo(2'-deoxyguanosine 5'-monophosphate)]2 can be obtained on a preparative scale [Eq. (3)], and the X-ray structural analysis reveals discrete dimers
having 24-membered chelate rings. Moreover, this analysis reveals features that are consistent with the solution NMR results, which include two magnetically distinct sets of Cp ligands related by a pseudo-inversion center (Fig. 7). The Cp2Mo2+ moiety is metrically unexceptional and shows simultaneous covalent binding to N7 of the guanine base and O(phosphate) of different nucleotides as suggested by the NMR results. The only conformational changes 5'dGMP undergoes upon coordination to Cp2Mo2+ include a C3'-endo sugar pucker and a high syn-glycosidic conformation [80]. Studies with Cp2TiCl2 [61] reveal that complexation to nucleobases is weak (510% complexation) in water and that interactions with nucleotides (pH 24) exhibit 1H and 31P NMR characteristics that imply simultaneous binding to base and O(phosphate). However, the lability of the Cp-Ti bond precludes isolation and full characterization of any complex. Moreover, at pH >6, little or no complexation of Cp2TiCl2(aq) to nucleotides or nucleobases is observed [61]. NMR studies also show that Cp2ZrCl2 and Cp2HfCl2, which have water-labile M-Cp ligation, do not form discrete complexes with nucleic acid constituents [61]. Cp2Tin+ complexes of purine and theophylline have been synthesized in nonaqueous media and characterized by Xray crystallography. For the Cp2Ti(IV)Cl(purinato) complex [81], the Cp2TiCl+ moiety binds to the N9 position, and for the Cp2Ti(III)-(theophyllinato) complex [82], the Cp2Ti+ moiety binds to both the O6 and N7 sites. Both complexes have metrically unexceptional Cp2Ti parameters and show the planes of the purine and theophylline ligands to be situated in the equatorial plane of the Cp2Ti2+ wedge, reminiscent of the aforementioned Cp2Mo(nucleobase) complexes. However, the nonaqueous method of preparing these complexes and their hydrolytic instability
< previous page
page_69
next page >
< previous page
page_70
next page > Page 70
Fig. 7. Perspective view of the nonhydrogen atoms of [Cp2Mo(5'-dGMP)]2 showing the 24-membered chelate ring and dimeric molybdenum coordination. The water molecules surrounding the structure are omitted for clarity, and all atoms are thermal vibrational ellipsoids drawn to include 30% probability. (Reproduced by permission from Ref. 60.) leave in doubt whether they are accurate representations of titanocene-nucleobase/nucleotide/DNA interactions under physiological conditions. 3.3 Metallocene-Oligo/Polynucleotide Interactions The crystallographic and NMR results for Cp2MoCl2(aq) clearly show that the Cp2Mo2+ moiety can covalently bind to both the base and
< previous page
page_70
next page >
< previous page
page_71
next page > Page 71
phosphate groups of 2'-deoxymononucleotides. Simultaneous binding to these sites in short oligodeoxynucleotides by other metals, including platinum, is well documented [8386]. This raises the question of whether Cp2MoCl2(aq) can bind to nucleic acids in a similar fashion. Accordingly, the interaction of Cp2MoCl2(aq) with the synthetic, selfcomplementary hexadeoxynucleotide d(ApGpGpCpCpT) and with the 5'-phosphorylated analog d(pApGpGpCpCpT) was studied by NMR [87]. This particular sequence was selected because it was used in a similar study [88,89] with cisplatin and would allow direct comparisons. The NMR spectra (D2O, pH = 7.0) of d(ApGpGpCpCpT) show at most minor changes upon addition of Cp2MoCl2(aq). However, the spectra of d(pApGpGpCpCpT) show significant pertubations at the 5' end that are consistent with simultaneous Cp2Mo2+ coordination to the N7 and O(phosphate) positions of the 5'-terminal adenosine (Figs. 8 and 9). These features, which include a 33 ppm downfield displacement of the 5'-terminal phosphate, 0.29 and +0.10 ppm displacements of the H8 and H2 1H signals, respectively, and the appearance of nonequivalent Cp ligands, are reminiscent of those observed upon complexation of Cp2Mo2+ to dAMP [60]. Together with the spectroscopic invariance of the remaining nucleobases, these results suggest that binding of Cp2MoCl2(aq) is through the phosphate-base and not through base-base coordination (e.g., to N7 sites of adjacent guanosines) as was found for cisplatin [88,89]. Consistent with this observation is that Cp2MoCl2(aq) has no effect on the electrophoretic mobility of supercoiled DNA plasmids [87]. This again is in contrast with the results of analogous experiments performed with cisplatin, where binding to DNA induces unwinding and consequently a reduction in the electrophoretic mobility of DNA plasmids [90]. Oxygen(phosphate) binding by Cp2Mo2+ has been observed in studies with larger (150200 base pairs) DNA fragments [91]. 31P NMR studies show that the interaction of Cp2MoCl2(aq) with sonicated calf thymus DNA yields new, minute signals (
< previous page
page_72
next page > Page 72 Page 73
Fig. 8. 1H (400-MHz) NMR spectra showing the interaction of Cp2MoCl2(aq) (3 equivalents) with d(ApGpGpCpCpT) (panels A and C) and with the 5'-phosphate derivative d(pApGpGpCpCpT) (panels B and D) in D2O (pD =7.0, 20°C). The resonances at 8.05 and 8.45 ppm in the spectrum of d(pApGpGpCpCpT) + Cp2MoCl2(aq) are unidentified byproduct(s).
< previous page
page_72
next page >
< previous page
page_74
next page > Page 74
Fig. 9. 31P (161 MHz) NMR spectra of the synthetic hexadeoxynucleotide (A) d(ApGpGpCpCpT) and the 5'-phosphate derivative (B) d(pApGpGpCpCpT), and the products of their reactions with 3 equivalents of Cp2MoCl2(aq) (panels C and D). The resonance at ~12 ppm in the spectrum of d(pApGpGpCpCpT) + Cp2MoCl2(aq) (panel D) is an unidentified byproduct. imply a phosphate-bound Mo-DNA complex that is likely located at a terminal phosphate. (M = Ti, Zr, Hf, Nb, V) interactions with salmon testis DNA were studied with inductively coupled plasma (ICP) spectroscopy [93]. Cp2MCl2-DNA adducts of unknown binding mode were detected subsequent to DNA precipitation for Ti, Zr, Hf, and Nb, whereas no evidence for a similar DNA complex was observed for Cp2VCl2. These results suggest a weak correlation between antitumor
< previous page
page_74
next page >
< previous page
page_75
next page > Page 75
activity and metal-DNA binding as assayed by precipitation. In accordance with the hydrolytic results for Cp2TiCl2, it was found that at pH 5.3 and 7.0, Cp2Ti-DNA and CpTi-DNA adducts are formed, respectively. However, whether these particular adducts are connected with the inhibition of DNA synthesis remains unknown. The reasons behind the differences in Cp2Mo2+ and cis-(NH3)2-Pt2+-DNA coordination modes lie in part in the steric interactions involving the respective nonchloride ligands. One complex is square planar whereas the other is pseudotetrahedral; with respect to binding to the DNA double helix, one is flat with small amine ligands that are directed away from the binding site whereas the other possesses transversely oriented ligands of considerable steric bulk. Molecular graphics studies [60,91] show that the planar platinum complex is readily accommodated within the DNA major groove in an orientation that facilitates binding to adjacent nucleobases. However, the same is not true for pseudotetrahedral metallocene complexes. The sterically more demanding Cp ligands of metallocenes appear to preclude this mode of binding to DNA; metallocenes are simply too large to fit into the major groove (Fig. 10). In contrast, minimal nonbonded contacts are encountered when the metallocene is located at the 5' terminus, binding through the terminal phosphate and a nucleobase. 4 Metallocene Interactions with DNA-Processing Enzymes 4.1 Endonucleases, Polymerases, Kinases, and Topoisomerases In light of the aforementioned results on Cp2MCl2(aq) (M = V and Mo) interaction with oligonucleotides, it seems unlikely that the metallocene dichlorides disrupt cellular function by DNA binding in a mode analogous to that of cisplatin. In addition, incubation of plasmid DNA with Cp2MoCl2(aq) or Cp2VCl2(aq) has no effect on the DNA electrophoretic mobility and does not significantly affect endonuclease, ligase, or polymerase activity on these plasmids [87]. In contrast, cisplatin binding to DNA perturbs the activity of DNA restriction endonucleases and polymerases by blocking certain sites of attack [90,94].
< previous page
page_75
next page >
< previous page
page_76
next page > Page 76
Fig. 10. Comparison of Cp2Mo2+ and cis-(NH3)2Pt2+ bonding to two purine nucleobases via the N7 site. The steric interactions imposed by Cp ligands (A) force bulky ligands such as purines to assume a coplanar orientation in the equatorial girdle (B). Such bonding would require a substantial conformational change in the DNA backbone. (C) In contrast, two mutually tipped (noncoplanar) purine ligands can be readily accommodated in the plane(s) perpendicular to the molecular square plane of cis-(NH3)2Pt2+. Therefore, alternative biomolecular targets for Cp2MCl2 have been sought. One identified target is protein kinase C (PKC), an enzyme that plays a role in regulating cellular proliferation [95,96]. There is considerable evidence that PKC activation leads to tumor promotion [9799], while a number of antitumor agents are believed to be active, at least in part, via PKC inhibition [100102]. Thus, disruptions of normal PKC functions impair signal transduction necessary for replication, transcription, and mitosis, which are functions suppressed by the metallocene dihalides. It was shown (Fig. 11) that the ability of Cp2VCl2(aq) and Cp2MoCl2(aq) to inhibit PKC activity [87] is comparable to that of the PKC inhibitory antitumor agents Tamoxifen and Suramin [103]. Control experiments show that the products of Cp2MCl2 hydrolysis (i.e., HCl) and NaCl do not inhibit PKC and that cofactors such as Ca2+, ATP, or αphosphatidyl-L-serine do not significantly decrease the inhibition of PKC activity by Cp2VCl2(aq) or Cp2MoCl2(aq). However, air oxidation of the metallocenes prior to the experiment or the addition of effective chelating ligands such as phosphonoformic acid incapacitates Cp2MoCl2(aq) and Cp2VCl2(aq) as PKC inhibitors [87]. Further implications for metallocene therapy in light of these PKC inhibitory effects lie in the connection of PKC with drug resistance [104] and metastasis [105]. Higher PKC activity is found in drug-resistant tumor cell lines, and activators of PKC can reduce the chemosensi-
< previous page
page_76
next page >
< previous page
page_77
next page > Page 77
Fig. 11. The inhibition of protein kinase C (PKC) by MDC (Cp2MoCl2), VDC (Cp2VCl2), and tamoxifen. PKC activity is measured by the rates of 32P incorporation from [γ-32P]-ATP into substrate. The depression of PKC activity, or the extent of inhibition, is expressed as the ratio: (PKC activity in the presence of the specific inhibitor concentration)/(activity of uninhibited PKC). tivities of a variety of mammalian tumor cells, while PKC inhibitors can effect the reverse [104]. Since the emergence of drug-resistant tumor cells [104] limits the efficacy of most currently available antitumor agents, the potential of metallocenes to lower drug resistance through PKC inhibition, when used in combination therapy, could be of great value. In addition, increased PKC activity has been linked to the metastasis of tumor cells while PKC inhibition reverses the effect [106]. It was also shown [87] that Cp2MoCl2(aq) and Cp2VCl2(aq) inhibit enzymes involved in the topological processing of DNA. One such enzyme is DNA topoisomerase II, which controls the DNA conformation and superstructure necessary in the final stage of replication by catenating the DNA polymer [107]. DNA topoisomerase II has been identified as the primary cellular target for a number of chemotherapeutic
< previous page
page_77
next page >
< previous page
page_78
next page > Page 78
drugs [108111] that inhibit cellular DNA synthesis by arresting the cells preferentially at the G2 or at the G1/G2 phases of the cell cycle [112,113], a phenomenon that was observed in the in vitro studies with metallocenes [19]. Electrophoretic experiments [87] reveal a concentration-dependent inhibition of bacterial DNA topoisomerase II (winding) activity on relaxed DNA plasmid treated with Cp2MCl2(aq) (M = Mo, V). Whether this inhibition also takes place with the eukaryotic enzymes remains to be demonstrated. 5 Conclusions and Outlook It is clear from the above discussion that metallocene dichloride complexes interact with DNA in a very different manner than does cisplatin, despite apparent similarities in molecular structure and in vivo/in vitro biological activity. The aqueous interaction of the metallocene dichlorides with polynucleotides is characterized by labile Cp2VCl2(aq)-phosphate interaction and less labile C2MoCl2(aq)-O(phosphate) + N(nucleobase) coordination. The metallocenes do not affect cellular function in the same manner as cisplatin in terms of blocking DNA-processing enzymes such as polymerase and restriction endonucleases. Therefore it appears that the carcinostatic activity of the metallocene complexes does not predominantly originate from DNA binding but is likely due to enzymatic inhibition that blocks cellular mitotic activity. Two possible target enzymes that have been identified are protein kinase C and topoisomerase II, which are inhibited by Cp2MoCl2(aq) and Cp2VCl2(aq) as effectively as other antitumor agents that act on these enzymes. Indeed, the present metallocenes may be the first examples of organometallic PKC and topoisomerase II inhibitors, the effects of which are consistent with selective inhibition of DNA synthesis and mitosis, while interacting with DNA in a manner different from that of cisplatin. Furthermore, the labile DNAmetallocene interactions observed may avoid unfavorable cisplatin toxicological characteristics and suggest that these organometallic complexes may be less mutagenic than cisplatin. In light of this hypothesis and the novelty of the metallocene coordination chemistry, it is apparent that these complexes are new and relatively unexplored carcinostatic agents having potential as chemotherapeutic agents.
< previous page
page_78
next page >
< previous page
page_79
next page > Page 79
Acknowledgments T. J. M. gratefully acknowledges support from the National Science Foundation (NSF CHE 9104112) for this work. A. H. L. thanks the National Institutes of Health, National Cancer Institute for stipend support (NIH Postdoctoral Fellowship, grant no. 5 F32 CA8636). Abbreviations
ATP
adenosine 5'-triphosphate
bipy
2,2'-bipyridyl
Cp
cyclopentadienyl
dAMP
2'-deoxyadenosine 5'-monophosphate
dCMP
2'-deoxycytidine 5'-monophosphate
dGMP
2'-deoxyguanosine 5'-monophosphate
dTMP
deoxythymidine 5'-monophosphate
EAT
Ehrlich ascites tumor
EELS
electron energy loss spectroscopy
FAB/MS
fast atom bombardment mass spectroscopy
HEp-2
human epidermoid carcinoma
HSAB
hard-soft-acid-base
ICP
inductively coupled plasma
MDC
molybdenocene dichloride
NMR
nuclear magnetic resonance
phen
o-phenanthroline
PKC
protein kinase C
VDC
vanadocene dichloride
References 1. H. Köpf and P. Köpf-Maier, Angew. Chem. Int. Ed., 18, 477 (1979). 2. M. E. Cooley, L. E. Davis, M. DeStefano, and J. Abrahm, Cancer Nurs., 17, 173 (1994). 3. P. J. Bednarski, R. Gust, T. Spruss, N. Knebel, A. Otto, M. Farbel, and R. Koop, Cancer Treat. Rev., 17, 221 (1990). 4. G. Chu, J. Biol. Chem., 269, 787 (1994).
< previous page
page_79
next page >
< previous page
page_80
next page > Page 80
5. M. Castagnola, B. Floris, G. Illuminati, and G. Ortaggi, J. Organomet. Chem., 60, C17 (1973). 6. E. W. Neuse, R. Loonat, and J. C. A. Boeyens, Trans. Met. Chem., 9, 12 (1984). 7. P. Köpf-Maier, Z. Naturforsch [C], 40, 843 (1985). 8. P. Köpf-Maier, T. Klapötke, Arzneim-Forsch., 39, 369 (1989). 9. F. P. Purchnik and S. A. Duraj in Organometallic Chemistry of the Transition Elements (J. P. Tackler, ed.), Plenum Press, New York, 1990, pp. 509574. 10. J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc., 98, 1729 (1976). 11. R. H. Crabtree in The Organometallic Chemistry of Transition Metals, John Wiley and Sons, New York, 1988, p. 108. 12. G. Wilkinson, P. L. Pauson, J. M. Birmingham, and F. A. Cotton, J. Am. Chem. Soc., 75, 1011 (1953). 13. M. Dub in Organometallic Compounds, Vol. I, Springer-Verlag, Berlin, 1966, p. 105. 14. C. R. Lucas and M. L. H. Green, J. Chem. Soc. Chem. Commun., 1005 (1972). 15. M. Bottrill, P. D. Gavens, J. W. Kelland, and J. McMeeking in Comprehensive Organometallic Chemistry, Vol. 3 (G. Wilkinson, F. Gordon, and A. Stone, eds.), Pergamon Press, New York, 1982, p. 331. 16. T. M. Klapötke, H. Köpf, I. C. Tornieporth-Oetting, P. S. White, and P. Köpf-Maier, Angew. Chem. Int. Ed., 33, 1518 (1994). 17. P. Köpf-Maier, J. Struct. Biol., 105, 35 (1990). 18. P. Köpf-Maier and D. Krahl, Naturwiss, 68, 273 (1981). 19. H. Köpf and P. Köpf-Maier in Platinum, Gold and Other Transition Metal Chemotherapeutic Agents (S. J. Lippard, ed.), ACS Symp. Series Vol. 209, 1983, p. 315. 20. P. Köpf-Maier and H. Köpf, Anticancer Res., 6, 227 (1986). 21. P. Köpf-Maier, F. Preiss, T. Marx, T. Klapötke, and H. Köpf, Anticancer Res., 6, 33 (1986). 22. P. Köpf-Maier and H. Köpf, Arzneim-Forsch., 37, 532 (1987). 23. P. Köpf-Maier, W. Wagner, and H. Köpf, Cancer Chemother. Pharmacol., 5, 237 (1981).
< previous page
page_80
next page >
< previous page
page_81
next page > Page 81
24. P. Köpf-Maier, Cancer Chemother. Pharmacol., 23, 225 (1989). 25. P. Köpf-Maier, Dev. Oncol., 54, 601 (1988). 26. P. Köpf-Maier, W. Wagner, B. Hesse, and H. Köpf, Eur. J. Cancer, 17, 665 (1981). 27. P. Köpf-Maier, P. Moorman, and H. Köpf, Eur. J. Cancer, 21, 853 (1985). 28. P. Köpf-Maier and A. Hesse, J. Cancer Res. Clin. Oncol., 108, 254 (1984). 29. P. Köpf-Maier, W. Wagner, and L. Liss, J. Cancer Clin. Oncol., 102, 21 (1981). 30. P. Köpf-Maier, J. Cancer Clin. Oncol., 114, 250 (1988). 31. P. Köpf-Maier, J. Cancer Clin. Oncol., 113, 342 (1987). 32. P. Köpf-Maier and H. Köpf, Struct. Bond., 70, 103 (1988). 33. P. Köpf-Maier in Metal Complexes in Cancer Chemotherapy (B. K. Keppler, ed.), VCH, Weinheim, 1993, p. 261. 34. J. H. Toney, L. N. Rao, M. S. Murthy, and T J. Marks, Breast Cancer Res. Treat., 6, 185 (1985). 35. J. H. Toney, M. S. Murthy, and T. J. Marks, Chem.-Biol. Interact., 56, 45 (1985). 36. M. S. Murthy, L. N. Rao, L. Y. Kuo, J. H. Toney, and T. J. Marks, Inorg. Chim. Acta, 152, 117 (1988). 37. P. Köpf and P. Funke-Kaiser, Toxicology, 38, 91 (1986). 38. P. Köpf-Maier and S. Gerlach, J. Cancer Res. Clin. Oncol., 111, 243 (1986). 39. P. Köpf-Maier and S. Gerlach, Anticancer Res., 6, 235 (1986). 40. P. Köpf-Maier and T. Klapötke, Arzneim.-Forsch., 39, 488 (1989). 41. P. Köpf-Maier, E. Neuse, T. Klapötke, and H. Köpf, Cancer Chemother. Pharmacol., 24, 23 (1989). 42. P. Köpf-Maier and T. Klapötke, Cancer Chemother. Pharmacol., 29, 361 (1992). 43. P. Köpf-Maier and T. Klapötke, J. Cancer Res. Clin. Oncol., 118, 216 (1992). 44. P. Köpf-Maier and H. Köpf, J. Organomet. Chem., 342, 267 (1988). 45. P. Köpf-Maier, B. Hesse, R. Voigtlander, and H. Köpf, J. Cancer Res. Clin. Oncol., 97, 31 (1980).
< previous page
page_81
next page >
< previous page
page_82
next page > Page 82
46. P. Köpf-Maier, T. Klapötke, and H. Köpf, Inorg. Chim. Acta, 153, 119 (1988). 47. A. Moustatih, M. M. Fiallo, and A. Garnier-Suillerot, J. Med. Chem., 32, 336 (1989). 48. P. Köpf-Maier, W. Kahl, N. Klouras, G. Hermann, and H. Köpf, Eur. J. Med. Chem., 16, 275 (1981). 49. S. L. Bruhn, J. H. Toney, and S. J. Lippard, Prog. Inorg. Chem., 38, 477 (1990). 50. W. I. Sundquist and S. J. Lippard, Coord. Chem. Rev., 100, 293 (1990). 51. K. M. Comess and S. J. Lippard, in Molecular Aspects of Anticancer Drug-DNA Interaction (S. Neidle and M. Waring, eds.), CRC Press, Boca Raton, 1993, p. 134. 52. S. J. Lippard and J. M. Berg in Principles of Bioinorganic Chemistry, University Science Books, Mill Valley, CA, 1994, p. 175. 53. P. Köpf-Maier, J. Cancer Res. Clin. Oncol., 103, 145 (1982). 54. P. Köpf-Maier, J. Cancer Res. Clin. Oncol., 114, 250 (1988). 55. P. Köpf-Maier and D. Krahl, Chem.-Biol. Interact., 44, 317 (1983). 56. P. Köpf-Maier and H. Köpf, Naturwiss., 67, 415 (1980). 57. P. Köpf-Maier, W. Wagner, and H. Köpf, Naturwiss., 68, 272 (1981). 58. H. Köpf and P. Köpf-Maier, Nachr. Chem. Tech. Lab., 29, 272 (1981). 59. J. H. Toney and T. J. Marks, J. Am. Chem. Soc., 107, 947 (1985). 60. L. Y. Kuo, M. G. Kanatzidis, M. Sabat, A. L. Tipton, and T. J. Marks, J. Am. Chem. Soc., 113, 9027 (1991). 61. J. H. Murray and M. M. Harding, J. Med. Chem., 37, 1936 (1994). 62. R. G. Pearson (ed.), Hard and Soft Acids and Bases, Dowden, Hutchinson, and Ross, Stroudsburg, PA, 1973. 63. J. W. Reishus and D. S. Martin, Jr., J. Am. Chem. Soc., 83, 2457 (1961). 64. B. Lippert in Progress in Inorganic Chemistry (S. J. Lippard, ed.), John Wiley and Sons, New York, 1989, p. 1. 65. J. H. Toney, C. P. Brock, and T. J. Marks, J. Am. Chem. Soc., 108, 7263 (1986). 66. T. J. Swift in NMR of Paramagnetic Molecules (G. N. La Mar, W.
< previous page
page_82
next page >
< previous page
page_83
next page > Page 83
Harrocks, and R. H. Holm, eds.), Academic Press, New York, 1973, p. 77. 67. H. Schöllhorn, U. Thewalt, and B. Lippert, J. Am. Chem. Soc., 111, 7213 (1989). 68. M. J. Calhorda, M. A. A. F. de C. T. Carrondo, M. H. Garcia, and M. G. Hursthouse, J. Organomet. Chem., 324, 209 (1988). 69. M. J. Calhorda, M. A. A. F. de C. T. Carrondo, M. H. Garcia, and C. C. Ramao, J. Organomet. Chem., 320, 63 (1987). 70. M. A. A. F. de C. T. Carrondo, M. J. Calhorda, and M. G. Hursthouse, Acta Crystallogr., C43, 880 (1987). 71. M. A. A. F. de C. T. Carrondo and A. M. T. S. Domingos, J. Organomet. Chem., 253, 53 (1983). 72. R. A. Forder, G. D. Gale, and C. K. Prout, Acta Crystallogr., B31, 297 (1975). 73. C. K. Prout, T. S. Cameron, R. A. Forder, S. R. Critchley, B. Denton, and G. V. Rees, Acta Crystallogr., B30, 2290 (1974). 74. L. G. Marzilli, Prog. Inorg. Chem., 23, 256 (1982). 75. R. Faggiani, B. Lippert, C. J. L. Lock, and R. A. Sperzanzini, J. Am. Chem. Soc., 103, 1111 (1981). 76. K. Aoki, J. Am. Chem. Soc., 20, 335 (1981). 77. P. C. Gagnon and A. L. Beauchamp, Acta Crystallogr., B33, 1448 (1977). 78. P. DeMeester, D. M. Goodgame, A. C. Skapski and Z. Warne, Biochim. Biophys. Acta, 324, 301 (1973). 79. H. Schöllhorn, R. Beyerle-Pfnur, U. Thewalt, and B. Lippert, J. Am. Chem. Soc., 108, 3680 (1986). 80. D. W. Young, P. Tollin, and H. R. Wilson, Nature, 248, 513 (1974). 81. A. L. Beauchamp, D. Cozak, and A. Mardhy, Inorg. Chim. Acta, 92, 191 (1984). 82. D. Cozak, A. Mardhy, M. J. Oliver, and A. L. Beauchamp, Inorg. Chem., 25, 2600 (1986). 83. S. S. Massoud and H. Sigel, Eur. J. Biochem., 179, 451 (1989). 84. H. Sigel, Chem. Soc. Rev., 22, 255 (1993). 85. H. Sigel, S. S. Massoud, and R. Tribolet, J. Am. Chem. Soc., 110, 6857 (1988). 86. X. Jia, G. Zon, and L. G. Marzilli, Inorg. Chem., 30, 228 (1991).
< previous page
page_83
next page >
< previous page
page_84
next page > Page 84
87. A. Liu, L. Y. Kuo, and T. J. Marks, manuscript in preparation. 88. J. P. Caradonna, S. J. Lippard, M. J. Gait, and M. Singh, J. Am. Chem. Soc., 104, 5793 (1982). 89. J. P. Caradonna and S. J. Lippard, Inorg. Chem., 27, 1454 (1988). 90. G. L. Cohen, J. A. Ledner, W. R. Bauer, H. M. Ushay, C. Caravana, and S. J. Lippard, J. Am. Chem. Soc., 102, 2487 (1980). 91. M. M. Harding, G. J. Harden, and L. D. Field, FEBS Lett., 322, 291 (1993). 92. D. G. Gorenstein, Meth. Enzymol., 211, 245 (1992). 93. M. L. McLaughlin, J. M. Cronan, Jr., T. R. Schaller, and R. D. Snelling, J. Am. Chem. Soc., 112, 8949 (1990). 94. F. Bernges and E. Holler, Biochemistry, 27, 6398 (1988). 95. Y. Nishizuka, Science, 258, 607 (1992). 96. Y. Nishizuka, Adv. Med. Chem., 1, 1 (1992). 97. A. Farago and Y. Nishizuka, FEBS Lett., 268, 350 (1990). 98. A. S. B. Kekule, U. Lauer, L. Weiss, B. Luber, and P. H. Hofschneider, Nature, 361, 742 (1993). 99. C. J. Marasco, C. Piantadosi, K. L. Meyer, S. Morris-Natschke, K. S. Ishaq, G. W. Small, and L. W. Daniel, J. Med. Chem., 33, 985 (1990). 100. D. Barak, M. Shitabata, and R. Rein, Theochem., 76, 419 (1991). 101. M. D. Minana, V. Felipo, F. Cortes, and S. Grisolia, FEBS Lett., 284, 60 (1991). 102. J. E. River and G. R. Marshall (eds.), Pept.; Chem. Struct. Biol. Proc. 11th Am. Pept. Symp., ESCOM, Leiden, 1989, p. 373. 103. C. E. Hensey, D. Boscoboinik, and A. Azzi, FEBS Lett., 258, 156 (1989). 104. N. E. Ward and C. A. O'Brian, Cancer Lett., 58, 189 (1991). 105. R. Gopalakrishna and S. H. Barsky, Proc. Natl. Acad. Sci. USA, 85, 612 (1988). 106. J.-M. Herbert and J.-P. Maffrand, Biochem. Pharmacol., 42, 163 (1991). 107. N. R. Cozzarelli and J. C. Wang (eds.), DNA Topology and Its Biological Effect, Cold Spring Harbor Laboratory Press, New York, 1990. 108. J. C. Wang, Biochim. Biophys. Acta, 909, 1 (1985).
< previous page
page_84
next page >
< previous page
page_85
next page > Page 85
109. A. L. Bodley and L. F. Liu, Biotechnology, 1315 (1988). 110. J. C. Wang, Annu. Rev. Biochem., 54, 665 (1985). 111. L. F. Liu, Annu. Rev. Biochem., 58, 351 (1989). 112. C. Rius, A. R. Zorrilla, C. Cabana, F. Mata, C. Bernabeu, and P. Aller, Mol. Pharmacol., 39, 442 (1991). 113. H. Takano, K. Kohno, M. Ono, Y. Uchida, and M. Kuwano, Cancer Res., 51, 3951 (1991).
< previous page
page_85
next page >
< previous page
page_87
next page > Page 87
4 Evidences for a Catalytic Activity of the DNA Double Helix in the Reaction between DNA, Platinum(II), and Intercalators Marc Boudvillain, Rozenn Dalbiès, and Marc Leng Centre de Biophysique Moléculaire, CNRS, Rue Charles Sadron, F-45071 Orleans Cedex 02, France
88
1. Introduction
89
2. Reactivity of cis- and trans-DDP with DNA
89
2.1. Formation of the Adducts
92
2.2. Distortion of the DNA Double Helix
3. Rearrangement of trans-{Pt(NH3)2[d(GXG)-N7-G,N7-G]} Intrastrand Crosslinks
92
3.1. Platinated Single-stranded DNA
93
3.2. Platinated Double-stranded DNA
95
3.3. Is the DNA Double Helix a Catalyst?
3.4. Intrastrand Crosslinks vs. Interstrand Crosslinks
4. Reaction between cis-DDP, DNA, and Intercalators
4.1. Catalytic Activity of the DNA Double Helix
95
96
96
98
4.2. Influence of the Salt
99
5. Conclusion
< previous page
92
page_87
next page >
< previous page
page_88
next page > Page 88 101
Acknowledgments
102
Abbreviations
102
References
1 Introduction cis-Diamminedichloroplatinum(II) (cis-DDP) is largely used for cancer chemotherapy alone or in combination with other drugs. It is active against several tumors with a remarkable efficiency against ovarian and testicular cancers. It is generally accepted that cellular DNA is the target of the drug with formation of covalent adducts. Although the adducts block DNA replication and transcription, there is not yet a clear understanding of the mechanism of action of cis-DDP. Subsequent to the formation of the damaging lesions (the first necessary step), multistep reactions involving, probably, several pathways lead to cell death (for general reviews see [14]). In vitro and in vivo, cis-DDP binds preferentially to purine residues and, after a few hours, bifunctional adducts are mainly formed. The major adducts are intrastrand crosslinks at the d(GpG) and d(ApG) sites. Minor adducts are interstrand crosslinks between two G residues on opposite strands at the d(GpC)·d(GpC) sites, intrastrand crosslinks between two G residues at the d(GpXpG) sites (X being any base) and between G and A residues at the d(GpA) sites [15]. A question still under debate concerns trans-diamminedichloroplatinum(II) (trans-DDP), the stereoisomer of cis-DDP. This compound is less mutagenic and less cytotoxic than cis-DDP and is ineffective as an antitumor drug, although it reacts with DNA. In their reaction with DNA, the two isomers present similarities and differences. The exchange of chloro groups of the two isomers is rate limiting in the initial attack of DNA. The preferred site of initial binding is the N7 atom of G residues. Among the differences, one is that stereochemical limitations preclude trans-DDP forming intrastrand crosslinks between adjacent base residues [6]. In vitro and at relatively high level of platination, the most prevalent adducts are intrastrand crosslinks between G residues
< previous page
page_88
next page >
< previous page
page_89
next page > Page 89
and between G and C residues separated in both cases by at least one residue, and interstrand crosslinks between complementary G and C residues [610]. The interstrand crosslinks are formed by the trans isomer more slowly than by the cis isomer [10]. There are contradictory results on the rate of the intrastrand crosslinking reaction [79]. It is generally accepted that the bifunctional adducts are stable under close to physiological conditions and even in the presence of NaCl. The bound platinum can be removed by some reagents having a high affinity for the ''soft" platinum(II) center such as cyanide ions and sulfur-containing nucleophiles [6]. The purpose of this chapter is to summarize some results showing the active participation of the DNA double helix itself in transformation and/or removal of the adducts under physiological conditions. Two examples will be mainly discussed. The first one concerns the DNA double-helix-promoted linkage isomerization reaction of the trans-DDP-(G1,G3) intrastrand crosslinks at the d(GpXpG)·d(CpYpC) sites (Y is a base residue). The intrastrand crosslinks are stable (there are a few exceptions) within single-stranded DNA but they rearrange into interstrand crosslinks as soon as the platinated DNA is hybridized with its complementary strand. The second example, more complex, concerns the reaction of cisDDP with DNA in the presence of intercalators (Am). It has been proposed that DNA acts as a catalyst in that it favors the reaction between Am and cis-DDP but also promotes the rearrangement of the monofunctional adduct cis[Pt(NH)3)2(dG) Am]n+ into bifunctional adducts. 2 Reactivity of cis- and trans-DDP with DNA 2.1 Formation of the Adducts Numerous results [2,6] support that formation of the adducts in the reaction between DNA and cis- or trans-DDP proceeds in a two-solvent-assisted reaction in sequence as summarized in Fig. 1. At pH 6.5 and 37°C, the pseudofirst-order rate constants for the first DNA binding step are k = (10.2 ± 0.7) × 105 sec1 (t1/2 = 1.9 ± 0.1 hr) for cisDDP and k = (9.6 ± 0.7) × 105 sec1 (t1/2 = 2 ± 0.1 hr) for trans-DDP, rates which are the same as for the rate of hydrolysis of the first chloride ion [9]. The
< previous page
page_89
next page >
< previous page
page_90
next page > Page 90
Fig. 1. Formulas of cis- (cis-DDP) and trans-diamminedichloroplatinum(II) (trans-DDP). Main steps of the reaction between cis- and trans- DDP and DNA.
rate-determining step for closure of the cis-DDP monofunctional adducts to bifunctional adducts at d(GpG) and d(ApG) sites is also the hydrolysis of the second chloride ion (t1/2 = 2.1 hr) [9,11]. Longer halflives (2.812 hr) are found for closure of monofunctional adducts to interstrand crosslinks and even longer than 20 hr for closure to intrastrand crosslinks at d(GpTpG) sites [12]. It seems likely that the slower reaction is due to an unfavorable local conformation of DNA rather than to a slower hydrolysis of the chloride ion. Results are contradictory for closure of the trans-DDP monofunctional adducts and for the nature of the bifunctional adducts. Bancroft et al. [9] find that the rate-determining step is hydrolysis of the second chloride ion in both native and denatured DNA. Eastman et al. [8] find that the rate of closure is slower in native DNA than in denatured DNA (about 50% of the adducts are monofunctional in native DNA after 24 hr of reaction). The apparent disagreement could be related to different input drug-to-nucleotide ratios. As the level of platination increases, DNA becomes more and
< previous page
page_90
next page >
< previous page
page_91
next page > Page 91
more distorted, which could affect the rate of the crosslinking reaction and the nature of the adducts. This could also reflect a nonrandom distribution of the platinum residues along the DNA molecule. These results show that DNA itself influences the reaction with the drugs. Other results underline the active participation of DNA. The reactivity of cis-DDP increases with increasing length of DNA but more than expected on the basis of a net charge effect [13]. DNA is a negatively charged surface which attracts cis-[Pt(NH3)2(H2O)Cl]+ cations. In addition to the increased local concentration, a high degree of mobility of the cations along the polymer backbone favors their binding to their target sequences [13]. The fact that the electrostatic potential is not constant along the DNA molecule might modulate this effect [14]; the high frequency of cis-DDP binding at d(GpG) sites in natural DNA excludes an equal reactivity for all the G residues [14]. In addition, without considering the non-B structures (A form, Z form, cruciform, triplex, etc.), B-DNA, the most stable form of the double helix, presents local geometrical variations with the base sequence, which in turn affects the DNA hydration layers [1517]. Moreover, DNA is not a static but a dynamic structure [18,19]. It is subjected to thermally driven fluctuations resulting in transient conformations involving distortions such as bending, twisting, base pair opening, and local strand separation. Without these fluctuations, the distances between the reactive sites would prevent the crosslinking reaction. Nevertheless, the geometrical constraints of the double helix and the relative orientation of the monofunctional adducts and their targets appear unfavorable for the formation of 1,3-intrastrand crosslinks. The crosslinking reaction is at least 10 times slower for the cis-DDP 1,3-intrastrand crosslinks at the d(GpTpG) sites (the distance between the two N7 is about 8.5 Å but the N7 atoms are well accessible in the major groove) than for the cisDDP intrastrand crosslinks at the d(ApG) sites [12]. The reaction is expected to be slower for the trans-DDP 1,3intrastrand crosslinks because of the orientation of the entering group in the final Pt(II) complex. A more severe situation concerns the trans-DDP intrastrand crosslinks between G and A residues or between G and C residues. The N1 (A) or the N3 (C) are inside the double helix and hydrogen-bonded. The formation of the trans-DDP interstrand crosslinks implies a rotation of the platinated G residues around the glycosyl bond from an anti conformation to a syn conformation [20].
< previous page
page_91
next page >
< previous page
page_92
next page > Page 92
2.2 Distortion of the DNA Double Helix The distortions induced in the double helix by the major cis-DDP adducts have been characterized by several methods [2,4,6,2125], but crystallographic structures of the platinated duplexes are not yet available. Fewer studies have been devoted to the distortions induced by the trans-DDP adducts. Recent works show that the cis- and transDDP interstrand crosslinks distort the double helix differently. The cis-DDP adducts [between the G residues in the d(GpC)·d(GpC) sites] bend the double helix toward the major groove and unwind it (about 45° and 80°, respectively); the major distortion occurs at the level of the adduct and the C residues become reactive with some chemical probes [26]. The trans adducts (between complementary G and C residues) bend the double helix toward the major groove and unwind it (about 26° and 12°, respectively); the distortion spreads over four base pairs as revealed by chemical probes [20]. 3 Rearrangement of trans-{Pt(NH3)2[d(GXG)-N7-G7,N7-G]} Intrastrand Crosslinks 3.1 Platinated Single-Stranded DNA Oligonucleotides containing a single d(GXG) sequence (X stands for A, T, or C) react with trans-DDP to yield trans{Pt(NH3)2[d(GXG)-N7-G,N7-G]} crosslinks. The adducts are stable at neutral pH and in various conditions (in 10 mM thiourea and at 37°C, the half-life is about 32 hr). There is one exception [27,28]: when the base residue adjacent to the adduct on its 5' side is a C residue, a linkage isomerization reaction occurs which transforms the (G,G)-intrastrand crosslink to the 1,4-trans-{Pt(NH3)2[d(CGXG)-N3-C,N7-G]} intrastrand crosslink. The scheme of the reaction can be written as follows:
The rate of the reaction depends on temperature (the half-lives of the 1,3-trans-{Pt(NH3)2[d(GAG)-N7-G,N7-G]} crosslink within a 12-mer are about 4 hr at 60°C and 40 hr at 37°C, respectively). At a given
< previous page
page_92
next page >
< previous page
page_93
next page > Page 93
temperature, the rate is independent of the nature and concentration of the salt (NaCl or NaClO4) in the rage 10400 mM. The equilibrium constant for the rearrangement is about 10, in favor of the 1,4-intrastrand crosslink [27,28]. In first approximation, all of the results are the same for several oligonucleotides [from d(pCGXG) to 22-mer containing a single central sequence d(CGXG)]. The chemical nature of X and of the residues flanking the d(CGXG) sequence have no major effect on the reaction [28]. The 5' C residue adjacent to the adduct plays a key role in the reaction; its replacement by a G, A, or T residue prevents the reaction. No rearrangement was observed when a C residue is adjacent to the adduct on the 3' side. There is not yet a complete explanation of the rearrangement. Considering the usual mechanism of labile ligand substitution in platinum(II) complexes, two extreme pathways can be proposed: (1) a two-step solvent-assisted process involving the opening of the bifunctional crosslink to a monofunctional adduct and subsequent closure to a bifunctional crosslink (see Fig. 1); (2) a direct nucleophilic attack of the platinum(II) center by a base residue. Mechanism (2) could be relevant in explaining the results taking into account a large negative entropy of activation of the reaction (mean value, ∆S≠ = 75 J·mol1·K1) and the noninfluence of chloride ions on the reaction rate [27,28]. Mechanism (2) requires that the platinated oligonucleotides adopt a conformation allowing the nucleophilic N atom of the base residue to attack the platinum(II) center in a well-defined direction. Nuclear magnetic resonance (NMR) experiments are in progress to test this proposal. 3.2 Platinated Double-Stranded DNA The formation of a double helix by pairing a single-stranded DNA containing a single trans-{Pt(NH3)2[d(GXG)-N7G,N7-G]} crosslink to its complementary strand promotes a novel linkage isomerization reaction which transforms the 1,3-intrastrand crosslink to an interstrand crosslink [29] as schematically represented in Fig. 2. In fact, the nature of the interstrand crosslinks depends on Y. If Y is a pyrimidine base (Py), the interstrand crosslinks are between the 5' G (upper strand) and its complementary C. If Y is a purine base (Pu), there are two interstrand crosslinks, the major one between the 5' G
< previous page
page_93
next page >
< previous page
page_94
next page > Page 94
Fig. 2. Rearrangement of the trans-DDP 1,3-intrastrand crosslink into an interstrand crosslink promoted by the pairing of the single-stranded DNA containing a single 1,3-trans-{Pt(NH3)2[d(GXG)-N7-G,N7-G]} intrastrand crosslink with its complementary strand. and its complementary C and the minor one (about 20%) between the 5' G and Y, but this has to be confirmed. The rate of the isomerization reaction depends on the nature of Y (Pu or Py) more than on the nature of X. Whatever X (Pu or Py), the half-lives of the 1,3-intrastrand crosslinks are in the range 16 hr, Y being a Py, and in the range 424 hr, Y being a Pu. They depend slightly on the nature of the base pairs adjacent to the adducts. Although they are not directly involved in the reaction, the adjacent base pairs play an important role. The replacement of one of these base pairs by a noncomplementary base pair (mismatch) prevents the rearrangement, which suggests that the local conformation of the double helix influences the reaction. Finally, all of these results do not depend on the nature and the concentration of the salt (NaCl or NaClO4) as long as the duplex is formed [29]. The rate of the rearrangement depends on temperature, which allows one to deduce the activation parameters (mean values: ∆H = 82 kJ·mol1, ∆S = 67 J·mol1·K1). The values of these parameters and the nontrapping effect of chloride ions support that the linkage isomerization reaction results from a direct nucleophilic attack of the C residue complementary to the platinated 5' G residue on the platinum residue (mechanism (2) in Sec. 3.1). Such a mechanism requires that the
< previous page
page_94
next page >
< previous page
page_95
next page > Page 95
C residue comes close to the platinum residue along a line perpendicular to the platinum square plane. The distortions induced in the double helix by the 1,3-intrastrand crosslink are not yet well characterized. An NMR study of the platinated duplex d(CCTCG*AG*TCTCC)·d(GGAGACTCGAGG) has been done but the sample was heated [30]; there are some doubts as to the purity of the sample. In another report [31], chemical probes suggest a local denaturation over four base pairs including the three base pairs at the level of the adduct and the 5' base pair adjacent to the adduct. This local denaturation could allow the C residue to be in the position required for the reaction. Nevertheless it seems likely that despite the local denaturation some structural organization is necessary since the reaction is slower in RNA-DNA hybrids and does not occur in duplexes containing a mismatch on the 5' or 3' side of the adduct [29]. 3.3 Is the DNA Double Helix a Catalyst? The results in Sec. 3.2 show that the 1,3-intrastrand crosslinks formed in the reaction between trans-DDP and singlestranded DNA containing a single d(GXG) sequence rearrange into interstrand crosslinks as soon as the platinated DNA is hybridized with its complementary strand. It is interesting to compare the reaction scheme proposed in Fig. 2 and the classical minimal reaction scheme of a Menten-Michaelis reaction:
In this scheme, the enzyme (E) binds the substrate (S), lowers the energy of the transition state, and promotes the catalytic event. The two reaction schemes are formally similar if one identifies the substrate with the adduct within the single-stranded DNA and the enzyme with the double helix, respectively. 3.4 Intrastrand Crosslinks vs. Interstrand Crosslinks A question was to know whether in the reaction between double-stranded DNA and trans-DDP, the intrastrand crosslinks are first
< previous page
page_95
next page >
< previous page
page_96
next page > Page 96
formed and subsequently rearrange into interstrand crosslinks. Several duplexes having d(GXG)·d(CYC) as central sequence and containing a single monofunctional adduct trans-{Pt(NH3)2(dG)Cl}+ were incubated at 37°C. Analysis of the results shows that the closure of the monofunctional adducts to bifunctional crosslinks yields mainly interstrand crosslinks and the reaction is slow [29,32]. Thus, the formation of the interstrand crosslinks does not proceed through the formation of 1,3-intrastrand crosslinks in double-stranded DNA after reaction with trans-DDP. It has been further shown by means of an enzymatic probe that (G1,G3)-, (G1,G4)-, and (C1,G4)-intrastrand crosslinks are not the major adducts in DNA after reaction with trans-DDP (level of platination, 0.005 adduct per nucleotide residue, after incubation of the drug with DNA during 24 hr at 37°C). Under these conditions, the major adducts are the monofunctional adducts (about 80% of the bound platinum) and the interstrand crosslinks (about 15%) [32]. The apparent discrepancy between the results [79] on the nature of the bifunctional adducts and on their rate of formation in trans-DDP-modified DNA could be due to the different level of platination. 4 Reaction between cis-DDP, DNA, and Intercalators 4.1 Catalytic Activity of the DNA Double Helix As already pointed out, the reaction between DNA and cis-DDP yields mainly intrastrand crosslinks at the d(GpG) and d(ApG) sites. When the reaction of platination is done in the presence of intercalators (Am) such as proflavine, ethidium bromide, N-methyl-2,7-diazapyrenium (compounds which form reversible complexes with DNA), the major adduct is a new adduct in which the platinum residue is linked to a G residue and to Am [reaction(2)]:
If the double-stranded DNA is replaced by single-stranded DNA, the new adduct is not formed. Under the same experimental conditions but without DNA, cis-DDP reacts poorly with Am. Thus, double-stranded DNA promotes the binding of cis-DDP to Am by acting as a matrix
< previous page
page_96
next page >
< previous page
page_97
next page > Page 97
which enables a favorable orientation of the reactants. The importance of the relative orientation of the reactants at the binding site is also supported by the fact that the new type of adduct is not formed if cis-DDP is replaced by transDDP or if proflavine is replaced by acridine or 9-aminoacridine [3337]. The new adduct is relatively unstable in that as a function of time Am is slowly released in solution and the amount of bifunctional adducts increases. On the other hand, the new adduct is stable if the modified DNA is digested by DNase I. A systematic study of the stability of the adduct has been done with N-methyl-2,7-diazapyrenium (MDAP) as Am. The complex cis-[Pt(NH3)2(N7-MDAP)Cl]2+, prepared in organic solvent, reacts with G residues in single-stranded DNA and yields cis-[Pt(NH3)2(N7-MDAP)(N7-dG)]3+. The adduct is stable under various experimental conditions but becomes unstable as soon as the platinated DNA is hybridized with its complementary strand. Two independent reactions occur:
Reaction (3) corresponds to the cleavage of the Pt-G bond and cis-[Pt(NH3)2(N7-MDAP)(H2O)]3+ (Pt-MDAP) is released in solution. Reaction (4) corresponds to the cleavage of the Pt-MDAP bond and MDAP is released in solution. The yields of the two reactions depend on several parameters [39]. Under some experimental conditions, the major reaction is (3) and one can summarize reactions (2) and (3):
According to reaction (5), the DNA double helix promotes the binding of cis-DDP and MDAP. Subsequently, a local conformation change in the double helix initiates the release of Pt-MDAP and the recovery of DNA. It has been proposed that the DNA double helix behaves as a catalyst in this reaction [38]. The local conformational change deserves more study. There is evidence that in the first step of reaction (5), MDAP is intercalated between the base pairs and then after binding to cis-DDP, MDAP is located in the groove of the double helix. Another view of the possible catalytic activity of the double helix is in the comparison between reaction (3) and the reaction describing the
< previous page
page_97
next page >
< previous page
page_98
next page > Page 98
duplex-promoted rearrangement of trans-DDP 1,3-intrastrand crosslinks into interstrand crosslinks. The monofunctional cis-[Pt(NH3)2-(N7-MDAP)(N7-dG)]3+ adducts are stable within single-stranded DNA but become unstable in the duplex, which promotes the release of Pt-MDAP. Both the schemes are formally similar to the minimal scheme of a Menten-Michaelis reaction [reaction (2), Sec. 3.3] and are in favor of a catalytic activity of the DNA double helix. A last comment about reaction (3) is the release of Pt-MDAP in solution. This Pt-MDAP complex can further react with any of the G residues in the duplex, which results in a monofunctional cis-[Pt(NH3)2-(N7-MDAP)(N7-dG)]3+ adduct. Thus, as a function of time, the amount of monofunctional cis-[Pt(NH3)2(N7-MDAP)(N7-dG)3+ adducts is constant but their location along the DNA molecule differs. However, the cycle, release of Pt-MDAP and reformation of the monofunctional adducts, stops because of reaction (4). Reaction (4) (cleavage of the bond between MDAP and the platinum residue) generates a monofunctional cis[Pt(NH3)2(dG)H2O)]2+ adduct which can further react with the neighboring base residues. The consequence is that the adducts are the same as those formed in the reaction between DNA and cis-DDP. A scheme of the results relative to reactions (3) and (4) is presented in Fig. 3. A systematic study of the binding of cis-[Pt(NH3)2(Am)Cl]n+ in which Am is a heterocyclic amine (intercalator or not) and of the stability of the adducts as a function of the conformation of DNA has shown that the adducts are stable within single-stranded helices and become unstable when the duplexes are formed. The yields of reactions (3) and (4) depend on the chemical nature of Am and on the conformation of the duplexes [39]. 4.2 Influence of the Salt All of the experiments described in Sec. 4.1 have also been done in high salt concentration (1 M NaCl or NaClO4). In 1 M NaClO4, the adducts are stable within single- or double-stranded DNA. In 1 M NaCl and even after 24 hr of incubation at 37°C, the adducts are stable within single-stranded DNA. Within double-stranded DNA, the adducts are unstable and the major reaction is the cleavage of the Pt-Am bond. This cleavage generates a monofunctional adduct which can further react with the
< previous page
page_98
next page >
< previous page
page_99
next page > Page 99
Fig. 3. Scheme of the reactions promoted by the duplex formed by hybridization between a single-stranded DNA containing a single cis-[Pt(NH3)2(N7-MDAP)(N7-dG)]3+ adduct and its complementary strand. neighboring bases after removal of NaCl. This reaction allows study of the influence of the flanking base pairs on the closure of a single monofunctional adduct to a bifunctional adduct [12]. 5 Conclusion DNA acts as an active partner in the reaction with cis- or trans-DDP in that it favors or hinders the formation of the adducts and their transfor-
< previous page
page_99
next page >
< previous page
page_100
next page > Page 100
mation. Whether the double helix can be considered as a catalyst is still a matter of discussion. It is safe to state that the double helix acts as a promoter and it seems important to devote further studies on these promoted reactions because of their fundamental interest in the chemistry of platinum(II) complexes as well as because of their potential applications. Recently, it was shown that the platinum-triamine complexes cis-[Pt(NH3)2(Am)Cl]+, in which Am is an amine ligand derived from pyrimidine, purine, and piperidine, are active against murine and human tumor systems [40]. This new series of platinum(II) antitumor agents, which initially form monofunctional adducts on DNA, violate the structure-activity relationship established for platinum complexes. According to the results presented in Sec. 4, one expects a two-step reaction (the cleavage of the Pt-Am bond which generates monofunctional cis-DDP adducts and then the formation of bifunctional crosslinks with adjacent bases) yielding the same adducts as those formed in the reaction between DNA and cis-DDP. It is not yet known which adducts (intrastrand and/or interstrand crosslinks) are involved in the antitumor activity of cis-DDP. It might be possible to form preferentially one kind of adduct by the right choice of Am, with Am favoring the recognition of sequences for intrastrand or interstrand crosslinks. It is important to explain the apparent disagreement in the rate of formation of trans-DDP adducts and their composition [79]. These experiments deal with samples containing different percentages of platinum residues. We find that in trans-DDP-modified DNA (0.005 platinum per nucleotide), the major adducts are the monofunctional adducts and the interstrand crosslinks, about 80% and 15%, respectively [32]. We exclude (G1,G3)- and (G1,G4)intrastrand crosslinks as the major adducts. We exclude also that (G1,G3)-intrastrand crosslinks are first formed and then transformed into interstrand crosslinks. One can argue that in vivo even the interstrand crosslinks are hardly formed. The interstrand crosslinking reaction is slow and it is likely that most of the monofunctional adducts are trapped by compounds such as glutathione [79]. It is tempting to speculate that trans-DDP is clinically ineffective because it does not form bifunctional adducts with cellular DNA. Chemical modifications of the nonleaving groups which could favor the formation of bifunctional adducts, might be interesting to make derivatives of trans-DDP clinically active, as in fact shown in two recent reports [41,42].
< previous page
page_100
next page >
< previous page
page_101
next page > Page 101
Many of the small molecules used in human therapy, as cis-DDP, are not specific, i.e., they bind to several genes. The antisense or the antigene strategy appears very promising in the design of new specific drugs. In the antisense strategy, several studies have shown that oligonucleotides complementary to a given sequence of a mRNA, can control the stability and the function of the mRNA [43,44]. However, the binding of the oligonucleotides to their target sequences is reversible; thus it is difficult to completely block biological functions in a RNase-H-independent mechanism. To make the process irreversible, chemical and photoactivable reagents have been covalently linked to the oligonucleotides [43,44]. Non-sequence-specific reactions have often been observed for chemically induced crosslinks. Photochemical activation is difficult in the in vivo experiments. An advantage of the oligonucleotides containing trans-{Pt(NH3)2[d(GXG)-N7-G,N7-G]} intrastrand crosslinks is that under physiological conditions the crosslinks can be considered as stable as long as the oligonucleotides are single-stranded. The rearrangement into interstrand crosslinks occurs only when the oligonucleotides bind to their targets. The same goal can be achieved with the oligonucleotides containing cis-[Pt(NH3)2(Am)(dG)]n+ [reaction (4)]. However, the rates of the crosslinking reactions have to be increased to make this application useful in the in vivo experiments. Recent data (unpublished) indicate that under some conditions most of the intrastrand crosslinks are transformed in less than 1 hr. As concerns the antigene strategy, it is not yet known whether the triplex formed by the binding of the platinated strand to the complementary duplex promotes the interstrand crosslinking reaction. To conclude, it is tempting to speculate that the active participation of the DNA double helix in the reaction between DNA, cis- or trans-DDP, and heterocyclic amines is an example of reactions occurring in more elaborate systems such as the complexes between DNA, metal ions, and proteins. Acknowledgments This work was supported in part by la Ligue contre le Cancer, l'Association pour la Recherche sur le Cancer, la Fondation pour la Recherche Médicale, and the EU contracts (CHRX-CT92-0016, CHRX-CT 94-0482).
< previous page
page_101
next page >
< previous page
page_102
next page > Page 102
Abbreviations
Am
heterocyclic amine
DDP
diamminedichloroplatinum(II)
MDAP N-methyl-2,7-diazapyrenium NMR
nuclear magnetic resonance
Pu
purine
Py
pyrimidine
X
stands for A (adenine), G (guanine) or T (thymine) in several DNA sequences
References 1. S. L. Bruhn, J. H. Toney, and S. J. Lippard in Progress in Inorganic Chemistry: Bioinorganic Chemistry, Vol. 38 (S. J. Lippard, ed.), John Wiley and Sons, New York, 1990, p. 477 ff. 2. J. Reedijk, Inorg. Chim. Acta, 198, 873 (1992). 3. A. Eastman, Pharmacol. Ther., 34, 155 (1987). 4. M. Sip and M. Leng in Nucleic Acids and Molecular Biology, Vol 7 (F. Eckstein and D. M. J. Lilley, eds.), Springer-Verlag, Berlin, 1993, p. 1 ff. 5. A. M. Fichtinger-Shepman, J. L. and van de Veer, P. H. Lohman, and J. Reedijk, Biochemistry, 24, 707 (1985). 6. C. A. Lepre and S. J. Lippard in Nucleic Acids and Molecular Biology, Vol. 4 (F. Eckstein and D. M. J. Lilley, eds.), Springer-Verlag, Berlin, 1990, p. 9 ff. 7. A. Eastman and M. A. Barry, Biochemistry, 26, 3303 (1987). 8. A. Eastman, M. M. Jennerwein, and D. L. Nagel, Chem. Biol. Interact., 67, 71 (1988). 9. D. P. Bancroft, C. A. Lepre, and S. J. Lippard, J. Am. Chem. Soc., 112, 6860 (1990). 10. V. Brabec and M. Leng, Proc. Natl. Acad. Sci. USA, 90, 5345 (1993). 11. J. M. Malinge and M. Leng, Nucl. Acids Res., 16, 7663 (1988).
< previous page
page_102
next page >
< previous page
page_103
next page > Page 103
12. D. Payet, F. Gaucheron, M. Sip, and M. Leng, Nucl. Acids Res., 21, 5846 (1993). 13. S. K. C. Elmroth and S. J. Lippard, J. Am. Chem. Soc., 116, 3633 (1994). 14. A. Pullman and B. Pullman, Quart. Rev. Biophys., 14, 289 (1981). 15. P. J. Hagerman, Annu. Rev. Biochem., 59, 755 (1990). 16. O. Kennard and W. N. Hunter, Quart. Rev. Biophys., 22, 327 (1989). 17. A. A. Travers, Annu. Rev. Biochem., 58, 427 (1989). 18. J. Ramstein and R. Lavery, J. Biomol. Struct. Dynam., 7, 915 (1990). 19. J. L. Leroy, E. Charretier, M. Kochoyan, and M. Gueron, Biochemistry, 27, 8894 (1988). 20. V. Brabec, M. Sip and M. Leng, Biochemistry, 32, 11676 (1993). 21. V. Brabec, V. Kleinwächter, J. L. Butour and N. P. Johnson, Biophys. Chem., 35, 129 (1990). 22. J. Kozelka and J. C. Chottard, Biophys. Chem., 35, 165 (1990). 23. S. F. Bellon and S. J. Lippard, Biophys. Chem., 35, 179 (1990). 24. F. Herman, J. Kozelka, V. Stoven, E. Guittet, J. P. Girault, T. Huynh-Dinh, J. Igolen, J. Y. Lallemand, and J. C. Chottard, Eur. J. Biochem., 194, 119 (1990). 25. J. Kozelka, M-H. Fouchet, and J. C. Chottard, Eur. J. Biochem., 205, 895 (1992). 26. J. M. Malinge, C. Perez, and M. Leng, Nucl. Acids Res., 22, 3834 (1994). 27. K. M. Comess, C. E. Costello, and S. J. Lippard, Biochemistry, 29, 2102 (1990). 28. R. Dalbiès, M. Boudvillain, and M. Leng, Nucl. Acids Res., 23, 949 (1995). 29. R. Dalbiès, D. Payet, and M. Leng, Proc. Natl. Acad. Sci. USA, 91, 8147 (1994). 30. C. A. Lepre, L. Chassot, C. E. Castello, and S. J. Lippard, Biochemistry, 29, 811 (1990). 31. M. F. Anin and M. Leng, Nucl. Acids Res., 18, 4395 (1990). 32. M. Boudvillain, R. Dalbiès, and M. Leng Nucl. Acids Res., 23, 2381 (1995).
< previous page
page_103
next page >
< previous page
page_104
next page > Page 104
33. J. M. Malinge and M. Leng, Proc. Natl. Acad. Sci. USA, 86, 6317 (1986). 34. J. M. Malinge, A. Schwartz, and M. Leng, Nucl. Acids Res., 15, 1779 (1987). 35. J. M. Malinge, M. Sip, A. J. Blacker, J. M. Lehn, and M. Leng, Nucl. Acids Res., 18, 3887 (1990). 36. W. I. Sundquist, D. P. Bancroft, L. Chassot, and S. J. Lippard, J. Am. Chem. Soc., 110, 8559 (1988). 37. T. Ren, D. P. Bancroft, W. I. Sundquist, A. Masschelein, M. V. Keck, and S. J. Lippard, J. Am. Chem. Soc., 115, 11341 (1993). 38. F. Gaucheron, J. M. Malinge, A. J. Blacker, J. M. Lehn, and M. Leng, Proc. Natl. Acad. Sci. USA, 88, 3516 (1991). 39. D. Payet and M. Leng, in Structural Biology: the State of the Art, Vol. 2 (R. H. Sarma and M. H. Sarma, eds.), Adenine, Guilderland, NY, p. 325 ff. 40. L. S. Hollis, A. R. Amundsen, and E. W. Stern, J. Med. Chem., 32, 128 (1989). 41. N. Farrell, L. R. Kelland, J. D. Roberts, and M. Van Beusichen, Cancer Res., 52, 5065 (1992). 42. M. Coluccia, A. Nassi, F. Loseto, A. Boccarelli, M. A. Maiggio, D. Giordano, F. P. Intimi, P. Caputo, and G. Natile, J. Med. Chem., 36, 510 (1993). 43. C. Hélène and J. J. Toulmé, Biochim. Biophys. Acta, 1049, 99 (1990). 44. W. Marshall and M. H. Caruthers, Science, 259, 1564 (1993).
< previous page
page_104
next page >
< previous page
page_105
next page > Page 105
5 Trans-Diammineplatinum(II): What Makes It Different from cis-DDP? Coordination Chemistry of a Neglected Relative of Cisplatin and Its Interaction with Nucleic Acids Bernhard Lippert Fachbereich Chemie, Universität Dortmund, Otto-Hahn-Strasse 6, D-44227 Dortmund, Germany
106
1. Introduction
107
2. Basic Properties of trans-a2PtCl2
107
2.1. Synthesis of trans-(NH3)2PtCl2
2.2. Properties of trans-(NH3)2PtCl2 and Differentiation from Its cis Isomer
2.3. trans-(NH3)2PtCl2 Analogs; Complex Isomerization
109
110
2.4. Solvolysis of trans-(NH3)2PtCl2
110
2.4.1. Mono- and Diaqua Species
111
2.4.2. Acid-Base Equilibria
112
2.4.3. Selected Examples
113
3. Biological Effects
3.1. Toxicity, Antitumor Activity, Mutagenicity of trans-DDP
113
113
3.2. Novel Active trans Compounds
114
4. Reactions with Nucleic Acids
< previous page
108
page_105
next page >
< previous page
page_106
next page > Page 106
4.1. Kinetics of DNA Adduct Formation
4.2. Spectrum of DNA Adducts of trans-DDP
4.3. Macroscopic Effects of Transplatin Binding
4.4. Staining of tRNAs
5. Reactions with Defined Oligonucleotides
5.1. Single-Stranded Oligonucleotides
5.2. 1,3-Intrastrand Crosslinking in Double-Stranded Oligos
5.3. Interstrand G,C Crosslinking
5.4. Linkage Isomerization Reactions
6. Model Studies
6.1. Mono(nucleobase) Complexes
6.2. Bis(nucleobase) Complexes and Derivatives
6.2.1. Trans-a2PtL2 Compounds
6.2.2. Heteronuclear Derivatives of trans-[a2PtL2]n+
6.3. Mixed Nucleobase Complexes
6.3.1. Metal-Modified Base Pairs
6.3.2. Dimetalated Triples and Cyclic Quartets
6.4. Combining cis- and trans-DDP
114
114
115
116
116
116
116
118
118
119
120
122
122
123
125
125
126
128
129
6.5. Trans-a2Pt(IV) Nucleobase Complexes
6.6. Ternary Nucleobase/Amino Acid Complexes
6.7. Other trans-a2Pt(II) Nucleobase Complexes
129
129
130
6.8. Trans-a2Pd(II) Nucleobase Complexes
130
7. Summary
132
Abbreviations
133
References
1 Introduction The high specificity of reactions between biomolecules depends on the proper chiralities of the partners. Thus the optical isomers of a certain molecule may have completely different effects, one being reactive, the other one being totally unreactive or a competitive inhibitor or even a
< previous page
page_106
next page >
< previous page
page_107
next page > Page 107
toxic agent. In many cases these differences can be explained on the basis of the ''key-and-lock" principle. Chiral metal complexes, when interacting with chiral biomolecules, are no exception to it. The dissimilar effects of cis- vs. trans-(NH3)2PtCl2 in living cells represent a rare case of geometrical isomers of a metal compound causing pronounced differences in biological systems [1]. While the cis isomer exhibits remarkable activity in many tumor systems and is now a clinically important antitumor drug, the trans isomer has a much lower cytotoxic potency and for this reason is not useful as a drug. This observation holds for many analogs of these two compounds, even though there seem to be occasional exceptions. The obvious question, "What makes the two isomers so different?" cannot be answered at this stage. Despite the fact that many comparative studies have been undertaken, no really satisfactory picture has emerged as yet. With DNA being widely considered the crucial target of antitumor-active cis-a2Pt(II) compounds, research has concentrated in particular on possible differences in reactivity of the two isomers with this molecule, on differences in adducts, as well as differential repair of DNA lesions. In this chapter, the attempt is made to survey literature data both on the basic chemistry of trans-(NH3)2PtCl2 and its relatives and their reactions with DNA, oligonucleotides, and model nucleobases that may be relevant to the question posed in the title. 2 Basic Properties of trans-a2PtCl2 2.1 Synthesis of trans-(NH3)2PtCl2 There are a number of ways according to which trans-diamminedichloroplatinum(II), trans-(NH3)2PtCl2 (transplatin), can be prepared [2]. The first documented synthesis of transplatin is by J. Reiset ("Reiset's second chloride") who obtained it by heating dry [Pt(NH3)4]Cl2 at 250°C [3]. A major disadvantage of this methodextensive decomposition to Pt(0)can be overcome if the temperature is kept at 190195°C and a reduced pressure applied [4]. This modification also allows the preparation of isotopically labeled compounds, e.g., of transPt(15NH3)2Cl2. Alternatively, and probably still the most common procedure of preparation, is the action of concentrated HCl on an aqueous solution of
< previous page
page_107
next page >
< previous page
page_108
next page > Page 108
[Pt(NH3)4]Cl2, originally proposed by Peyrone [5] and later modified by Kauffman and Cowan [6]. It is still the method of choice to prepare transplatin analogs. 2.2 Properties of trans-(NH3)2PtCl2 and Differentiation from Its cis Isomer Basic-physico properties of trans-(NH3)2PtCl2, as determined prior to 1957 [2] and 1973 [7], respectively, have been compiled. They include, among others, solubility data, electronic spectra, infrared (IR) spectra, and a qualitative description of the bonding properties. More recent data have been obtained by luminescence, magnetic circular dichroism (CD), nuclear quadrupole resonance, and extended Hückel molecular orbital (EHMO) calculations [8]. Xray crystallography has revealed the solid state structure of trans-(NH3)2PtCl2 [9,10]: The compound crystallizes in the monoclinic system, space group P21/a with a = 7.99(1) Å, b = 6.00(1) Å, c = 5.45(1) Å, β = 95.2(2)°, U = 260.2 Å3, Z = 2 (120 ± 5 K). Bond lengths are 2.05(4) Å for PtN and 2.32(1) Å for PtCl. In many cases, a comparison with the corresponding cis analog has been made. A convenient method of differentiation of the two isomers has been Raman spectroscopy, due to differences in molecular symmetry (D2h for the trans isomer; C2v for the cis isomer) and characteristic differences in skeletal vibrational modes [11]. Alternative methods applied involve derivatization of either isomer (allyl alcohol [12] or thiourea [13,14]) and subsequent analysis by ultraviolet-visible (UV-vis) spectroscopy [12] or high-performance liquid chromatography (HPLC) [13,14]. Thiourea (tu) derivatization (''Kurnakow test" [15]) of cis- and trans-(NH3)2PtCl2 leads to yellow [Pt(tu)4]Cl2 and colorless trans[Pt(NH3)2(tu)2]Cl2, respectively. Both compounds have been crystallized and X-ray structurally characterized [16]. 195Pt nuclear magnetic resonance (NMR) spectroscopy, despite its improvement in sensitivity in recent years, does not readily differentiate between the two isomers, e.g., δ195Pt, 2101 ppm for trans-(NH3)2-PtCl2 and 2104 ppm for cis-(NH3)2PtCl2 [4]. Similarly, chemical shifts in the 15N NMR spectra differ by only 0.4 ppm (66.3 and 65.9 ppm, respectively [4]), but J(Pt-N) coupling constants are markedly different, 278 Hz for the trans isomer and 303 Hz for the cis form.
< previous page
page_108
next page >
< previous page
page_109
next page > Page 109
2.3 trans-(NH3)2PtCl2 Analogs; Complex Isomerization Analogs of transplatin are generated by changing the ammonia ligands, the halogens, or the square planar metal ions. In the following, only Pt compounds containing N and Cl donor atoms will be considered, with a brief look at the Pd(II) analog of transplatin. The simplest Pt analogs of transplatin are the methylamine compound, trans-(MeNH2)2PtCl2 [17], the dimethylamine compound, trans-(Me2NH)2PtCl2 [18], as well as the trimethylamine species, trans-(Me3N)2PtCl2 [19]. X-ray crystal structure analyses are available for the MeNH2 [17] and Me2NH [18] compounds. A number of analogs with larger amine ligands, e.g., trans-(cyclohexylamine)2PtCl2 [20], has likewise been characterized. There appears to be no X-ray crystal structure analysis available of a mixed amine complex of the type trans-(a)-(a')PtCl2 except for an example with a = NH3 and a' = 1-methylcytosine [21] (see also Sec. 6.7). Finally, there is a long list of transplatin analogs containing N-heterocyclic ligands. For example, trans-(py)2PtCl2 is prepared in analogy to transplatin from [Pt(py)4]Cl2 in aqueous HCl [22]. Alternatively, it can also be obtained from the corresponding cis compound, prepared from K2PtCl4 and pyridine (py) in aqueous solution, upon isomerization in nonaqueous solvents such as dimethylsulfoxide (DMSO) or dimethylformamide (DMF) in the presence of free ligand (py) [23,24]. With substituted pyrimidines (pym), an analogous isomerization has been reported [24] and verified by Xray analysis [25]. Interestingly, the reverse processisomerization from the trans to the cis formis accomplished in 4 M HCl in the heat [25]. This process involves protonation of the coordinated pym ligand and temporary displacement of pymH+. Isomerization reactions in solution (during simple recrystallization [26] or photochemically [27]) or in the solid state [28] are a common phenomenon in Pt(II) chemistry [29], but in the majority of cases cis isomers are more prone to isomerization than trans compounds. The Pd(II) analog of transplatin is prepared in analogy to the latter from [Pd(NH3)4]2+ upon precipitation with HCl [30]. While structural analogs of trans-(NH3)2PdCl2 with NH3 replaced by heterocyclic ligands [31], including nucleobases [32], are known and X-ray structurally characterized, it is well known that the corresponding diaqua
< previous page
page_109
next page >
< previous page
page_110
next page > Page 110
species, trans-[(NH3)2Pd(H2O)2]2+, undergoes ligand disproportionation to [(NH3)Pd(H2O)3]2+ and [(NH3)3Pd(H2O)]2+ as well as isomerization to cis-[(NH3)2Pd(H2O)2]2+ [33]. For nucleobase chemistry related to this aspect, see also Sec. 6.8. 2.4 Solvolysis of trans-(NH3)2PtCl2 2.4.1 Mono- and Diaqua Species Solvolysis of the Cl ligands of trans-(NH3)2PtCl2 in water takes place in two steps (Fig. 1). Reported equilibrium constants for K1 range from 23.9 × 105 M [34], 32 × 105 M [35], and 48 ± 12×105 M [36] to 62.2 × 105 M [37]. K2 values are estimated to be ≤2 × 105 M [35,36]. These values indicate that, compared to cisplatin, spontaneous hydrolysis of the trans isomer is considerably reduced, at least by a factor of 10 for the first step and a factor of 20 for the second one. For practical purposes this means that chloride hydrolysis of trans-(NH3)2PtCl2 is only significant in very dilute solutions free of added Cl. In an acidic aqueous solution containing transplatin at a concentration of 102 M, for example, some 20 ± 3% (depending on K1 values used) of trans-[(NH3)2PtCl(H2O)]+ exists if equilibrium is reached. From the kinetic measurements [3437] that were used for calculating the thermodynamic equilibrium constants, it is evident that, although the rate constant k1 for the first hydrolysis step is faster for the trans isomer as compared to the cis isomer (consequence of the higher kinetic trans effect of Cl over NH3), the reverse reaction (k1) is likewise faster for the trans isomer, thus leading in essence to a smaller thermodynamic equilibrium constant K1. The second hydrolysis step k2 is faster for the cis isomer. Reported rate constants for transplatin are as follows: k1 [s1], 9.8 × 105 [35], 1.9 × 105 [37], k1 [M1 s1], 3.05 ×
Fig. 1. Stepwise hydrolysis of trans-(NH3)2PtCl2.
< previous page
page_110
next page >
< previous page
page_111
next page > Page 111
102 [37]; k2 [s1],
< previous page
page_112
next page > Page 112
As with hydrolysis products of cisplatin or related cis-diammineplatinum(II) compounds [41], aqua and hydroxo species of trans-a2Pt(II) undergo condensation reactions with formation of di- and oligo-µ-OH complexes [4]. Since only single OH bridges between Pt centers are possible with trans compounds, they may be less robust than the cyclic compounds formed by the cis isomer. 2.4.3 Selected Examples The only X-ray structurally characterized hydrolysis product of transplatin, reported by Arpalahti and coworkers [42], is trans-(NH3)2-PtCl(OH)·H2O (Fig. 3). It was prepared by careful NaOH hydrolysis of trans-(NH3)2PtCl2. Two forms of a DMSO solvolysis product of transplatin, trans-[(NH3)2Pt(DMSO-S)Cl]+, differing in the counterion ((ClO4)0.8Cl0.2 [43] and Cl [44]), have likewise been characterized by single-crystal X-ray crystallography. The fast kinetics of its formation (half-life of trans-(NH3)2PtCl2 in DMSO, 37°C, 8 min), combined with the known cislabilizing effect of DMSO (which could lead to loss of NH3), warn against the use of DMSO for the poorly soluble transplatin in any biological experiments [43].
Fig. 3. X-ray structurally characterized hydrolysis product of trans-DDP: trans-(NH3)2PtCl(OH)·H2O. (Reproduced by permission from [42].)
< previous page
page_112
next page >
< previous page
page_113
next page > Page 113
3 Biological Effects 3.1 Toxicity, Antitumor Activity, Mutagenicity of trans-DDP Both cis- and trans-(NH3)2PtCl2 are cytotoxic to prokaryotes and eukaryotes, both isomers exhibit antitumor activity (which, however, is very weak in the case of trans-DDP), and both are mutagenic [45]. There is substantial evidence that both isomers inhibit DNA replication and that this is caused by interference with the template rather than an effect on the polymerase machinery. Cell killing appears to be a consequence of inhibition of DNA synthesis. However, in general, trans-DDP is less toxic by a factor of 520 than cis-DDP, and this is the main reason why transDDP is not useful as an antitumor agent (see, however, below). For example the ID50 doses for cytotoxicity in cultured L1210 leukemia cells are 2.3 µM for cis-DDP but 67 µM for trans-DDP [46], and T/Cx100 values for a P388 leukemia tumor system in mice are 205% and 116% for the two isomers, respectively [47]. Similarly, the mutagenic potency of trans-DDP as measured for the reversion of his+ mutants of E. coli, for example, is lower by a factor of ~10 [45]. 3.2 Novel Active trans Compounds Several exceptions to the rule-of-thumb of two good leaving groups in cis orientation necessary for antitumor activity of Pt(II) complexes have been reported in recent years. These ''exceptions," which frequently show activity against cisplatin-resistant cell lines, are of the following types: (1) trans-L2PtCl2 with L = planar heterocyclic ligand [4749], (2) bis(platinum(II)) complexes with bridging diamine ligands of composition trans-{[PtCl(NH3)2]2(µNH2(CH2)xNH2)}Cl2 [50], and (3) trans-L2PtCl2 with L = imino ether in a head-tail arrangement ("trans-EE") [51]. While it is too early to speculate on possible modes of action of these compounds, it appears safe to assume that isomerization reactions to the corresponding cis forms do not take place. Some binding studies with mono- [52] and dinucleotides [53] as well as DNA [51,54] have been performed.
< previous page
page_113
next page >
< previous page
page_114
next page > Page 114
4 Reactions with Nucleic Acids 4.1 Kinetics of DNA Adduct Formation If hydrolysis of the first chloride of (NH3)2PtCl2 is the rate-determining step in reactions with biomolecules, transplatin is expected to react faster to the monofunctional DNA adduct than cisplatin (cf. Sec. 2.4.1). Experimentally determined rate constants for reactions of trans-(NH3)2PtCl2 with double-stranded DNA are indeed in the range for Cl hydrolysis, k being between (5.4 ± 0.9) × 105 sec1 [55] and 9.6 × 105 sec1 [56]. These values do not differ significantly from those of cis-DDP (k = (10.2 ± 0.7) × 105 sec1 [56]), however. Reported lifetimes of the monoadducts (before closure to bisadducts) vary between t1/2 = 3.1 hr [56] and 24 hr [57] to 30 hr [58]. Possibly these discrepancies are due to differences in methods applied, 195Pt NMR spectroscopy [56] and mono-adduct trapping by nucleophiles [57,58]. For comparison, closure of the cisplatin monoadducts to the crosslinks occurs somewhat faster, with k = (9.2 ± 1.4) × 105 sec1 to (4.55) × 105 sec1, corresponding to t1/2 = 2.1 ± 0.3 hr [56] and 4 hr [59], respectively. 4.2 Spectrum of DNA Adducts of trans-DDP The characterization of DNA adducts of trans-(NH3)2Pt(II) [60,61] is not as detailed and complete as that of cisplatin adducts. Monofunctional adducts with guanine dominate after short incubation times: Within 1 hr, 85% of all transplatin bound to double-stranded DNA ends up at guanine [60], presumably at the N(7) position [61]. Bifunctional crosslinks, analyzed after removal of monofunctionally bound trans-(NH3)2Pt(II) by glutathione and subsequent enzymatic digestion, are distributed as follows: ~50% Pt(dG)(dG), 40% Pt(dG)2, and 10% Pt(dA)(dG) for double-stranded DNA, and ~60% Pt(dG)2, 35% Pt(dA)(dG), and 5% Pt(dG)(dC) for single-stranded DNA (rb = 0.01). Due to enzymatic cleavage, these values do not provide numbers of actual adducts but only give an indication of transplatin preference for nucleobase donor sites. Long-range adducts, e.g., between two G's separated by other bases, show up as Pt(dG)2 only. The same applies to 1,3 adducts in GNG or ANG sequences which have been detected using a replication mapping assay
< previous page
page_114
next page >
< previous page
page_115
next page > Page 115
[62]. As far as adenine-containing adducts are concerned, both N(1) and N(7) binding of trans-(NH3)2Pt(II) takes place. According to Eastman [63], 2% of all bifunctional adducts formed with duplex DNA represent interstrand crosslinks. This figure is higher by at least a factor of 2 as compared to cisplatin [63,64]. The preferred interstrand crosslink appears to be one between the two complementary bases of a GC pair [65]. DNA-protein crosslinks are likewise more frequent for the trans isomer by a factor of 46 [66]. Summarizing these findings, it is clear at this stage that the high preference of cisplatin for adjacent guanines (5060% intrastrand GG adducts) and neighboring guanines and adenines (2030% intrastrand AG adducts) is not observed for transplatin. The preferred types of adducts (1,2 intrastrand for cisplatin; sterically impossible for transplatin) are different for the two isomers and there appears to be a greater variation of transplatin as far as nucleobase binding sites are concerned. 4.3 Macroscopic Effects of Transplatin Binding DNA secondary structure modifications as a consequence of transplatin binding, evident from fluorescence, UV, and CD spectroscopy as well as electron microscopy and melting behavior, were summarized many years ago [67,68]. From this work it is evident that both cis- and trans-(NH3)2PtCl2, yet not (dien)Pt(II), distort DNA greatly and that both isomers, presumably via long-distance crosslinks, lead to an apparent DNA shortening at high platination levels. Paradoxically, binding of the trans isomer led to an increase in melting temperature, very much as dienPt(II) despite the structural distortion evident from spectroscopic data in the case of transplatin. It is likely that the high number of monofunctional adducts, which thermally stabilize DNA by charge neutralization as well as favorable H bonding between ammine protons and phosphate oxygens, overcompensate the distorting effects of the crosslinks, which are fewer in number (see, however, Sec. 5.2). Today it is generally assumed that transplatin crosslinks disrupt DNA secondary structure more profoundly than do bifunctional cisplatin adducts. Work with antibodies, raised to recognize nucleosides in a distorted DNA [69], supports this view.
< previous page
page_115
next page >
< previous page
page_116
next page > Page 116
4.4 Staining of tRNAs In several cases trans-(NH3)2PtCl2 was applied to obtain heavy-atom derivatives of tRNAphe crystals. Relatively short soaking times (13 days) of the monoclinic form lead to selective binding of a trans-(NH3)2PtCl+ entity to the N(7) position of Gm34 in the anticodon loop [70,71]. Both NH3 groups are involved in hydrogen bonding: one group forms an H bond to O(6) of the nucleobase; the second NH3 hydrogen bonds to 5'-phosphate oxygens. Longer soaking (23 days) and use of the orthorhombic form of the tRNAphe change the picture [72]: trans-(NH3)2Pt(II) then shows a clear preference for the N(1) position of adenine-73 with only weak binding to guanine G18, Gm34, and G43. Again, NH3 groups are involved in H bonding in the case of A73 binding to adjacent nucleobases (G1 and C74), yet not to phosphate oxygens. Interestingly, cisplatin binds quite differently, with a preference for G15 and G18 [72]. 5 Reactions with Defined Oligonucleotides 5.1 Single-Stranded Oligonucleotides The 1,3 crosslinks of trans-(NH3)2Pt(II) with the trinucleotides d(GTG) [43,7375] and d(GCG) [76] as well as the hexanucleotide d(AGGCCT) [77] have been obtained and were studied by NMR spectroscopy and/or HPLC. In all cases these adducts were the major compounds formed although not the exclusive ones. According to 1H NMR spectroscopy (pH dependence of nonexchangeable base protons), Pt binding is always via the purine-N(7) positions and the intervening base is destacked from the two platinated bases. In the case where sugar conformations were determined, it was found that the sugar at the 5' purine changes from C2'-endo to predominantly 3'-endo as a consequence of backbone distortion. 5.2 1,3-Intrastrand Crosslinking in Double-Stranded Oligos Reaction of the single-stranded dodecanucleotide d(CCTCGAGTCTCC) with trans-(NH3)2PtCl2 yields the 1,3 adduct with Pt bound to the two
< previous page
page_116
next page >
< previous page
page_117
next page > Page 117
central guanines in the GAG sequence [78]. Structural changes are similar to those mentioned above, e.g., with A destacked. This metalated 12 mer anneals with the complementary, unmetalated 12 mer d(GGAGACTCGAGG) to give a duplex at low temperature. A model, generated by molecular dynamics calculations, suggests that the two crosslinked guanines are (1) in a head-head arrangement, (2) slightly propeller-twisted, and (3) strongly tilted with regard to the helix axis, pulling the complementary cytosines with one guanine and disrupting H bonding with the other one. The intervening adenine is completely rotated out of the helix and sticks in the minor groove (Fig. 4). The overall effect of this crosslink on the helix nevertheless appears to be relatively small as far as kinking (~ 18°) is concerned. Moreover, the melting temperature of the duplex is not, as with cis-DDP 1,2-intra-
Fig. 4. Stereo views of B-DNA model of d(CCTCGAGTCTCC)· d(GGAGACTCGAGG) (top) and trans-(NH3)2Pt(II) metalated form (bottom). The Pt crosslink is between G(5) and G(7) of d(CCTCGAGTCTCC), and the model has been derived from molecular dynamics. (Reproduced by permission from [78].)
< previous page
page_117
next page >
< previous page
page_118
next page > Page 118
strand crosslinks, lowered but rather is slightly higher. Interestingly, the ammine protons of the Pt entity become involved in H bonding with the complementary strands. In a similar study on GTG 1,3 crosslinks [79] in longer duplex oligonucleotides, a marked decrease in melting temperature on trans-DDP binding has been observed. Thus the intervening base and its location within the duplex may be important. 5.3 Interstrand G,C Crosslinking A series of single-stranded oligonucleotides (1922 mers) containing a central d(TGCT) sequence were reacted with trans-DDP to give a monofunctional adduct with the central guanine [80]. Upon annealing with the respective complementary strand, interstrand crosslink formation was observed. On the basis of results obtained with various DNA probes it was concluded that crosslinking occurs with the complementary cytosine of the second strand and gel mobility studies revealed that the GC interstrand crosslink unwinds DNA by ~12° and bends the double helix by ~26°. Molecular mechanics calculations suggest that in this crosslink guanine switches from anti to syn conformation, thereby allowing H-bond formation between guanine-O6 and cytosine-N(4)H2. This arrangement has also been observed in the X-ray crystal structure analyses of two model compounds containing simple model nucleobases [81,82] (see also Sec. 6.3). 5.4 Linkage Isomerization Reactions Although Pt(II) and in particular Pt(IV) complexes are generally considered to form kinetically inert complexes, there are several well-documented examples of linkage isomerization reactions in model systems [8386]. There are now also known two cases of linkage isomerization of trans-(NH3)2Pt(II) with oligonucleotides. The first one, observed in the single-stranded dodecamer d(TCTACGCGTTCT) by Comess et al., involves isomerization of a 1,3-G,G adduct to a 1,4-C,G adduct [87]. The 1,3 crosslink forms at pH 3.8 as the major species but rearranges to the 1,4 crosslink at pH ≥4.5. This
< previous page
page_118
next page >
< previous page
page_119
next page > Page 119
reaction is most unusual considering the high kinetic preference of any Pt electrophile for guanine-N7 and is as yet unexplained. In a second example, reported by Dalbiès et al. [88], rearrangement of a 1,3 adduct of a GAG sequence in a 22 mer to an interstrand adduct upon addition of the complementary strand (deoxyribonucleotide or ribonucleotide) takes place. This finding is in marked contrast to the behavior of the platinated dodecamer discussed in Sec. 5.2, which also contains a 1,3 adduct on a GAG sequence yet forms a duplex with retention of the original Pt lesion [78]. DNA sequence may be the clue to explain these apparent discrepancies. Feldmann and Lippert's studies on the reactivity of trans-(NH3)2Pt d(GNG) crosslinks with dCMP provide no evidence on an inherent instability of these 1,3 adducts [89]. 6 Model Studies Early solution studies, in particular by Tobias and coworkers, on mononucleotide interactions with Pt(II) electrophiles, including trans-DDP [9092], provided a picture of the principal Pt binding sites and established the kinetic rather than thermodynamic control of these reactions. Recent oligonucleotide and DNA work (cf. Secs. 4.2 and 5) established that apart from G-N7, at least also A-N7, A-N1, and C-N3 are binding sites of trans-DDP, while from the above model studies adenine binding had been ruled out for trans-DDP, which once more underlines the significance of kinetics. For a long time, structural work on Pt-nucleobase compounds has concentrated almost exclusively on cisplatin compounds (for reviews, see, e.g., [9395]). For example, while there are crystal structure analyses available for diand trinucleotide complexes of cis-(NH3)2Pt(II) [96,97] and dienPt(II) [98] for several years, there is none for a trans-(NH3)2Pt(II) complex with either a mono- or oligonucleotide or even a nucleoside. Structural work, carried out mainly in our group, has utilized model nucleobases such as 9-methyl- or 9-ethylguanine (9-MeGH, 9-EtGH), 9methyladenine (9-MeA), 1-methylcytosine (1-MeC), 1-methylthymine (1-MeTH), and 1-methyluracil (1-MeUH). Occasionally modified forms were applied, e.g., 1,5-dimethylcytosine (1,5-DimeC), 1,3-dimethyluracil (1,3-DimeU), 7,9-dimethylguanine (7,9-DimeG), or
< previous page
page_119
next page >
< previous page
page_120
next page > Page 120
7,9-dimethylhypoxanthine (7,9-DimeHyp). Despite the simplicity of these bases as compared to the corresponding nucleosides and nucleotides, these studies have led to a large number of interesting results, both with regard to structure and in some cases to reactivity. As with the cis isomer, the principal adducts of trans-a2Pt(II) are of the following types: (1) four 1:1 compounds, i.e., a2PtLX with L = G, A, C, T; (2) four bis(nucleobase) compounds of type a2PtL2; and (3) six mixed-base compounds, i.e., a2PtGA, a2PtGC, a2PtGT, a2PtAT, a2PtAC, and a2PtTC. The actual number of possibilities is higher if linkage isomers (e.g., purine-N1 or N7), nucleobase bridging (e.g., purine-N1, N7) or additional variations (e.g., loss of amine ligands) are considered. A list of all X-ray structurally characterized complexes of trans-a2Pt(II) (a = NH3 or CH3NH2) containing any of the above model nucleobases is given in Table 1. A large number of the other possible adducts have been prepared and spectroscopically identified yet have not been obtained as single crystals suitable for X-ray crystallography. In the following, selected features of some of the compounds will be presented. 6.1 Mono(nucleobase) Complexes Of all the 1:1 complexes prepared (L = 1-MeC [99], 9-EtGH [105,112], 1-MeT [109], 1,3-DimeU [100]), the 1,3dimethyluracil compound, trans-(CH3NH2)2Pt(1,3-DimeU-C5)Cl, is the most unusual one in that rather than binding to an exocyclic oxygen, as seen with various other metal ions [113], Pt(II) is bound via the C(5) position with displacement of the aromatic proton. Formation of this organometallic compound proceeds slowly, in contrast to Hg(II), which quickly forms an analogous compound [100,103]. Formation of the C-bound species can be easily followed by 1H NMR spectroscopy: The H(6) uracil proton (singlet) displays a characteristic 3J coupling with the 195Pt isotope of 74 Hz. The solution chemistry of trans-[a2Pt(9-EtGH-N7)(H2O)]2+ resembles that of the corresponding cis compound [114]; in the absence of additional nucleobases, it rapidly self-condensates with deprotonation of the guanine N(1) position [105], e.g.
< previous page
page_120
next page >
< previous page
next page >
page_121
Page 121 TABLE 1 Model Nucleobase Complexes of trans-a2Pt(II) Characterized by X-Ray Crystal Structure Analysis Ratio L:Pt 1:1
2:1
Compound
X
Orient. of La Ref.
[a2PtLCl]X L = 1-MeC-N3
CH3NH2 Cl
L = 1,3-DimeU-C5
CH3NH2
99 100
[a2PtL2]X2 L = 1-MeC-N3
2:1
a
NH3
NO3
h-t
101
NH3
ClO4
h-t
102
CH3NH2 ClO4
h-t
104
CH3NH2 ClO4
h-h
104
CH3NH2 NO3, PF6
h-h
104
L = 9-EtGH-N7
CH3NH2 Cl
h-t
105
L = 9-MeA-N7
NH3
ClO4
h-t
106
L = 7,9-DimeG-N1
CH3NH2 NO3
h-t
107
L = 7,9-DimeHyp-N1
NH3
NO3
h-t
108
NH3
NO3
H
81
H
82
W-C
107
W-C
109
[a2PtLL']Xn L = 1-MeC-N3 L' = 9-EtGH-N7 L = 1-MeC-N3
CH3NH2 Cl
L' = 9-MeGH-N7 L = 1-MeC-N3
NH3
PF6
L' = 7,9-DimeG-N1 L = 1-MeT-N3
CH3NH2 ClO4
L' = 9-MeA-N1 L = 1-MeT-N3
NH3
ClO4
H
109
NH3
ClO4
H
110
NH3
NO3
106
CH3NH2 ClO4
111
L' = 9-MeA-N7 L = 1-MeC-N3 L' = 9-MeA-N7 L = 9-MeGH-N7 L' = 9-MeA-N7 1:2
[Cla2PtLPta2Cl]Xn L = 9-MeA-N7,N1
ah-t = head-tail; h-h = head-head; H = Hoogsteen-modified; W-C = Watson-Crickmodified.
< previous page
page_121
next page >
< previous page
page_122
next page > Page 122
The trans-[a2PtLCl]Cl compounds of both L = 9-EtGH-N7 [105] and 1-MeC-N3 [115] undergo in aqueous solution rather unexpected symmetrization reactions of the kind:
An analogous disproportionation reaction (in ''more concentrated solution") has also been reported for trans[a2Pt(H2O)Cl]+ [4]. The potential significance of these reactions is at hand: They imply that monofunctionally bound trans-DDP may remove itself from DNA as a consequence of the trans effect of the Cl ligand. At the same time these reactions strongly suggest that any ligand with a reasonably strong trans effect, e.g., any S donor, should be capable of removing mono adducts from DNA rather easily. Even though we have shown that Cl is capable of displacing an NH3 group trans to itself in cisplatin-nucleobase adducts [101], such a reaction, if occurring on DNA, would not lead to removal of Pt from DNA. This difference could account for and explain findings that repair of cisplatin DNA adducts requires an excision or SOS-type mechanism [45], whereas repair of trans-DDP monoadducts appears to occur simply chemically. We note that nucleobase displacement from trans-[a2PtLCl]Cl has been observed in a mass spectrometric study as early as 1974 [116]. However, these findings appear to have been widely ignored due to the nonphysiological conditions of the experiment. 6.2 Bis(nucleobase) Complexes and Derivatives 6.2.1 Trans-a2PtL2 Compounds The solid state structures of these compounds (cf. Table 1) in most cases show head-tail orientations of the two nucleobases which then are coplanar and form dihedral angles of 7580° with the PtN4 coordination planes. Exceptions are the heteronuclear derivatives from trans-[a2PtL2]2+ (L = 1-MeC) discussed in Sec. 6.2.2 and two examples of trans-[(CH3NH2)2Pt(1-MeC-N3)2]2+, which were obtained from the heteronuclear Pt,Hg derivative upon removal of Hg(II) and subsequent rapid crystallization [104]. There the 1-MeC planes are markedly propellertwisted [55.9(11)°]. 1H NMR spectra of the bis(guanine) and the bis(adenine) compounds in D2O are rather simple, probably due to rapid ligand rotation
< previous page
page_122
next page >
< previous page
page_123
next page > Page 123
about the Pt-N(7) bonds. In contrast, with trans-[a2Pt(1-MeC-N3)2]2+ two sets of 1H NMR resonances of H(5) and H(6) protons are observed, corresponding to a distribution of head-head and head-tail rotamers in a 1:3 ratio. 6.2.2 Heteronuclear Derivatives of trans-[a2PtL2]n+ An outstanding feature of 1-MeU-N3 and 1-MeT-N3 complexes of cis-a2Pt(II) is their propensity to self-aggregate via exocyclic oxygens and/or to bind additional metal ions via these sites [93]. As a consequence, di- and multinuclear complexes form, a large variety of which have been characterized. As far as electronic properties are concerned, trans-a2PtL2 compounds (L = 1-MeU-N3 or 1-MeT-N3) should behave similarly. However, the extremely low solubility of trans-a2PtL2 has prevented the development of an analogous chemistry. Only in one case, with Ag+, could the solubility problem be overcome and a crystalline polymeric compound of composition trans-(NH3)2Pt(1-MeU)2Ag2(NO3)2(H2O)·H2O isolated and X-ray structurally characterized [117,118]. In this compound two Ag's are bound pairwise by O(4) and O(2) oxygens of the 1-MeU ligands, with O(4) bridging Ag ions from adjacent PtAg2 units. Starting from the better soluble uridine (urd) compound, with Cu(II) an orange-colored derivative of composition trans-[a2PtL2CuL2-Pta2]2+ (L = urd) had been isolated, but its crystal structure could not be obtained [118]. The unexpected color of this complex initiated subsequent work with related 2-pyridonate compounds, which eventually afforded two X-ray crystal structure analyses of Pt2Cu derivatives and provided an insight into the reasons for the unexpected color [119]. The restrictions set by the N,O bite of these ligands and the inability of the trans[a2PtL2M]n+ entity to adjust the Pt . . . M separation via a tilting of the Pt and the M coordination planes, as is possible with the cis-[a2PtL2M]n+ analog [120], are responsible for generally shorter Pt . . . M distances and/or more severe distortions of the metal ion geometries in the trans compounds (Fig. 5). A large number of heteronuclear derivatives of trans-[a2PtL2]2+ (L = 1-MeC-N3 or 1,5-DimeC-N3) have been prepared and in many cases structurally characterized. In these compounds, the L ligands are oriented head-head and the exocyclic amino groups of both cytosines are singly deprotonated each and binding the heterometal (Fig. 6). Hetero-
< previous page
page_123
next page >
< previous page
page_124
next page > Page 124
Fig. 5. Basic types of heteronuclear (Pt, M) complexes derived from (a) cis-a2PtL2 (b) trans-a2PtL2 (L = uracil or thymine) with more flexible Pt . . . M separations in the case of the cis isomer (due to tilting of Pt and M planes). metals studied so far include Pd(II) [121123], Cu(II) [124], Hg(II) [125], and Ag(I) [126]. Surprisingly, deprotonation of the exocyclic amino groups proceeds readily, in the case of Pd(II) even at acidic pH. In the mixed Pt,Pd compounds, the coordination planes of the two metals are perpendicular to each other, and Pd(II) utilizes Pt(II) as a ligand. This way Pt becomes five-coordinate and short Pt-Pd bonds of around 2.50 Å form. A typical example is given in Fig. 7. According to EHMO calculations, the Pt-Pd bond in these compounds is best described as a two-electron/two-orbital donor-acceptor interaction [123]. The formal bond order of 1 in mixed Pt,Pd compounds of this kind is progressively reduced as the d8 metal ion Pd(II) is replaced by Cu(II) (d9) and eventually Hg(II) (d10). Pt-Cu distances are
Fig. 6. Basic types of heteronuclear (Pt, M) complexes derived from trans-a2Pt(1-MeC)2 with (a) M = Pd(II) (b) M = Cu(II) (c) M = Hg(II). Coordination geometries of the metals are square planar (a), square planar or square pyramidal (b), and trigonal bipyramidal (c).
< previous page
page_124
next page >
< previous page
page_125
next page > Page 125
Fig. 7. ORTEP drawing of the cation of trans-[(NH3)2Pt(1-MeC)2Pd(1-MeU)]-NO3·3H2O. The two metal coordination planes are virtually perpendicular to each other and the PtPd bond is 2.515(4) Å. (Reproduced by permission from [121].) still quite short though, ~2.502.56 Å, but Pt-Hg separations are substantially longer, 2.772.84 Å. Nevertheless, even in the latter case Pt ''feels" the vicinity of Hg, as evident from the 199 Hg satellites in the 195Pt NMR spectrum (1J = 2783 Hz) [125]. trans-[a2Pt(1-MeC)2PdCl]+ ("PtPdCl") is a particularly versatile material for the preparation of many derivatives with Cl replaced by simple anions, small molecules, and larger molecules such as nucleobases [122], amino acids and dipeptides [127]. 195Pt NMR spectroscopy has proven very useful in establishing the nature of the donor atom of Pd and trans to Pt [122]. Work is in progress to compare nucleobase DNA and protein binding properties of "PtPdCl" with that of simple, mononuclear species such as (dien)Pd(II). 6.3 Mixed Nucleobase Complexes 6.3.1 Metal-Modified Base Pairs We have used the term "metal-modified base pairs" to describe a situation in which a proton involved in H bonding between two nucleobasesirrespective whether the bases are identical, complementary,
< previous page
page_125
next page >
< previous page
page_126
next page > Page 126
noncomplementary, or chargedis replaced by a metal of suitable geometry with the planarity of the bases maintained (Fig. 8) [109]. Trans-a2Pt(II) is suitable in this sense, whereas cis-a2Pt(II) is not since it forces two bases bound to it into large dihedral angles, thereby disrupting any H bonds that might exist between these bases if arranged in a coplanar fashion. X-ray structurally characterized examples are included in Table 1. Two examples, formally corresponding to WatsonCrick and Hoogsteen analogs of adenine-thymine pairs, are depicted in Fig. 9 [109]. A special feature, observed also occasionally in related compounds [107,128], is the involvement of a solvent (H2O) molecule in the H bonding between two nucleobases. If additional H bonding of a metalated base pair with another, free nucleobase takes place, (mono)metalated base triples are formed. An example is trans-{[(CH3NH2)2Pt(1-MeC-N3)(9-MeGH-N7)]·(1MeC)}2+ [82], the analog of CH+=C≡G, which is formed in triplex DNA; there is a regular Watson-Crick base pair between 1-MeC and the platinated 9-MeGH, which is connected with the other 1-MeC in a Hoogsteen-like fashion. A possible relevance of this compound is discussed in Chap. 6 of this book. 6.3.2 Dimetalated Triples and Cyclic Quartets X-ray work on N(1),N(7)-diplatinated purines reveals that Pt-N vectors in these compounds are roughly at right angles [111,129]. This fact, in combination with the close-to-planar orientation of nucleobases (L, L') in trans[a2PtL2]n+ and trans-[a2PtLL']n+ complexes, lead to an essentially coplanar arrangement of three or more nucleobases connected in such a way. If intramolecular H bonding between exocyclic groups of
Fig. 8. Schematic representation of ''metal modification" of a nucleobase pair. (Reproduced by permission from [109].)
< previous page
page_126
next page >
< previous page
page_127
next page > Page 127
Fig. 9. ORTEP drawings of cations of (a) trans-[(CH3NH2)2Pt(1-MeT-N3)(9-MeA-N1)]ClO4·3.25H2O, and (b) trans-[(NH3)2Pt(1-MeT-N3)(9-MeA-N7)]ClO4·2.5H2O, corresponding to metal analogs of Watson-Crick (a) and Hoogsteen pairs (b) of T and A. In (a) an H2O molecule is involved in H bonding between the bases. (Reproduced by permission from [109].)
< previous page
page_127
next page >
< previous page
page_128
next page > Page 128
metallated nucleobases is possible, the entity gains additional rigidity [106,111]. A special case is realized if four metallated purines are combined in a cyclic fashion (Fig. 10). While the existence of metalated, cyclic base quartets remains to be established, it is noted that in mixed A2G2 quartets intracomplex H bonding between the exocyclic groups in the 6 position of the purines would be extremely favorable. 6.4 Combining cis- and trans-DDP Synergistic cytotoxicity between different cis-a2Pt(II) antitumor drugs has been reported [130]. Possible additive or synergistic effects of a combination of cis- and trans-DDP appear never to have been investigated, even though such a study might lead to interesting results. As far as model nucleobase chemistry is concerned, such combinations are feasible and may eventually lead to novel macrocyclic ring systems containing metal ions and nucleobases [131]. Based on the model compound cis{(NH3)2Pt[(9-MeA)Pt(NH3)3]2}(NO3)6·2H2O, which contains the two 9-MeA bases in a head-head arrangement, coordinated
Fig. 10. Two principal arrangements (I, II) of metalated, cyclic purine quartets. Particularly favorable intramolecular H bonding is expected for I if A's and G's alternate. (Reproduced by permission from [111].)
< previous page
page_128
next page >
< previous page
page_129
next page > Page 129
to cis-(NH3)2Pt(II) via the N(1) positions and to the (NH3)3Pt(II) entities via the N(7) positions, we have proposed such a possibility and are presently trying to verify it. The specific features of this compound are the right angles between the cis-(NH3)2Pt(N1)2 plane and the adenines, between the adenines themselves, and between the adenines and the (NH3)3Pt(N7) planes. 6.5 Trans-a2Pt(IV) Nucleobase Complexes A limited number of nucleobase complexes of type trans,trans,trans-[a2PtL2X2]n+, obtained via oxidation of the corresponding Pt(II) compound trans-[a2PtL2]n+ and in one case via subsequent linkage isomerization, respectively, are described. They include five 1-MeC [132,133] as well as two 1-MeU [118,134] compounds. X groups are OH, OH2, , and N(4)H of cytosine nucleobases. A similar guanine complex, trans,trans,trans-[(CH3NH2)2Pt(9MeGH)2(OH)2]2+, has likewise been prepared [135]. Unlike corresponding bis(guanine) complexes of cisplatin, which ''spontaneously" become reduced in solution [136], this species is robust and shows no signs of reduction in the long term. 6.6 Ternary Nucleobase/Amino Acid Complexes As mentioned in Sec. 4.2, DNA-protein crosslinks are more frequently observed for trans-DDP than for cis-DDP. Model studies involving isolated nucleobases (model nucleobases, nucleotides) and amino acids, N-acetyl amino acids as well as dipeptides, have been carried out in very few cases, with the emphasis on spectroscopic characterization [137139], on the kinetics of formation [140] and on structural characterization [99]. Available data on these somewhat simple models do not yet provide a picture on the possible significance of such adducts, either as far as biology is concerned or with regard to coordination chemistry aspects of interest. 6.7 Other trans-a2Pt(II) Nucleobase Complexes Well-characterized nucleobase complexes of the type trans-X2PtLL', with X = halogen and L and/or L' = nucleobase, are very rare. One
< previous page
page_129
next page >
< previous page
page_130
next page > Page 130
example is trans-K2[PtI2(1-MeU-N3)2]·6H2O, a compound crystallizing in a star-like supramolecular fashion with anionic trans-[PtI2(1-MeU)2]2 entities threaded to hexagons of K+ ions, which in turn are connected via H2O bridges [141]. Another example is trans-Cl2Pt(1-MeC-N3)(NH3), a compound formed from cis-[(NH3)2Pt(1-MeCN3)Cl]Cl upon loss of NH3 [21]. This reaction is still the only well-documented case of NH3 displacement by Cl in a cis-a2Pt(II) compound. Trans-Cl2Pt(1-MeC-N3)(NH3) has a considerable synthetic potential for the preparation of mixed ligand complexes of the type Pt(NH3)(1-MeC)(L)X [142], for tris(nucleobase) complexes of the composition trans-Pt(NH3)-(1-MeC)(L)2 or Pt(NH3)(1-MeC)(L)(L') [143,144], and, via the corresponding diaqua species trans[(H2O)2Pt(1-MeC-N3)(NH3)]2+, to di- and oligomeric 1-MeC bridged compounds [145]. 6.8 Trans-a2Pd(II) Nucleobase Complexes While a number of bis(nucleobase) complexes of the type trans-X2PdL2 (X = halogen) have been described and structurally assigned [32,146148], there appears to be just one example for X = NH3, for which an X-ray crystal structure is available, trans-[(NH3)2Pd(1-MeC-N3)2](NO3)2 [149]. In solution this compound undergoes complex rearrangement reactions, very similar to those of trans-[(NH3)2Pd(H2O)2]2+ (cf. Sec. 3.2 and [33]). 7 Summary The question raised, i.e., trans-diammineplatinum(II)what makes it different from its cis isomer?'' permits a number of answers, relating to various aspects of the chemistry of the two compounds, but the crucial one concerning antitumor activity is yet to be answered. In principle, any of the following reasons or combinations thereof might account for the observed differences in biological effects, but some are more likely than others: 1 Target Molecules It is generally accepted that DNA is the most important target molecule of cisplatin and that inhibition of DNA synthesis correlates with antitumor activity. This fact does not contradict
< previous page
page_130
next page >
< previous page
page_131
next page > Page 131
observations on extensive reaction with other biomolecules. The mutagenic effects of trans-a2Pt(II), although weaker than for cis-DDP, point toward DNA as being an important target for trans-DDP as well. 2 DNA Adducts As a consequence of the inherent difference in geometry, the two isomers form different adducts. Specifically, transDDP cannot form 1,2-intrastrand crosslinks, which represent by far the most abundant adducts of cis-DDP. On the other hand, the trans isomer displays a greater variation in nucleobase donor sites (cf. Sec. 4.2). 3 DNA Stability While the 1,2-intrastrand adducts of cis-DDP, regardless if GG or AG, consistently cause DNA kinking and lead to thermal destabilization, the effects of trans-DDP adducts seem to be of a greater variability, causing thermal stabilization or destabilization. For intrastrand 1,3 adducts in puXpu sequences, the intervening base X appears to be important in this respect (cf. Sec. 4.3). 4 DNA Repair In living cells, higher doses of trans-DDP as compared to cis-DDP are required to bind an equal number of Pt atoms per nucleotide [66]. On the other hand, bifunctional DNA adducts of either isomer inhibit DNA replication to the same extent [66,150,151]. This finding has been interpreted in terms of a differential repair of adducts of the two isomers [66], but an alternative explanation has also been offered [152]. It is to be noted that repair of monofunctional trans-DDP lesions does not require enzymatic repair but rather may be accomplished by any nucleophile within the cell exercising a reasonably high trans influence, e.g., an S-donor of glutathione. 5 Intrinsic Reactivity Differences in hydrolysis kinetics of the two isomers (unfavorable equilibrium concentration of trans-[(NH3)2PtCl(OH2)]+ [37]) and in reaction rates of various hydrolysis species with DNA constituents [36] could, in principle, explain a difference in biological effects. On the other hand, the kinetics of reactions of the dichloro species of both isomers with DNA (in the absence of any repair agents) appear not to be that largely different to produce a strong point for differential reactivity of this species. Model studies have shown at least one more distinct difference between mono(nucleobase) adducts of both isomers: While cis-[a2PtLCl]+ can lose the amine trans to Cl, in trans[a2PtLCl]+ Cl is capable of displaying L (cf. Sec. 6.1). Both reactions are not very fast with Cl, but any good nucleophile replacing Cl would do so with high efficiency. In the case of trans-DDP, such a reaction leads to removal of Pt from DNA, unlike in the case of cis-DDP (even though we
< previous page
page_131
next page >
< previous page
page_132
next page > Page 132
are aware that cis effects may also be operative). With respect to amine displacement from monoadducts of cis-DDP, it is interesting to speculate on the fate of (toxic) NH3. Is it capable of undergoing condensation reactions with biomolecules? In theory, such a scenario might provide possibilities for the role of the amine ligands in cis(amine)2Pt(II) compounds alternative to those commonly accepted (e.g., role of NH protons in stabilizing DNA adducts [153]). 6 Transport and Distribution Despite differences in water solubility of the two isomers (trans-DDP less soluble), this property is unlikely to be important at physiological concentrations. There is also no evidence that either isomer is actively transported across cell membranes, but at least in one case, trans-DDP, has been found to penetrate better than does the active cis isomer [150]. Partition coefficients may, however, become significant with analogs of either isomer containing markedly lipophilic ligands. In conclusion, apart from the question of differential repair of DNA adducts of the two isomers of DDP as a consequence of structural differences in DNA distortion and their recognition, intrinsic reactivity differences of the two geometrical isomers should be given attention because they might be a clue to the question of activity vs. inactivity. Abbreviations
A
adenine, unspecified
a
amine
C
cytosine, unspecified
CD
circular dichroism
cisplatin, cis-DDP
cis-(NH3)2PtCl2
dien
diethylenetriamine
Dime
dimethyl
1,5-DimeC
1,5-dimethylcytosine
7,9-DimeG
7,9-dimethylguanine
7,9-DimeHyp
7,9-dimethylhypoxanthine
1,3-DimeU
1,3-dimethyluracil
DMF
dimethylformamide
DMSO
dimethylsulfoxide
< previous page
page_132
next page >
< previous page
page_133
next page > Page 133
EHMO
extended Hückel molecular orbital
Et
ethyl
G
guanine, unspecified
his
histidine
HPLC
high-performance liquid chromatography
Hyp
hypoxanthine
ID50
50% inhibitory dose
Ino
inosine
L, L'
ligand(s), unspecified
Me
methyl
9-MeA
9-methyladenine, neutral
1-MeC
1-methylcytosine, neutral
1-MeC
1-methylcytosine, deprotonated at N4
9-MeGH
9-methylguanine, neutral
1-MeTH
1-methylthymine
1-MeU
1-methyluracil, anion
1-MeUH
1-methyluracil, neutral
1-MeU-N3
1-MeU, metalated at N3, etc.
NMR
nuclear magnetic resonance
pu
purine
py
pyridine
pym
pyrimidine
rB
ratio Pt:nucleotide
T
thymine, unspecified
T/C
median survival time of treated vs. control animals
transplatin, trans-DDP
trans-(NH3)2)PtCl2
tu
thiourea
U
uracil, unspecified
urd
uridine
UV-vis
ultraviolet-visible
References 1. B. Rosenberg, L. VanCamp, J. E. Trosko, and V. H. Mansour, Nature, 222, 385 (1969).
< previous page
page_133
next page >
< previous page
page_134
next page > Page 134
2. Gmelin Handbuch der Anorg. Chemie, Vol. 68D, Verlag Chemie, Weinheim, 1957, pp. 236241. 3. (a) J. Reiset, Ann. Chim. Phys., 11, 417 (1844); (b) J. Reiset, Compt. Rend., 18, 1103 (1844). 4. T. G. Appleton, A. J. Bailey, K. J. Barnham, and J. R. Hall, Inorg. Chem., 31, 3077 (1992). 5. M. Peyrone, Ann. Chem. Pharm., 51, 1 (1846). 6. G. B. Kauffman and D. O. Cowan, Inorg. Synth., 7, 239 (1963). 7. F. R. Hartley, The Chemistry of Platinum and Palladium, Applied Science, London, 1973. 8. H. H. Patterson, J. C. Tewksbury, M. Martin, M.-B. Krogh-Jespersen, J. A. lo Menzo, H. O. Hooper, and A. K. Viswanath, Inorg. Chem., 20, 2297 (1981). 9. M. A. Porai-Koshits, Trudy Inst. Krist. Akad. Nauk. SSSR, 9, 229 (1954). 10. G. H. W. Milburn and M. R. Truter, J. Chem. Soc. (A), 1609 (1966). 11. H. Poulet, P. Delorme, and J. P. Mathieu, Spectrochim. Acta, 20, 1855 (1964). 12. C. P. Hicks and M. Spiro, J. Chem. Soc., Chem. Commun., 131 (1981). 13. J. Arpalahti and B. Lippert, Inorg. Chim. Acta, 138, 171 (1987). 14. J. D. Woollins, A. Woollins, and B. Rosenberg, Polyhedron, 2, 175 (1983). 15. N. S. Kurnakow, J. Russ. Phys. Chem., 25, 565 (1893). 16. J. Arpalahti, B. Lippert, H. Schöllhorn, and U. Thewalt, Inorg. Chim. Acta, 153, 51 (1988). 17. J. Arpalahti, B. Lippert, H. Schöllhorn, and U. Thewalt, Inorg. Chim. Acta, 153, 45 (1988). 18. J. A. Anderson, J. W. Carmichael, and A. W. Cordes, Inorg. Chem., 9, 143 (1970). 19. P. L. Goggin, R. J. Goodfellow, and F. J. S. Reed, J. Chem. Soc., Dalton. Trans., 1298 (1972). 20. G. Zanotti, A. Del Pra, G. Bombieri, and A. M. Tamburro, Acta Crystallogr., B34, 2138 (1978). 21. B. Lippert, C. J. L. Lock, and R. A. Speranzini, Inorg. Chem., 20, 808 (1981).
< previous page
page_134
next page >
< previous page
page_135
next page > Page 135
22. G. B. Kauffman, Inorg. Synth., 7, 249 (1963). 23. P. C. Kong and F. D. Rochon, Can. J. Chem., 56, 441 (1978). 24. P. C. Kong and F. D. Rochon, Can. J. Chem., 57, 526 (1979). 25. F. D. Rochon, P. C. Kong, and R. Melanson, Can. J. Chem., 59, 195 (1981). 26. See, e.g.: C. J. L. Lock, and M. Zvagulis, Acta Crystallogr., B36, 2140 (1980). 27. J. R. Perumareddi and A. W. Adamson, J. Phys. Chem., 72, 414 (1968). 28. J. R. Bales, C. J. Coulson, D. W. Gilmour, M. A. Mazid, S. Neidle, R. Kuroda, B. J. Peart, C. A. Ramsden, and P. J. Sadler, J. Chem. Soc., Chem. Commun., 432 (1983). 29. Yu. N. Kukushkin, Platinum Metals Rev., 35, 28 (1991). 30. (a) H. D. K. Drew and G. H. Wyatt, J. Chem. Soc., 56 (1934). (b) G. B. Kauffman and J. H. Tsai, Inorg. Synth., 8, 234 (1966). 31. See, e.g.: (a) O. Yamauchi, M. Takani, K. Toyoda, and H. Masuda, Inorg. Chem., 29, 1856 (1990). (b) F. J. Martin-Gil and J. Martin-Gil, Inorg. Chim. Acta, 137, 131 (1987). 32. M. Quirós, J. M. Salas, M. Purificación Sánchez, A. L. Beauchamp, and X. Solans, Inorg. Chim. Acta, 204, 213 (1993). 33. T. G. Appleton, J. R. Hall, S. F. Ralph, and C. S. M. Thompson, Aust. J. Chem., 41, 1425 (1988). 34. V. D. Panasuyk and N. F. Malashok, Russ. J. Inorg. Chem., 13, 1405 (1986). 35. F. Aprile and D. S. Martin, Jr., Inorg. Chem., 1, 551 (1962). 36. M. Mikola and J. Arpalahti, Inorg. Chem., 33, 4439 (1994). 37. S. E. Miller, K. J. Gerard, and D. A. House, Inorg. Chem. Acta, 190, 135 (1991). 38. K. A. Jensen, Z. Anorg. Allg. Chem., 242, 87 (1939). 39. S. J. Bernes-Price, T. A. Frankiel, U. Frey, J. D. Ranford, and P. J. Sadler, J. Chem. Soc., Chem. Commun., 789 (1992). 40. T. G. Appleton, J. R. Hall, S. F. Ralph, and C. S. M. Thompson, Inorg. Chem., 28, 1989 (1989). 41. See, e.g.: (a) R. Faggiani, B. Lippert, C. J. L. Lock, and B. Rosenberg, J. Am. Chem. Soc., 99, 777 (1977); (b) R. Faggiani, B. Lippert, C. J. L. Lock, and B. Rosenberg, Inorg. Chem., 17, 1941
< previous page
page_135
next page >
< previous page
page_136
next page > Page 136
(1978); (c) J.-P. Macquet, S. Cros, and A. L. Beauchamp, J. Inorg. Biochem., 25, 197 (1985); (d) F. D. Rochon, A. Morneau, and R. Melanson, Inorg. Chem., 27, 10 (1988). 42. J. Arpalahti, R. Sillanpää, and M. Mikola, J. Chem. Soc., Dalton Trans., 1499 (1994). 43. W. I. Sundquist, K. J. Ahmed, L. S. Hollis, and S. J. Lippard, Inorg. Chem., 26, 1524 (1987). 44. J.-M. Delafontaine, P. Khodadad, P. Toffoli, and N. Rodier, Acta Crystallogr., C41, 702 (1985). 45. N. P. Johnson, P. Lapetoule, H. Razaka, and G. Villani in Biochemical Mechanisms of Platinum Antitumour Drugs (D. C. H. McBrien and T. F. Slater, eds.), IRL Press, Oxford, 1986, pp. 128. 46. J. P. Macquet and J. L. Butour, J. Natl. Cancer Inst., 70, 899 (1983). 47. N. Farrell, T. T. B. Ha, J.-P. Souchard, F. L. Wimmer, S. Cros, and N. P. Johnson, J. Med. Chem., 32, 2240 (1989). 48. M. Van Beusichem and N. Farrell, Inorg. Chem., 31, 634 (1992). 49. N. Farrell, L. R. Kelland, J. D. Roberts, and M. Van Beusichem, Cancer Res., 52, 5065 (1992). 50. N. Farrell, Y. Qu, and M. P. Hacker, J. Med. Chem., 33, 2179 (1990). 51. M. Coluccia, A. Nassi, F. Loseto, A. Boccarelli, M. A. Mariggio, D. Giordano, F. P. Intini, P. Caputo, and G. Natile, J. Med. Chem., 36, 510 (1993). 52. Y. Qu and N. Farrell, J. Am. Chem. Soc., 113, 4851 (1991). 53. M. J. Bloemink, J. Reedijk, N. Farrell, Y. Qu, and A. I. Stetsenko, J. Chem. Soc., Chem. Commun., 1002 (1992). 54. Y. Zou, B. van Houten, and N. Farrell, Biochemistry, 32, 9632 (1993). 55. H. M. Ushay, T. D. Tullius, and S. J. Lippard, Biochemistry, 20, 3744 (1981). 56. D. P. Bancroft, C. A. Lepre, and S. J. Lippard, J. Am. Chem. Soc., 112, 6860 (1990). 57. A. Eastman and M. A. Barry, Biochemistry, 20, 3303 (1987). 58. J.-L. Butour and N. P. Johnson, Biochemistry, 25, 4534 (1986). 59. J.-M. Malinge and M. Leng, Nucl. Acids Res., 16, 7763 (1988).
< previous page
page_136
next page >
< previous page
page_137
next page > Page 137
60. A. Eastman and M. A. Barry, Biochemistry, 26, 3303 (1987). 61. A. Eastman, M. M. Jennerwein, and D. L. Nagel, Chem.-Biol. Interact., 67, 71 (1988). 62. A. P. Pinto and S. J. Lippard, Proc. Natl. Acad. Sci. USA, 82, 4616 (1985). 63. A. Eastman, Biochem. Biophys. Res. Commun., 105, 869 (1982). 64. J. J. Roberts and F. Friedlos, Biochim. Biophys. Acta, 655, 146 (1981). 65. V. Brabec and M. Leng, Proc. Natl. Acad. Sci USA, 90, 5345 (1993). 66. R. B. Ciccarelli, M. J. Solomon, A. Varshavsky, and S. J. Lippard, Biochemistry, 24, 7533 (1985). 67. J.-P. Macquet and J.-L. Butour, Biochimie, 60, 901 (1978). 68. J.-P. Macquet and J. L. Butour, Eur. J. Biochem., 83, 375 (1978). 69. W. I. Sundquist, S. J. Lippard, and B. D. Stollar, Biochemistry, 25, 1520 (1986). 70. A. Jack, J. E. Ladner, D. Rhodes, R. S. Brown, and A. Klug, J. Mol. Biol., 111, 315 (1977). 71. C. D. Stout, H. Mizuno, S. T. Rao, P. Swaminathan, J. Rubin, T. Brennan, and M. Sundaralingam, Acta Crystallogr., B34, 1529 (1978). 72. J. R. Rubin, M. Sabat, and M. Sundaralingam, Nucl. Acids Res., 111, 6571 (1983). 73. J. L. van der Veer, G. J. Ligtvoet, H. van den Elst, and J. Reedijk, J. Am. Chem. Soc., 108, 3860 (1986). 74. N. Boogaard and J. Reedijk, J. Inorg. Biochem., 43, 428 (1991). 75. N. Boogaard, C. Altona, and J. Reedijk, J. Inorg. Biochem., 49, 129 (1993). 76. D. Gibson and S. J. Lippard, Inorg. Chem., 26, 2275 (1987). 77. C. A. Lepre, K. G. Strothkamp, and S. J. Lippard, Biochemistry, 26, 5651 (1987). 78. C. A. Lepre, L. Chassot, C. E. Costello, and S. J. Lippard, Biochemistry, 29, 811 (1990). 79. M.-F. Anin and M. Leng, Nucl. Acids Res., 18, 4395 (1990). 80. V. Brabec, M. Síp, and M. Leng, Biochemistry, 32, 11676 (1993). 81. A. P. Hitchcock, C. J. L. Lock, W. M. C. Pratt, and B. Lippert in
< previous page
page_137
next page >
< previous page
page_138
next page > Page 138
Platinum, Gold, and Other Metal Chemotherapeutic Agents (S. J. Lippard, ed.), American Chemical Society, Washington, D.C., ACS Symp. Series 209, 1983, p. 209 ff. 82. I. Dieter-Wurm, M. Sabat, and B. Lippert, J. Am. Chem. Soc., 114, 357 (1992). 83. J. L. van der Veer, H. van den Elst, and J. Reedijk, Inorg. Chem., 26, 1536 (1987). 84. K. Inagaki, A. Tomita, and Y. Kidani, Bull. Chem. Soc. Jpn., 61, 2825 (1988). 85. T. V. O'Halloran and S. J. Lippard, Inorg. Chem., 28, 1289 (1989). 86. B. Lippert, H. Schöllhorn, and U. Thewalt, J. Am. Chem. Soc., 108, 6616 (1986). 87. K. M. Comess, C. E. Costello, and S. J. Lippard, Biochemistry, 29, 2102 (1990). 88. R. Dalbiès, D. Payet, and M. Leng, Proc. Natl. Acad. Sci. USA, 91, 8147 (1994). 89. G. Feldmann and B. Lippert, unpublished results. 90. G. Y. H. Chu, S. Mansy, R. E. Duncan, and R. S. Tobias, J. Am. Chem. Soc., 100, 593 (1978). 91. S. Mansy, G. Y. H. Chu, R. E. Duncan, and R. S. Tobias, J. Am. Chem. Soc., 100, 607 (1978). 92. M. R. Moller, M. A. Bruck, T. O'Connor, F. J. Armatis, Jr., E. A. Knolinski, N. Kottmair, and R. S. Tobias, J. Am. Chem. Soc., 102, 4589 (1980). 93. B. Lippert, Prog. Inorg. Chem., 37, 1 (1989). 94. T. J. Kistenmacher, J. D. Orbell, and L. G. Marzilli in Platinum, Gold, and Other Metal Chemotherapeutic Agents (S. J. Lippard, ed.), American Chemical Society, Washington, D.C., ACS Symp. Series 209, 1983, p. 191 ff. 95. Various articles in Nucleic Acid-Metal Ion Interactions, Vol. 1 of Metal Ions in Biology (T. G. Spiro, ed.), John Wiley and Sons, New York, 1980. 96. S. E. Sherman, D. Gibson, A. H.-J. Wang, and S. J. Lippard, J. Am. Chem. Soc., 110, 7368 (1988). 97. G. Admiraal, J. L. van der Veer, R. A. G. de Graaff, J. H. den Hartog, and J. Reedijk, J. Am. Chem. Soc., 109, 592 (1987). 98. G. Admiraal, M. Alink, C. Altona, F. J. Dijt, C. J. van Garderen,
< previous page
page_138
next page >
< previous page
page_139
next page > Page 139
R. A. G. de Graaff, and J. Reedijk, J. Am. Chem. Soc., 114, 930 (1992). 99. F. J. Pesch, H. Preut, and B. Lippert, Inorg. Chim. Acta, 169, 195 (1990). 100. M. Höpp, A. Erxleben, I. Rombeck, and B. Lippert, Inorg. Chem., in press. 101. B. Lippert, C. J. L. Lock, and R. A. Speranzini, Inorg. Chem., 20, 808 (1981). 102. B. E. Brown and C. J. L. Lock, Acta Crystallogr., C44, 611 (1988). 103. G. Fusch, E. C. Fusch, and B. Lippert, to be published. 104. D. Holthenrich, I. Sóvágó, G. Fusch, A. Erxleben, E. C. Fusch, I. Rombeck, and B. Lippert, Z. Naturforsch., 50b, in press. 105. F. J. Pesch, M. Wienken, H. Preut, A. Tenten, and B. Lippert, Inorg. Chim. Acta, 197, 243 (1992). 106. A. Schreiber, M. S. Lüth, A. Erxleben, E. C. Fusch, and B. Lippert, submitted. 107. S. Metzger, A. Erxleben, E. C. Fusch, and B. Lippert, to be published. 108. J. D. Orbell, K. Wilkowski, L. G. Marzilli, and T. J. Kistenmacher, Inorg. Chem., 21, 3478 (1982). 109. O. Krizanovic, M. Sabat, R. Beyerle-Pfnür, and B. Lippert, J. Am. Chem. Soc., 115, 5538 (1993). 110. R. Beyerle-Pfnür, B. Brown, R. Faggiani, B. Lippert, and C. J. L. Lock, Inorg. Chem., 24, 4001 (1985). 111. A. Schreiber, E. C. Hillgeris, and B. Lippert, Z. Naturforsch., 48b, 1603 (1993). 112. G. Raudaschl and B. Lippert, Inorg. Chim. Acta, 80, L49 (1983). 113. B. A. Cartwright, M. Goodgame, K. W. Johns, and A. C. Skapski, Biochem. J., 175, 337 (1978). 114. G. Raudaschl-Sieber, L. G. Marzilli, B. Lippert, and K. Shinozuka, Inorg. Chem. 24, 989 (1985). 115. O. Krizanovic, F. J. Pesch, and B. Lippert, Inorg. Chim. Acta, 165, 145 (1989). 116. I. A. G. Roos, A. J. Thomson, and J. Eagles, Chem.-Biol. Interact., 8, 421 (1974).
< previous page
page_139
next page >
< previous page
page_140
next page > Page 140
117. H. Schöllhorn, U. Thewalt, and B. Lippert, J. Chem. Soc., Chem. Commun., 769 (1984). 118. I. Dieter, B. Lippert, H. Schöllhorn, and U. Thewalt, Z. Naturforsch., 45b, 731 (1990). 119. A. Schreiber, O. Krizanovic, E. C. Fusch, B. Lippert, F. Lianza, A. Albinati, S. Hill, D. M. L. Goodgame, H. Stratemeier, and M. A. Hitchman, Inorg. Chem., 33, 6101 (1994). 120. G. Frommer, F. Lianza, A. Albinati, and B. Lippert, Inorg. Chem., 31, 2434 (1992). 121. M. Krumm, B. Lippert, L. Randaccio, and E. Zangrando, J. Am. Chem. Soc., 113, 5130 (1991). 122. M. Krumm, E. Zangrando, L. Randaccio, S. Menzer, and B. Lippert, Inorg. Chem., 32, 700 (1993). 123. C. Mealli, F. Pichierri, L. Randaccio, E. Zangrando, M. Krumm, D. Holthenrich, and B. Lippert, Inorg. Chem., 34, 3418 (1995). 124. G. Fusch, E. C. Fusch, and B. Lippert, to be published. 125. M. Krumm, E. Zangrando, L. Randaccio, S. Menzer, A. Danzmann, D. Holthenrich, and B. Lippert, Inorg. Chem., 32, 2183 (1993). 126. D. Holthenrich, M. Krumm, E. Zangrando, F. Pichierri, L. Randaccio, and B. Lippert, J. Chem. Soc., Dalton Trans., in press. 127. I. Sóvágó, A. Kiss, and B. Lippert, J. Chem. Soc., Dalton Trans., 489 (1995). 128. S. Menzer, M. Sabat, and B. Lippert, J. Am. Chem. Soc., 114, 4644 (1992). 129. G. Frommer, H. Schöllhorn, U. Thewalt, and B. Lippert, Inorg. Chem., 29, 1417 (1990). 130. H. Baba, Y. Maehara, H. Takeuchi, S. Inutsuka, M. Yamamoto, K. Endo, and K. Sugimachi, Int. J. Oncol., 4, 329 (1994). 131. S. Jaworski, S. Menzer, B. Lippert, and M. Sabat, Inorg. Chim. Acta, 205, 31 (1993). 132. H. Schöllhorn, R. Beyerle-Pfnür, U. Thewalt, and B. Lippert, J. Am. Chem. Soc., 108, 3680 (1986). 133. B. Lippert, H. Schöllhorn, and U. Thewalt, J. Am. Chem. Soc., 108, 6616 (1986). 134. F. Lianza, A. Albinati, and B. Lippert, submitted. 135. M. Höpp and B. Lippert, unpublished results.
< previous page
page_140
next page >
< previous page
page_141
next page > Page 141
136. H.-K. Choi, A. Terzis, R. C. Stevens, R. Bau, R. Haugwitz, V. L. Narayanan, and M. Wolpert-De Filippes, Biochem. Biophys. Res. Commun., 156, 1120 (1988). 137. V. Aletras, N. Hadjiliadis, A. Lymberopoulou-Karaliota, I. Rombeck, and B. Lippert, Inorg. Chim. Acta, 227, 17 (1994) and references cited. 138. V. Aletras, N. Hadjiliadis, and B. Lippert, Polyhedron, 11, 1359 (1992). 139. T. G. Appleton, F. J. Pesch, M. Wienken, S. Menzer, and B. Lippert, Inorg. Chem., 31, 4410 (1992). 140. D. Gibson, G. M. Arvanitis, and H. M. Berman, Inorg. Chim. Acta, 218, 11 (1994). 141. O. Renn, B. Lippert, and I. Mutikainen, Inorg. Chim. Acta, 218, 117 (1994). 142. T. Wienkötter, M. Sabat, and B. Lippert, to be published. 143. B. Lippert, Inorg. Chim. Acta, 56, L23 (1981). 144. R. Faggiani, C. J. L. Lock, and B. Lippert, Inorg. Chim. Acta, 106, 75 (1985). 145. T. Wienkötter, M. Sabat, and B. Lippert, Inorg. Chem., 34, 1022 (1995). 146. E. Sinn, C. M. Flynn, Jr., and R. B. Martin, Inorg. Chem., 16, 2403 (1977). 147. P.-C. Kong and F. D. Rochon, Can. J. Chem., 59, 3293 (1981). 148. M. I. Gel'fman and N. A. Kustova, Russ. J. Inorg. Chem., 14, 985 (1969). 149. M. Krumm, I. Mutikainen, and B. Lippert, Inorg. Chem., 30, 884 (1991). 150. B. Salles, J.-L. Butour, C. Lesca, and J.-P. Macquet, Biochem. Biophys. Res. Commun., 112, 555 (1983). 151. W. J. Heiger-Bernays, J. M. Essigmann, and S. J. Lippard, Biochemistry, 29, 8461 (1990). 152. J. J. Roberts and F. Friedlos, Cancer Res., 47, 31 (1987). 153. J. Reedijk, Inorg. Chim. Acta, 198200, 873 (1992).
< previous page
page_141
next page >
< previous page
page_143
next page > Page 143
6 Metal Ions in Multiple-Stranded DNA Michal Sabat1 and Bernhard Lippert2 1Department of Chemistry, University of Virginia, McCormick Road, Charlottesville, VA 22901, USA 2Fachbereich Chemie, Universität Dortmund, Otto-Hahn-Strasse 6, D-44227 Dortmund, Germany
144
1. Introduction
145
2. Structure and Stability of DNA Triplexes
3. Interactions between Metal Ions and DNA Triplexes
150
3.1. General Considerations
151
3.2. Monovalent Ions
152
3.3. Divalent Ions
152
3.4. Transition Metal Ions
154
4. Quadruplexes
154
4.1. Telomeres and Aptamers
156
4.2. Role of Metal Ions
4.3. Other Four-Stranded Helices and Superstructures
158
158
5. Four-Way DNA Junctions
160
6. Strand Crosslinking by Metal Ions
< previous page
150
page_143
next page >
< previous page
page_144
next page > Page 144 160
6.1. Interstrand Crosslinking
161
6.2. Possible Applications
7. Metal Complexes as Probes in Multiple-stranded DNA
163
163
7.1. Fe(EDTA) Complexes
165
7.2. Bis(1, 10-phenanthroline)Cu(I)
165
7.3. Osmium(VIII) Complexes
167
7.4. Other
168
8. Summary and Future Outlook
168
Acknowledgments
169
Abbreviations
169
References
1 Introduction The role of metal ions in the structure, stability, and reactivity of double-helical DNA has been well documented [1,2]. Much less is known about metal involvement in noncanonical DNA structures such as multiple-stranded species (triplexes and quadruplexes), bulges, junctions, or other branched structures. This chapter presents a survey of current data and draws some general conclusions on the interactions between metal ions and multiple-stranded DNA. We will start by characterizing some aspects of the structure and stability of the triplex DNA. The role of monovalent and divalent metal ions in intra- and intermolecular triple-stranded structures will then be discussed. Our present understanding of the role of metal ions in DNA quadruplexes and four-way junctions will be scrutinized and metal ion involvement in interstrand crosslinking will be discussed. Finally, a section of this chapter will be devoted to the metal complexes used as probes in exploring structure and reactivity of the multiple-stranded DNA.
< previous page
page_144
next page >
< previous page
page_145
next page > Page 145
2 Structure and Stability of DNA Triplexes Triple-helical DNA was discovered by Felsenfeld, Davies, and Rich [3] only a few years after the structure of duplex DNA was revealed. Originally treated as a scientific curiosity without apparent biological relevance, DNA triplexes have recently become the subject of intense research. This renewed interest was stimulated mostly by the realization that triple-stranded DNA structures may be involved in gene expression, recombination, and several other important biological processes [46]. Generally, a triplex consists of a homopurine-homopyrimidine Watson-Crick duplex (PuPy) which binds the third strand (homopyrimidine or purine-rich) through the formation of Hoogsteen-type hydrogen bonds. DNA triplexes can be divided into two major groups according to the base type of the third strand. These can be either pyrimidine·purinepyrimidine (Py·PuPy) triplexes or relatively less explored purine·purinepyrimidine (Pu·PuPy) triplestranded structures (Fig. 1). Triple-stranded DNA may occur in either intramolecular or intermolecular form. The existence of intramolecular DNA triplexes (also known as H-DNA) was postulated by Frank-Kamenetskii and coworkers [7]. They showed that upon superhelical stress, homopurine·homopyrimidine duplex DNA exhibiting mirror repeat symmetry may form a Py·PuPy triplex by folding back half of the pyrimidine strand on the PuPy double helix. Alternatively, a Pu·PuPy triplex can be formed when the polypurine strand folds back on the PuPy duplex (Fig. 2). Intermolecular triplexes play an important role in sequence-specific DNA recognition [8]. Oligonucleotides of appropriate sequence and length can bind to the major groove of duplex DNA forming local triple-helical structures. The addition of a third strand may then result in inhibition of DNA replication or gene expression [9]. The structure and stability of triple-helical DNA have been the subject of several excellent reviews [4,8,1014]. Formation of the triplexes is affected by several factors including structure and stability of the base triplets, pH and ionic strength of the solution, and type of countercations. In addition, the conformation of the duplex, the length
< previous page
page_145
next page >
< previous page
page_146
next page > Page 146
Fig. 1. Schematic representation of (a) Pu·PuPy (b) Py·PuPy triple DNA helices. (Reproduced with permission from [4].) of the oligonucleotide constituting the third strand, and, in the case of intramolecular triplexes, the topology and density of supercoiled DNA play an important role [10,12,14]. A Py·PuPy base triplet consists of the Watson-Crick PuPy (AT or GC) base pair and an additional Py base (T or C+) forming a Hoogsteen pair with the central purine base (Fig. 3a). Since the cytosine involved in the Hoogsteen pair forms two hydrogen bonds, it must be protonated. Consequently, the C+·GC triplexes are stable in acidic solutions.
< previous page
page_146
next page >
< previous page
page_147
next page > Page 147
Fig. 2. Intramolecular DNA triplex. (Reproduced with permission from [4].)
Fig. 3. Nucleobase triplets: (a) C+·GC triplet (b) G·GC triplet. Arrows indicate possible metal ion binding sites.
< previous page
page_147
next page >
< previous page
page_148
next page > Page 148
A Pu·PuPy base triplet is composed of the Watson-Crick PuPy base pair and the Hoogsteen Pu·Pu base pair in which the additional purine base is either guanine or adenine (Fig. 3b). Naturally, there exist several other triplet basepairing schemes. A good survey of their geometries has appeared in a number of recent publications [1518]. Ab initio quantum chemical calculations [19] indicate that the relative stabilities of base triplets decrease in the series:
where H and rH denote the Hoogsteen and reverse Hoogsteen arrangement of the third base. Orientation of the third strand relative to the duplex is of primary importance for the triplex formation [20]. When all nucleotides of the third strand are in the anti configuration, Hoogsteen hydrogen bonding results in a parallel orientation of the third strand with respect to the homopurine strand of the duplex. Reverse Hoogsteen base pairing imposes an antiparallel orientation. The orientations are reversed for the third strand with its nucleotides in the syn configuration. Experimental data accumulated so far indicate that oligopyrimidines bind in a parallel orientation and oligopurines bind in an antiparallel fashion relative to the homopurine strand of the target double-helical DNA [8]. Under physiological conditions, DNA in most organisms exists in a negatively supercoiled form [12]. The energy of the supercoiling may be used in intramolecular triplex (H-DNA) formation. In addition to the requirements listed previously, there are two other conditions which must be satisfied to make feasible the transition from a supercoiled DNA duplex to a triple-stranded structure. First, there should be a continuous strand of purines in the duplex. Second, the DNA fragment undergoing the transformation should contain mirror repeat symmetry. The mirror repeat is a fragment of DNA that has the same base sequence reading in both the 3' and the 5' direction (from a central point) in one strand of DNA. Four isomers labeled Hy3, Hy5, Hu5, and Hu3 are possible for the mirror repeat symmetry containing duplex folding into a triplex (Fig. 4). As we will see in the following section, metal ions play a crucial role in the formation and stability of these intramolecular triplex isomers. Several experimental techniques have been applied to study the structure and conformation of triple-helical DNA. Early X-ray fiber
< previous page
page_148
next page >
< previous page
page_149
next page > Page 149
Fig. 4. Four possible isomers of intramolecular DNA triplexes. H denotes H-DNA. y and u refer to the pyrimidine-rich and purine-rich strands, respectively, as the third strand in the triplex. The numbers 5 and 3 indicate the 5' and 3' end of the third strand that binds to the duplex. (Reproduced with permission from [4].) diffraction studies of the polynucleotide d(T)n·d(A)nd(T)n indicated that the duplex portion of the structure adopts an A-like conformation with C3'-endo sugar pucker [21]. However, a recent reinvestigation of the same polynucleotide by infrared spectroscopy and molecular modeling [22,23] suggested a B-like duplex DNA with sugars in the C2'endo conformation. The B-form model was also deduced from two-dimensional 1H nuclear magnetic resonance (NMR) studies of a 31-mer DNA oligonucleotide [24]. The fragment folds back to form a stable intramolecular triplestranded structure composed of seven triplets of the T·AT and C+·GC type. Almost all of the sugars of the fragment have puckers normally found for B-DNA. Further evidence was provided by molecular dynamics calculations on the triplex d(TC)5·d(GA)5d(C+T)5 [25] which implied that the conformation of the duplex portion is closer to that of the B-form DNA.
< previous page
page_149
next page >
< previous page
page_150
next page > Page 150
3 Interactions between Metal Ions and DNA Triplexes 3.1 General Considerations Nucleic acids exist as polyelectrolytes with negative charges associated with their phosphate groups. In doublehelical structures, repulsive forces between neighboring phosphate anions may induce unwinding of the duplexes in the absence of stabilizing counterions [2]. Compared to double-helical DNA, triple-stranded structures have a relatively higher charge density. Thus the main function of metal cations as well as other positively charged species, such as spermidine [26], may involve the stabilization of the triplex by either nonspecific (atmospheric) interactions or specific (localized) bonding to the phosphate groups. Monovalent metal ions tend to bind to the phosphate groups of nucleic acids in an atmospheric manner [27,28], although there are many known examples of more localized binding as well [29,30]. For divalent metal ions two additional types of binding are common [30,31]. In the outer-sphere binding mode a water molecule coordinated to the metal ion forms a hydrogen bond with the phosphate oxygen atom. The inner-sphere binding results in the formation of a direct coordination bond between the metal ion and the phosphate oxygen. Metal ions can also interact with the nucleic acid bases and the sugar oxygen atoms [32]. Base pairing limits the number of possible nucleobase metal binding sites in DNA duplexes. The situation is even more complex in DNA triplexes. The most common binding site, N7 of purines, usually exposed toward the major groove of the double helix, is engaged in Hoogsteen hydrogen bonding to the third strand. However, the additional purine strand in the Pu·PuPy triplexes contains N7 atoms of the purine bases quite open for metal binding. Furthermore, O6 sites of the guanine bases and exocyclic O4 sites of the pyrimidines of the third strand may potentially bind to metal ions. It is generally believed [3335] that alkali metal ions are sufficient for the stabilization of Py·PuPy triplexes. Triplestranded Pu·PuPy complexes require counterions with higher charge-to-ionic radius ratio (e.g., Mg2+, Ca2+). However, as shown in the following sections, there are several exceptions, and the function of metal ions in triplex stabilization appears to be more complicated than just the charge screening.
< previous page
page_150
next page >
< previous page
page_151
next page > Page 151
3.2 Monovalent Ions The importance of monovalent metal ions for the formation and stabilization of DNA triple helices was noted in the first report on such structures [3]. The poly(U)·poly(A)poly(U) triplex was synthesized in the presence of 100 mM NaCl. The addition of divalent ions such as Mg2+, Ca2+, and Zn2+ was found to drive the reaction to completion. Dervan et al. [36] studied the thermodynamics of formation of a triplex containing the 21-mer duplex 5'GCTAAAAAGAGAGAGAGATCG-3' and the 15-mer third strand 5'-TTTTTCTCTCTCTCT-3'. They found that a Na+ concentration of 200 mM was sufficient to bring about complete triplex formation, even in the absence of Mg2+ ions. The study indicated that the triplex is thermodynamically much less stable than its host duplex. The relationship between the energetics of triplex formation and the cationic solution environment has also been investigated in mixedvalence solutions [33,34]. Larger concentrations of monovalent ions generally result in higher stability of the triplestranded complexes. However, increasing the concentration of either Na+ or K+ ion lowers the extent of binding by multivalent cations such as Mg2+ or spermine4+, which in turn may lead to opposite effects on triplex stability as compared to those introduced by the presence of monovalent ions alone. Cheng and Van Dyke [37] investigated the influence of monovalent ions on intermolecular Pu·PuPy triple-helix formation. Several cations were found to interfere with the triplex formation. K+ and Rb+ ions were the most effective inhibitors, followed by , and Na+. Li+, and Cs+ had little or no effect on the formation. It was postulated that the observed interference could result from a competition with divalent ions required for triplex formation. Other explanations include the inhibition of a specific conformation assumed by the interacting oligonucleotides. Some monovalent ions could also stabilize other multistranded structures better than triplexes. For example, K+ ions are known to play an important structural role in guanine-containing quadruplexes (see Sec. 4). Some information on the possible location of Na+ ions in the Pu·PuPy triplexes was provided by molecular dynamics simulations of the homopolymeric G·GC 7-mer triple helix [38]. Although the phosphate groups were found to be the primary interaction site, some of the sodium ions were close to the bases as well.
< previous page
page_151
next page >
< previous page
page_152
next page > Page 152
3.3 Divalent Ions Formation of intramolecular triplexes at neutral pH requires DNA supercoiling and divalent cations [4]. The significance of Mg2+ for the intramolecular triple-stranded structures was demonstrated by Kohwi and KohwiShigematsu [39,40] who studied the sequence (dG)30·(dC)30. This sequence folded back to form an Hy3 C+·GC triplex in supercoiled DNA at low pH in the absence of Mg2+ (Fig. 5). However, a G·GC triple helix (Hu3) was formed at neutral as well as acidic pH when Mg2+ ions were present. Magnesium and calcium ions were also reported [41] to stabilize an Hy5 isomeric form of the (GAA)4TTCGC(GAA)4 inserts in certain recombinant plasmids. In the absence of the divalent ions an intramolecular triplex Hy3 was formed. Cation effects on the stability of intermolecular Pu·PuPy triplexes have been examined by Frank-Kamenetskii et al. [42] who observed that Mg2+ and Ca2+ stabilize only triplexes containing exclusively the G·GC base triplets and do not stabilize triplexes composed of both G·GC and A·AT triplets. Another alkaline earth ion, Ba2+, did not stabilize the Pu·PuPy triple helices at all. It is interesting to note that ab initio calculations [19] indicated that the G·GC triplet is significantly more stable than the C+·GC in the presence of Mg2+. The preferred sites for these ions in G·GC triplets were found to be located in the region between N7 and O6 atoms of the third-strand guanine base. 3.4 Transition Metal Ions Divalent transition metal ions have shown many interesting effects on the structure and stability of triple-helical DNA. In several instances their mode of action appears to be different than that of the divalent alkaline earth ions. One of the reasons for the different behavior could be the increased affinity for nucleobases. Azorin and coworkers [43] studied the effect of Zn2+ ions on the secondary structure of alternating d(GA·TC)n DNA fragments. At low pH, the sequences undergo a transition to intramolecular Py·PuPy triplexes. However, at neutral pH and in the presence of Zn2+ ions, a
< previous page
page_152
next page >
< previous page
page_153
next page > Page 153
Fig. 5. Intramolecular triplexes formed within poly(dG)poly(dC) sequences. (Reproduced with permission from [4].)
< previous page
page_153
next page >
< previous page
page_154
next page > Page 154
d(GA·TC)22 fragment was shown to form an intramolecular Pu·PuPy triple helix, often labeled *H-DNA. It was also found that shorter d(GA·TC)n sequences form *H-DNA with a higher difficulty. The results suggested that the first stage of the transition to *H-DNA is the formation of a denaturation bubble which, in the later steps, turns into the triple-stranded *H-DNA structure. Zn2+ ions could be responsible for local denaturation of the DNA double helix. They may also be involved in stabilization of the triplex conformation. Interestingly, the divalent metal ions used in the studies included Mg2+, Ca2+, Mn2+, Co2+, Ni2+, Cu2+, Zn2+, Cd2+, and Hg2+ but only the zinc, cadmium, and manganese ions were active in the triplex formation with their efficiency decreasing in the order Zn2+ > Cd2+ >> Mn2+. The transition metal ions Mn2+, Co2+, Ni2+, Zn2+, and Cd2+ can also stabilize intermolecular Pu·PuPy triplestranded structures of the type d(AG)n·d(GA)nd(TC)n [42]. As already indicated, Mg2+ and Ca2+ ions stabilize only the sequences which contain the G·GC triplets. In contrast, transition metal ions enhance the stability of the structures containing both the G·GC and A·AT triplets. An interesting hypothesis [44] has been put forward to explain the observed difference. According to this hypothesis, divalent metal cations can stabilize the Pu·PuPy triplexes by phosphate charge screening and enhancement of Hoogsteen hydrogen bonds. Transition metal ions with higher affinity to the purine bases are able to polarize both adenine and guanine (Fig. 6). The polarization would strengthen the Hoogsteen hydrogen bond in which the bases participate, leading to a greater overall stabilization of both G·GC and A·AT triplets. The effect of polarization by metal ions on the stability of Watson-Crick base pairs in the double-helical DNA has already been substantiated by recent theoretical calculations [45]. 4 Quadruplexes 4.1 Telomeres and Aptamers The existence of planar 6-oxopurine quartets, held together by four pairs of hydrogen bonds in a cyclic arrangement, appears to have been first proposed in the late 1950s for polyI [46] and a few years later for monomeric guanine nucleosides and nucleotides (for reviews, see
< previous page
page_154
next page >
< previous page
page_155
next page > Page 155
Fig. 6. Strengthening of the Hoogsteen-type hydrogen bond induced by metal ion binding to N7 of purines of the third strand. (Reproduced with permission from [44].) [47,48]). By use of X-ray fiber diffraction methods, this structural motif was eventually confirmed also for polyG [49,50]. Meanwhile a number of NMR solution studies on guanine-rich oligos is available, which conclusively demonstrate the existence of guanine quartets [51,52], as do two high-resolution single-crystal X-ray structures of the Oxytricha telomere repeat d(GGGGTTTTGGGG) [53] and of a parallel-stranded tetraplex formed by d(TGGGGT) [54]. Both X-ray and NMR solution studies reveal a surprising structural variability around the central guanine quartets (Fig. 7), which refers to strand and glycosidic bond orientation, (anti or syn, for a review, see [55]). Guanine quartets can form within a folded single oligonucleotide molecule [56,57] by association of four single strands (all parallel) [54,58] or by association of two DNA hairpins in antiparallel orientation [53,59]. Moreover, solid state and solution structure may differ with respect to loop topologies [60] and complicated equilibria between various forms may exist in solution. Interest in the biological significance of guanine quartets primarily stems from suggestions that the G-rich 3' single-strand overhangs of chromosome ends (''telomeres") can associate to quadruplex structures
< previous page
page_155
next page >
< previous page
page_156
next page > Page 156
Fig. 7. Hydrogen bonding pattern within a G quartet. The cation necessary for stabilization may be either in the center of the quartet or between adjacent quartets. [48,61]. Moreover, tetraplex formation has been implicated in the HIV-1 genome dimerization [62], in the recombination of certain regions of immunoglobulin genes [58], as well as in a number of other instances [60]. Finally, in vitro selected DNA oligomers sharing a highly conserved G-rich region of 1417 nucleotides, with a high affinity for the blood-clotting protein thrombin (''thrombin aptamers") [63], fold back in a way as to generate compact structures containing two guanine quartets [56,57]. 4.2 Role of Metal Ions Even at an early stage the importance of metal cations was recognized in the formation and stabilization of guanine quartets (with monomeric building blocks) as well as guanine quadruplexes (with oligonucleotides; "G4-DNA") [47]. While Na+ K+, and Rb+ salts were found to be essential for the formation of ordered G tetrads, with K+ being most effective, Li+ and Cs+ salts did not generate G tetrads. This finding clearly
< previous page
page_156
next page >
< previous page
page_157
next page > Page 157
pointed to a size effect of the cation rather than a charge effect and suggested that metal cations are integral parts of guanine quartets. Among other methods, 1H NMR was instrumental in proving this point [64]. While it has generally been accepted that the steric fit of the cation, either in the G4 plane or, more likely, in the cavity between stacked G quartets, is the sole determinant of the ion selectivity, recent free energy perturbation calculations have been interpreted in terms of an additional metal-specific electronic effect also being important [65]. Definitive evidence for the location of metal cations in G-quadruplex structures comes from the two single-crystal Xray structure analyses mentioned above [53,54]. Although the positions are not well defined due to disorder, in the Oxytricha telomere structure potassium ions are somewhat asymmetrically located between two levels of G quartets [53]. The parallel-stranded TG4T tetraplex structure has been solved at 1.2-Å resolution [54]. In this structure, sodium ions are identified as lying along the axis of the tetraplex, coordinated by eight O6 oxygens of guanines. Interestingly, aquated Ca2+ ions also present in the crystal are not directly involved in binding to the tetraplex but rather interact with the phosphate oxygens of the DNA backbone in an outer-sphere fashion via their aqua ligands. The G-quadruplex-forming effect of K+ can lead to the paradoxical situation of preventing the formation of G4-DNA in certain G-rich sequences by overstabilizing transient guanine quartets, which then do not permit formation of long G4 structures [66]. In contrast, Na+ and Rb+ do not show this effect but rather allow formation of extended G4 DNA structures. There are relatively few studies on the effects of divalent alkaline earth metal cations on G4-DNA formation [6771]. The G4-DNA stabilizing effect of these divalent cations follows the order Sr2+ > Ba2+ > Ca2+ > Mg2+ with Sr2+ being the most effective ion of all cations studied thus far [71]. Concentrations required for the formation of G tetrads in DNA oligomers containing terminal TGTG3TGTGTGTG3 sequences are in the millimolar range for Sr2+, compared to 100-fold higher concentrations in the case of alkali cations. Virtually nothing is known about the effect of other metal cations, in particular exogenous ones with a high binding preference for N7 of guanine. It is tempting to speculate on the possible effect of a cationic Pt species binding to these sensitive G regions. It is certainly feasible that quadruplex formation is seriously hampered or even prevented.
< previous page
page_157
next page >
< previous page
page_158
next page > Page 158
4.3 Other Four-Stranded Helices and Superstructures Supported by the presence of G quartets, quartet formation of bases other than guanine may be anticipated. For the RNA tetraplex (UGGGGU)4 the existence of a uracil quartet at either end of the four central G quartets has been postulated [72]. Similarly, cyclic thymine quartets have been suggested to be present in quadruplexes formed by d(G5T5) [73]. Quartet formation between pairs of A and T oligonucleotides has likewise been suggested [7477] and theoretical calculations on conformational parameters and potential energies of several feasible arrangements of (AT)4 quartets have been performed [76]. Finally, parallel-stranded guanine tetraplexes are able to form superstructures with two, three, and even four tetraplexes bonded front-to-back [78]. While a metal cation effect has been noted for the latter phenomenon, the possible role of metal cations in stabilizing, for example, (AT)4 structures appears not to have been studied as yet. 5 Four-Way DNA Junctions For processes leading to a rearrangement of DNA such as genetic recombination, a central intermediate, the so-called Holliday junction, has been postulated, in which the four strands of two recombining helices cross [79]. This junction, with the help of recombination proteins, can migrate along DNA before being cleaved to regenerate two separate DNA duplexes. According to a model proposed by Lilley and colleagues [8082], the helical arms of the fourway junction are arranged in a stacked, X-shaped structure (Fig. 8). It is important to recognize that two of the four DNA strands undergo a dramatic change of direction (bending) at the exchange point. This specific tertiary structure crucially depends on the presence of certain metal cations: The dipositive aqua cations of Mg2+, Ca2+, and Ni2+ and in particular the tripositive [Co(NH3)6]3+ stabilize this arrangement very efficiently, whereas the monovalent Na+ and K+ induce a partial folding of the junction only [80,83]. Similarly, if Mg2+ is complexed by ethylenediamine-N,N,N',N'-tetraacetic acid (EDTA) to give the anionic [Mg(EDTA)]2 species, the structure is completely different,
< previous page
page_158
next page >
< previous page
page_159
next page > Page 159
Fig. 8. Schematic representation of Holliday junction. Of the four DNA strands, two undergo sharp changes in direction and are most likely to require Mg2+ for stabilization. (Reproduced with permission from [80].) with the four arms unstacked and fully extended. These conclusions are the result of electrophoretic mobility studies of digests of Holliday junctions in the presence of a variety of cations [83]. The situation with DNA four-way junctions is thus reminiscent of that of tRNAs where Mg2+ stabilizes in particular regions where bends and turns occur. In the absence of Mg2+, both the characteristic tertiary structure and the biological activity of tRNAs are lost. From model building it is evident that negatively charged clefts, generated by phosphate oxygens near the four-way junction, call for specific metal binding as opposed to simple charge neutralization [80,83]. The high efficiency of [Co(NH3)6]3+ (effective concentration 2 µM as compared to 25 µM for spermine and 100 µM for Ca2+) points to hydrogen bonding between metal ligands (here NH3) and phosphate oxygens and/or nucleobase donor sites as being another important factor besides charge. Some insight into the possible role of hydrated Ca2+ ions in the stability of four-way junctions has been provided by a very recent crystallographic research on the B-DNA decamer CTCTCGAGAG [84]. The structure of this decamer shows a crossed arrangement of helices in the
< previous page
page_159
next page >
< previous page
page_160
next page > Page 160
crystal lattice, producing a feasible model for the Holliday junction. The interactions between the helices are facilitated by a cluster of hydrated Ca2+ ions, which bind in the minor groove on either side of a close contact with a neighboring phosphate group. 6 Strand Crosslinking by Metal Ions 6.1 Interstrand Crosslinking As pointed out in the preceding sections, cationic metal species in virtually all cases are essential for stabilizing multistranded nucleic acid structures. Metal cations discussed so far either are coordinatively saturated and inert, e.g., [Co(NH3)6]3+, or form kinetically labile adducts with nucleic acids (alkali and alkaline earth ions, Zn2+, . . .). Apart from outer-sphere phosphate binding, metal binding to donor sites of the heterocyclic part of nucleobases (e.g., O6 and N7 of guanine) appears to also take place. Superficially, the cations stabilizing G4-DNA may be considered as crosslinking strands but nevertheless should not be compared with those transition metal cations that form strong covalent bonds between bases of different strands. The earliest studies on crosslinking of two DNA strands involved Hg2+. There have been many suggestions concerning the structure of DNA-Hg2+ interstrand crosslinks. These include the ''slippage model" with selective TN3, T-N3 binding [86]; less selective binding to a variety of bases, including the amino groups of C and A [8789]; and insertion into A,T base pairs with T-O4, Λ-N6 binding [90]). Despite work that goes back to a time when the DNA double-helical structure was still unknown [85], and despite the many suggestions on how DNA-Hg2+ interstrand crosslinks might be alike on a molecular level, no conclusive picture has yet emerged. There is agreement, however, that Hg2+ binding is accompanied by loss of H+ from the DNA strands. Another metal ion, studied intensively with regard to its ability to form ordered helical polynucleotide structures with the metal ion substituting for a hydrogen bond between bases, is Ag+ [9197]. A parallel-stranded polyU double helix with hydrogen bonds replaced by covalent (U-N3)Ag(U-N3) bonds has been proposed by Shin and Eichhorn [97],
< previous page
page_160
next page >
< previous page
page_161
next page > Page 161
as has a four-stranded Ag-polyI helix containing cyclic (Ag-inosinate)4 entities. A quadruplex structure is consistent with results obtained from X-ray fiber diffraction patterns. The Ag+ cations either replace the N(1)H protons of the polyI quadruplex, and hence form intermolecular N1-Ag-O6 bonds [97], or, alternatively, crosslink N1 and N7 positions of four deprotonated inosine nucleobases [98]. Interstrand crosslinking of polynucleotides and DNA by other metal ions, e.g., Cu2+, has likewise been studied, as briefly reviewed in [1]. Interstrand crosslinking also takes place, when cis- or trans-(NH3)2-PtCl2 or their respective analogs react with DNA, both in vitro and in vivo. In the case of the cis isomer, interstrand crosslinks, presumably via two G-N(7) sites of a d(GC/CG) sequence, represent just a minor fraction (1% or less) of the total number of DNA adducts [99,100]. For the trans isomer, the spectrum of DNA adducts has not been studied to the same extent. However, it is generally assumed that interstrand crosslinks are more frequent there. These crosslinks appear to occur primarily between CN3 and G-N7 of complementary bases [101]. There is still controversy concerning the biological significance of interstrand crosslinks as opposed to the much more frequent intrastrand crosslinks for cell toxicity and antitumor activity of cis-(NH3)2PtCl2 [102,103]. 6.2 Possible Applications Sequence-specific binding of a synthetic oligonucleotide to a single-stranded DNA or a RNA molecule (''antisense technology" [104,105]) and a double-stranded DNA ("antigene technology" [106]), respectively, are novel approaches for therapeutic and/or molecular biology applications. The idea is to selectively inhibit transcription of DNA to RNA via triplex formation. Similarly, translation of RNA to protein may be inhibited via duplex formation or, more generally, a specific sequence of interest may be targeted by antisense RNA. In principle, it is possible to block a viral or oncogene this way. It has been estimated that relatively short oligos of some 1217 bases should be sufficient to uniquely define a sequence specificity within the human genome. Among the many obstacles connected with the realization of these elegant approaches
< previous page
page_161
next page >
< previous page
page_162
next page > Page 162
[8,107,108], the need for a long-lived (''permanent") link between the target sequence and the complementary synthetic oligonucleotide is of prime importance with respect to possible medical applications. Metal species capable of forming kinetically inert linkages between oligonucleotide strands might be useful in this respect. The idea of crosslinking two DNA strands by means of bifunctional bisplatinum species of type [BrPt(dien)(CH2)6(dien)Pt(H2O)]3+ appears to have been first pursued by Vlassov et al. [109]. In this experiment, it was demonstrated that an oligonucleotide, platinated with the bifunctional Pt species via one site only, was capable of platinating the complementary strand by "reaching around" the duplex. Similarly, Chu and Orgel have shown that various Pt(II) compounds are capable of crosslinking DNA sequences to complementary oligos modified with Sdonor groups (phosphorothioates [110] or cysteamine[111]). On the basis of model compounds, we have proposed that bifunctional Pt(II) species such as trans-(amine)2Pt(II) may also be capable of crosslinking two complementary oligonucleotide strands in a rational fashion [112]. Similarly, three strands may be connected, either via one metal crosslink and one set of H bonds between complementary nucleobases [113] (Fig. 9) or via two metal crosslinks [114]. There is a good chance that it is possible to crystallize even cyclic nucleobase quartets [114], in particular if interbase H bonding is favorable, such as in cyclic M4A2G2. The possible use of trans-a2Pt(II) in efficiently crosslinking oligonucleotide strands was also discussed by Leng and coworkers [115]. These authors showed that a trans-a2Pt(II) entity engaged in a 1,3-intrastrand crosslink between two guanines of a single-stranded DNA molecule undergoes linkage isomerization with formation of an interstrand adduct if the complementary strand is added. There have been other attempts to synthesize and test bifunctional bisplatinum compounds capable of crosslinking DNA either in an intra- or interstrand fashion [116118]. Although devised primarily with regard to pharmacological applications as novel antitumor agents with a spectrum of DNA adducts different from that of the antitumor agent cis(NH3)2PtCl2, some of these compounds may also be of use for antisense and/or antigene applications. Finally, it is possible to construct cis-(amine)2Pt(II) compounds that uniquely form inter- rather than intrastrand adducts [119].
< previous page
page_162
next page >
< previous page
page_163
next page > Page 163
Fig. 9. Base triplet consisting of a Watson-Crick pair between 1-methylcytosine and 9-methylguanine and a trans-(CH3NH2)2Pt(1-methylcytosine) entity joined in a Hoogsteen-like fashion. All three bases are essentially coplanar. (Reproduced with permission from [113]. Copyright 1992 American Chemical Society.) 7 Metal Complexes As Probes in Multiple-Stranded DNA 7.1 Fe(EDTA) Complexes Chemical probes containing transition metals have been widely used to study DNA structure and properties [120122]. The main function of one class of these probes, also referred to as chemical or artificial nucleases, is to cleave DNA either at some specific sites or nonspecifically at every base pair. Several chemical nucleases have been applied to elucidate the formation and properties of triple-helical DNA. Among these synthetic nucleases, of particular importance are oligonucleotides with
< previous page
page_163
next page >
< previous page
page_164
next page > Page 164
an Fe(II)-EDTA complex attached to their 5' ends. These have been extensively used by Dervan and his group [6,123] as affinity-cleaving probes. The terminal Fe(EDTA) moiety serves as a cleaving device by producing hydroxyl radicals in the presence of reducing agents and oxygen. The diffusing radicals are believed to degrade sugar residues through hydrogen abstraction at C4' atoms, thus inducing DNA backbone scission. A homopyrimidine oligonucleotide containing the Fe(EDTA) cleaving group should in principle be able to recognize the corresponding complementary sequence of duplex homopurine-homopyrimidine DNA and consequently produce a strand break at the target sequence. Fe(EDTA)-modified oligonucleotides have been used in several experiments leading to the formation of DNA triplexes. For example, a large linear DNA fragment containing 4.06 kbp was cleaved at a single homopurine site composed of 15 bp [124]. The size of the pyrimidine oligonucleotides used as probes varied from 11 to 15 bp. These probes were found to bind in the major groove parallel to the purine strand of duplex DNA, forming triple-helical Py·PuPy structures. In another experiment, an even larger (48 kbp) DNA fragment of the bacteriophage λ genome was cleaved at a single 18-bp homopurine site [125]. Cleavage can also be achieved by Fe(EDTA) attached to a polypurine strand in a Pu·PuPy triple helix [20]. The affinity cleavage experiments have been helpful in determination of factors affecting the triplex stability. Thus the influence of pH, organic solvent, added cations, temperature, as well as the effect of the length and sequence of oligonucleotides, have been evaluated by using various oligonucleotide-Fe(EDTA) probes [124]. It is also of interest to note that the optimum cleaving efficiencies for these probes were observed in the presence of millimolar concentrations of the [Co(NH3)6]3+ complex or spermine. An interesting application of the [Fe(EDTA)]2 complex in the study of the structure of four-way Holliday junctions has been reported [126]. The complex generates hydroxyl radicals in the presence of hydrogen peroxide, following the general pattern of the Fenton reaction [127]. Oligonucleotide analogs of the Holliday junction, called immobile DNA junctions, have been used in this experiment. One such junction, J1, composed of four 16-mers, has been tested for accessibility of its strands to hydroxyl radicals. The J1 cleavage pattern clearly showed that the junction possesses an overall twofold symmetry. The results
< previous page
page_164
next page >
< previous page
page_165
next page > Page 165
also suggested that the four arms of the junction are involved in extensive stacking within two helical domains. The Fe-EDTA complex was also recently used in characterization of DNA mesojunctions, i.e., DNA assemblies containing a combination of helices distributed both radially and circumferentially relative to central points of the junctions [128]. 7.2 Bis(1,10-phenanthroline)Cu(I) Bis(1,10-phenanthroline)Cu(I), introduced by Sigman [129] and used extensively in nucleic acid research [130,131], has also been tested as a probe of triplex DNA structure and conformation. The mechanism of DNA backbone scission induced by this complex involves several steps. First, the bis(1,10-phenanthroline)Cu(II) complex generated in situ is reduced to the active cuprous form by addition of thiols or ascorbic acid. The Cu(I) complex binds in a nonintercalative fashion to the minor groove of DNA. In the presence of hydrogen peroxide, a highly reactive complex with a hydroxyl radical coordinated to the Cu(II) atom is generated. This species attacks H1' and, to a lesser extent, H4' protons of the deoxyribose residues, leading to breaks in the DNA backbone. The bis(1,10-phenanthroline)Cu(I) complex has been applied by Hélène et al. [132] in the study of binding of 11-mer pyrimidine oligonucleotides to a 32-mer duplex. A number of conclusions have been reached. First, the probe does not bind to triple-helical DNA. Second, formation of the triple helix results in a conformational change on the 3' side of the bound oligonucleotide, as indicated by an enhanced efficiency on the purine strand. DNA footprinting studies have also provided some structural details of the junctions between the triple helix and the duplex structure. 7.3 Osmium(VIII) Complexes A variety of osmium(VIII) compounds have been utilized as versatile probes of DNA structure both in vitro and in vivo by Palecek and his group [12,133]. Osmium(VIII) tetroxide oxidizes thymine and, to much
< previous page
page_165
next page >
< previous page
page_166
next page > Page 166
lesser extent, cytosine and uracil to 5,6-cis-diols. Double-helical B-form DNA is quite inert toward the oxidation, as the stacking of nucleobases partially shields the 5,6 double bond of thymines. However, single-stranded DNA is oxidized easily. Furthermore, local changes in the double-helix geometry may expose the 5,6 double bond, making it accessible to the attack by the reagent. Thus, base unstacking, single-base mismatches and bulges, helix distortions in the vicinity of single-strand interruptions, and several other structural changes can be detected by using the osmium probes [12]. Addition of osmium(VIII) tetroxide across the 5,6 double bond results in the formation of an osmate ester, which can be stabilized by auxiliary ligands such as pyridine (py) or 2,2'-bipyridyl (bipy). Consequently, in several experiments the complexes OsO4(py)2 and OsO4(bipy) have been used instead of osmium(VIII) tetroxide. A recent report [134] elucidated some structural details of the OsO4(py)2-DNA complexes by applying molecular mechanics techniques. The detection of osmium binding sites can be achieved in a number of ways. These sites are recognized and cleaved by nuclease S1 [135]. They can also be detected at single-nucleotide resolution either by hot piperidine cleavage [136] or by utilizing the ability of the osmate esters to terminate transcription in vitro [137]. Numerous applications of Os(VIII) compounds have been characterized extensively in an excellent review by Palecek [12]. Here we analyze only some recent work on triple-helical, four-way junction and nodular DNA. Htun and Dahlberg [137] applied osmium probes in a study of the formation of intramolecular H-DNA triplexes in some naturally occurring sequences. The stability of these structures was affected by pH and the degree of negative supercoiling. Furthermore, at mildly alkaline pH, the sequences rearranged into a novel conformation, called J-DNA, which was different from both the B and H forms. OsO4 was used to examine the reactivity and the conditions of triplex formation in a restriction fragment from the human U1 gene [138]. Models of a triple-helical sequence at pH 5 and high superhelical density under the conditions of the OsO4 reaction have been developed. A (CT)16 fragment from the sea urchin histone gene has been probed using OsO4(bipy) in vivo in Escherichia coli by Palecek et al. [139]. The osmium binding pattern suggested the existence of triple-
< previous page
page_166
next page >
< previous page
page_167
next page > Page 167
helical structures, occurring mostly as the Hy5 isomer. Triplex-forming sequence (GA)7TA(GA)7 was also tested with the same osmium complex in living E. coli cells [140], indicating the presence of the Hy3 isomer. The osmium probes have been instrumental in determining the effects of metal ions on the conformation of four-way DNA junctions [141]. Metal ions were found to fold the junctions into compact conformations which protect all thymine bases against modifications by OsO4. Monovalent ions are only partially effective in this folding. OsO4 can easily modify junction thymine bases in the presence of 50 mM Na+. However, addition of 1 mM Mg2+ protects thymines from the reaction with OsO4 indicating that this alkaline earth ion completes the folding process. Panyutin and Wells [142] used OsO4 in the study of stabilization of nodular DNA by di- and trivalent metal ions. Formation of the DNA nodule has been observed for an alternating d(GA)nd(CT)n sequence. The nodule was composed of two triple-stranded stems connected by two single strands. A relevant model of this nodule was recently created by using molecular modeling techniques [143]. The close proximity of two triplexes requires metal ions for neutralization of the phosphate charges. It was found that the [Co(NH3)6]3+ complex cation was particularly effective in the stabilization of this interesting structure. 7.4 Other Nielsen [144] applied uranyl-mediated DNA photocleavage to test the accessibility of the phosphate groups in two plasmid structures involving 15-mer triple-helical fragments. The uranyl(VI) ion, , binds to the phosphate groups of the DNA backbone. The photocleavage induced by long-wavelength UV radiation is thought to occur through oxidation of the sugar moieties in the vicinity of the uranyl binding sites [121]. The results of the uranyl probing show that the third strand of the triple helix appears to be asymmetrically located in the major groove of the duplex with the strongest interactions between the strands occurring at the 5' end of the third strand. The accessibility of the phosphate groups in the triplexes containing both T·AT and C+·GC triplets was found to be altered, indicating a structural change in the backbone.
< previous page
page_167
next page >
< previous page
page_168
next page > Page 168
8 Summary and Future Outlook Metal ions constitute an integral part of noncanonical DNA conformations. Triple-helical structures involved in such biological events as gene expression and recombination require mono- or divalent ions for their stability. The presence of some of these metal ions is essential for telomeric DNA quadruplexes [145]. Transition metal ions are also known to stabilize many nonstandard DNA assemblies. Much remains to be established. First, more structural details provided by X-ray crystallography and NMR spectroscopy are needed to better understand the mechanisms governing stabilization. Several recently discovered cellular processes may involve various metal ionmultiple-stranded DNA interactions. For instance, some proteins bind preferentially to triplex DNA [146]. Virtually nothing is known about the significance of metal ions in proteintriplex DNA recognition. Interactions between duplex DNA and RNA polynucleotides are important in the design of artificial gene repressors [147]. Again, little is known about the role of metal ions in these processes. Metal species capable of forming kinetically inert linkages between oligonucleotide strands may have important applications in antisense and antigene technologies. Some models of these linkages have already been created. However, more structural information is required in order to understand the conformational changes induced by the metal entities. Exogenous metals such as chromium and platinum may bind to sensitive regulatory regions or ends of chromosomes. These processes could be of great importance for our understanding of metalinduced mutagenesis. Finally, the influence of metal ions on the conformation of four-way junctions has been characterized [141,148]. Metal ions are believed now to be an indispensable component of these systems, which are central intermediates in genetic recombination. Acknowledgments We thank Professors Cindy Klevickis, James Madison University, for many helpful comments and discussions and Richard R. Sinden, Texas A&M University, for providing the figures.
< previous page
page_168
next page >
< previous page
page_169
next page > Page 169
Abbreviations
A
adenine
A+
adenine protonated at N1
bipy
2,2'-bipyridyl
bp
base pair
C
cytosine
C+
cytosine protonated at N3
d
deoxy
dien
diethylenetriamine
EDTA
ethylenediamine-N,N,N',N'-tetraacetic acid
G
guanine
HIV
human immunodeficiency virus
NMR
nuclear magnetic resonance
Pu
purine base
py
pyridine
Py
pyrimidine base
T
thymine
U
uracil
References 1. G. L. Eichhorn, Adv. Inorg. Biochem., 3, 1 (1981). 2. L. G. Marzilli, T. J. Kistenmacher, and G. L. Eichhorn in Nucleic AcidMetal Ion Interactions (T. G. Spiro, ed.), John Wiley and Sons, New York, 1980, p. 179 ff. 3. G. Felsenfeld, D. R. Davies, and A. Rich, J. Am. Chem. Soc., 79, 2023 (1957).
4. R. R. Sinden, DNA Structure and Function, Academic Press, San Diego, 1994. 5. R. D. Wells, J. Biol. Chem., 263, 1095 (1988). 6. P. B. Dervan in Oligodeoxynucleotides. Antisense Inhibitors of Gene Expression (J. S. Cohen, ed.), CRC Press, Boca Raton, 1989, p. 197 ff. 7. S. M. Mirkin, V. I. Lyamichev, K. N. Drushlyak, V. N. Dobrynin, S. A. Filippov, and M. D. Frank-Kamenetskii, Nature, 330, 495 (1987).
< previous page
page_169
next page >
< previous page
page_170
next page > Page 170
8. N. T. Thuong and C. Hélène, Angew. Chem., 32, 666 (1993). 9. C. Hélène, Pure Appl. Chem., 66, 663 (1994). 10. H. Htun and J. E. Dahlberg in Structure and Methods, Vol. 3, DNA and RNA (R. H. Sarma and M. H. Sarma, eds.), Adenine Press, Schenectady, NY, 1990, p. 185 ff. 11. Y.-K. Cheng and B. M. Pettitt, Prog. Biophys. Mol. Biol., 56, 225 (1992). 12. E. Palecek, Crit. Rev. Biochem. Mol. Biol., 26, 151 (1991). 13. G. Yagil, Crit. Rev. Biochem. Mol. Biol., 26, 475 (1991). 14. H. Htun and J. E. Dahlberg, Science, 243, 1571 (1989). 15. H. W. T. Van Vlijmen, G. L. Rame, and B. M. Pettitt, Biopolymers, 30, 517 (1990). 16. Y.-K. Cheng and B. M. Pettitt, J. Am. Chem. Soc., 114, 4465 (1992). 17. C. A. Laughton and S. Neidle, Nucl. Acids Res., 20, 6535 (1992). 18. J. M. Piriou, C. Ketterle, J. Gabarro-Arpa, J. A. H. Cognet, and M. Le Bret, Biophys. Chem., 50, 323 (1994). 19. S.-P. Jiang, R. L. Jernigan, K.-L. Ting, J.-L. Syi, and G. Raghunathan, J. Biomol. Struct. Dynam., 12, 383 (1994). 20. P. A. Beal and P. B. Dervan, Science, 251, 1360 (1991). 21. S. Arnott and E. Selsing, J. Mol. Biol., 88, 509 (1974). 22. F. B. Howard, H. T. Miles, K. Liu, J. Frazier, G. Raghunathan, and V. Sasisekharan, Biochemistry, 31, 10671 (1992). 23. G. Raghunathan, H. T. Miles, and V. Sasisekharan, Biochemistry, 32, 455 (1993). 24. R. Macaya, E. Wang, P. Schultze, V. Sklenar, and J. Feigon, J. Mol. Biol., 225, 755 (1992). 25. C. A. Laughton and S. Neidle, J. Mol. Biol., 223, 519 (1992). 26. T. Thomas and T. J. Thomas, Biochemistry, 32, 14068 (1993). 27. M. T. Record, Jr., S. J. Mazur, P. Melancon, J.-H. Roe, S. L. Shaner, and L. Unger, Ann. Rev. Biochem., 50, 997 (1981). 28. D. M. York, T. Darden, D. Deerfield, II, and L. G. Pedersen, Int. J. Quantum Chem: Quantum Biol. Symp., 19, 145 (1992). 29. V. Swaminathan and M. Sundaralingam, CRC Crit. Rev. Biochem., 6, 245 (1979).
< previous page
page_170
next page >
< previous page
page_171
next page > Page 171
30. C. B. Black, H.-W. Huang, and J. A. Cowan, Coord. Chem. Rev., 135/136, 165 (1994). 31. S. A. Kazakov and S. M. Hecht in Encyclopedia of Inorganic Chemistry (R. B. King, ed.), John Wiley and Sons, New York, 1994, p. 2697 ff. 32. R. B. Martin and Y. H. Mariam in Metal Ions in Biological Systems, Vol. 8 (H. Sigel, ed.), Marcel Dekker, New York, p. 57 ff. 33. M. Rougee, B. Faucon, J. L. Mergny, F. Barcelo, C. Giovannangeli, T. Garestier, and C. Hélène, Biochemistry, 31, 9269 (1992). 34. S. F. Singleton and P. B. Dervan, Biochemistry, 32, 13171 (1993). 35. J. Voelker and H. H. Klump, Biochemistry, 33, 13502 (1994). 36. G. E. Plum, Y.-W. Park, S. F. Singleton, P. B. Dervan, and K. J. Breslauer, Proc. Natl. Acad. Sci. USA, 87, 9436 (1990). 37. A.-J. Cheng and M. W. Van Dyke, Nucl. Acids Res., 21, 5630 (1993). 38. V. Mohan, P. E. Smith, and B. M. Pettitt, J. Phys. Chem., 97, 12984 (1993). 39. Y. Kohwi and T. Kohwi-Shigematsu, Proc. Natl. Acad. Sci. USA, 85, 3781 (1988). 40. T. Kohwi-Shigematsu and Y. Kohwi in Structure and Methods, Vol. 3, DNA and RNA (R. H. Sarma and M. H. Sarma, eds.), Adenine Press, Schenectady, New York, 1990, p. 225 ff. 41. S. Kang, F. Wohlrab, and R. D. Wells, J. Biol. Chem., 267, 1259 (1992). 42. V. A. Malkov, O. N. Voloshin, V. N. Soyfer, and M. D. Frank-Kamenetskii, Nucl. Acids Res., 21, 585 (1993). 43. J. Bernues, R. Beltran, J. M. Casasnovas, and F. Azorin, Nucl. Acids. Res., 18, 4067 (1990). 44. V. N. Potaman and V. N. Soyfer, J. Biomol. Struct. Dynam., 11, 1035 (1994). 45. E. H. S. Anwander, M. M. Probst, and B. M. Rode, Biopolymers, 29, 757 (1990). 46. A. Rich, Biochim. Biophys. Acta, 29, 502 (1958). 47. W. Saenger in Principles of Nucleic Acid Structure, Springer-Verlag, New York, 1984, pp. 315 ff. 48. W. Guschlbauer, J.-F. Chantot, and D. Thiele, J. Biomol. Struct. Dynam., 8, 491 (1990).
< previous page
page_171
next page >
< previous page
page_172
next page > Page 172
49. S. Arnott, R. Chandrasekaran, and C. M. Marttila, Biochem. J., 141, 537 (1974). 50. S. B. Zimmermann, G. H. Cohen, and D. R. Davies, J. Mol. Biol., 92, 181 (1975). 51. F. Aboul-ela, A. I. H. Murchie, and D. M. J. Lilley, Nature, 360, 280 (1992) and references cited. 52. F. W. Smith and J. Feigon, Biochemistry, 32, 8682 (1993) and references cited. 53. C. Kang, X. Zhang, R. Moyzis, and A. Rich, Nature, 356, 126 (1992). 54. G. Laughlan, A. I. H. Murchie, D. G. Norman, M. H. Moore, P. C. E. Moody, D. M. J. Lilley, and B. Luisi, Science, 265, 520 (1994). 55. W. I. Sundquist in Nucleic Acids and Molecular Biology, Vol. 5 (F. Eckstein and D. M. J. Lilley, eds.), SpringerVerlag, Berlin, 1991, p. 1 ff. 56. R. F. Macaya, P. Schultze, F. W. Smith, J. A. Roe, and J. Feigon, Proc. Natl. Acad. Sci. USA, 90, 3745 (1993). 57. K. Y. Wang, S. H. Krawczyk, N. Bischofberger, S. Swaminathan, and P. H. Bolton, Biochemistry, 32, 11285 (1993). 58. D. Sen and W. Gilbert, Nature, 334, 364 (1988). 59. Y. Oka and C. A. Thomas, Nucl. Acids Res., 15, 8877 (1987). 60. J. Feigon, F. W. Smith, R. F. Macaya, and P. Schultze in Structural Biology: The State of the Art (R. H. Sarma and M. H. Sarma, eds.), Adenine Press, Schenectady, NY, 1994 p. 124 ff. 61. E. Blackburn, Nature, 350, 569 (1991). 62. W. I. Sundquist and S. Heaphy, Proc. Natl. Acad. Sci. USA, 90, 3393 (1993). 63. L. C. Bock, L. C. Griffin, J. A. Latham, E. H. Vermaas, and J. J. Toole, Nature, 355, 564 (1992). 64. T. J. Pinnavaia, C. L. Marshall, C. M. Mettler, C. L. Fisk, H. T. Miles, and E. D. Becker, J. Am. Chem. Soc., 100, 3625 (1978). 65. W. S. Ross and C. C. Hardin, J. Am. Chem. Soc., 116, 6070 (1994). 66. D. Sen and W. Gilbert, Nature, 334, 410 (1990). 67. F. J. Chantot and W. Guschlbauer, FEBS Lett., 4, 173 (1969). 68. J. S. Lee, Nucl. Acids. Res., 18, 6057 (1991).
< previous page
page_172
next page >
< previous page
page_173
next page > Page 173
69. C. C. Hardin, T. Watson, M. Corregan, and C. Bailey, Biochemistry, 31, 833 (1992). 70. F.-M. Chen, Biochemistry, 31, 3769 (1992). 71. E. A. Venczel and D. Sen, Biochemistry, 32, 6220 (1993). 72. C. Cheong and P. B. Moore, Biochemistry, 31, 8406 (1992). 73. M. H. Sarma, J. Luo, K. Umemoto, R-d. Yuan, and R. H. Sarma, J. Biomol. Struct. Dynam., 9, 1131 (1992). 74. S. McGavin, H. R. Wilson, and G. G. Barr, J. Mol. Biol., 22, 187 (1966). 75. S. McGavin, J. Mol. Biol., 55, 293 (1971). 76. A. A. Chernyi, Yu. P. Lysov, I. A. Il'icheva, A. S. Zibrov, A. K. Shchyolkina, O. F. Borisova, O. K. Mamaeva, and V. L. Florentiev, J. Biomol. Struct. Dynam., 8, 513 (1990). 77. O. F. Borisova, Yu. B. Golova, B. P. Gottikh, A. S. Zibrov, I. A. Il'icheva, Yu. P. Lysov, O. K. Mamayeva, B. K. Chernov, A. A. Chernyi, A. K. Shchyolkina, and V. L. Florentiev, J. Biomol. Struct. Dynam., 8, 1187 (1991). 78. D. Sen and W. Gilbert, Biochemistry, 31, 65 (1992). 79. R. Holliday, Genet. Res., 5, 282 (1964). 80. D. M. J. Lilley in Nucleic Acids and Molecular Biology, Vol. 4 (F. Eckstein and D. M. J. Lilley, eds.), SpringerVerlag, Berlin, 1990, p. 55 ff. 81. E. von Kitzing, D. M. J. Lilley, and S. Diekman, Nucl. Acids Res., 18, 2671 (1990). 82. R. M. Clegg, A. I. H. Murchie, A. Zechel, C. Carlberg, S. Diekman, and D. M. J. Lilley, Biochemistry, 31, 4846 (1992). 83. D. R. Duckett, A. I. H. Murchie, S. Diekman, E. von Kitzing, B. Kemper, and D. M. J. Lilley, Cell, 55, 79 (1988). 84. D. S. Goodsell, K. Grzeskowiak, and R. E. Dickerson, Biochemistry, 34, 1022 (1995). 85. S. Katz, J. Am. Chem. Soc., 74, 2238 (1952). 86. S. Katz, Biochim. Biophys. Acta, 68, 240 (1963). 87. G. L. Eichhorn and P. Clark, J. Am. Chem. Soc., 85, 4020 (1963). 88. R. B. Simpson, J. Am. Chem. Soc., 86, 2059 (1964). 89. U. S. Nandi, J. C. Wang, and N. Davidson, Biochemistry, 4, 1687 (1965).
< previous page
page_173
next page >
< previous page
page_174
next page > Page 174
90. N. Å. Frøystein and E. Sletten, J. Am. Chem. Soc., 116, 3240 (1994). 91. T. Yamane and N. Davidson, Biochim. Biophys. Acta, 55, 609 (1962). 92. T. Yamane and N. Davidson, Biochim. Biophys. Acta, 55, 780 (1962). 93. R. H. Jensen and N. Davidson, Biopolymers, 4, 17 (1966). 94. M. Daune, C. A. Dekker, and H. K. Schachman, Biopolymers, 4, 51 (1966). 95. G. L. Eichhorn, J. J. Butzow, P. Clark, and E. Tarien, Biopolymers, 5, 283 (1967). 96. S. K. Arya and J. T. Yang, Biopolymers, 14, 1847 (1975). 97. Y. A. Shin and G. L. Eichhorn, Biopolymers, 19, 539 (1980). 98. F. Bélanger-Gariépy and A. L. Beauchamp, J. Am. Chem. Soc., 102, 3461 (1980). 99. J. J. Roberts and F. Friedlos, Biochim. Biophys. Acta, 655, 146 (1981). 100. A. Eastman, Biochemistry, 24, 5027 (1985). 101. V. Brabec, M. Sip, and M. Leng, Biochemistry, 32, 11676 (1993). 102. R. B. Ciccarelli, M. J. Solomon, A. Varshavsky, and S. J. Lippard, Biochemistry, 24, 7533 (1985). 103. J. J. Roberts and F. Friedlos, Cancer Res., 47, 31 (1987). 104. P. C. Zamecnik and M. L. Stephenson, Proc. Natl. Acad. Sci. USA, 75, 280 (1978). 105. M. L. Stephenson and P. C. Zamecnik, Proc. Natl. Acad. Sci. USA, 75, 285 (1978). 106. C. Hélène, Anti-Cancer Drug Design, 6, 569 (1991). 107. E. Uhlmann and A. Peyman, Chem. Rev., 90, 543 (1990). 108. C. A. Stein and Y.-C. Cheng, Science, 261, 1004 (1993). 109. V. V. Vlassov, V. V. Gorn, E. M. Ivanova, S. A. Kazakov, and S. V. Mamaev, FEBS Lett. 162, 286 (1983). 110. B. C. F. Chu and L. E. Orgel, Nucl. Acids Res., 18, 5163 (1990). 111. B. C. F. Chu and L. E. Orgel, Nucl. Acids Res., 17, 4783 (1989). 112. O. Krizanovic, M. Sabat, R. Beyerle-Pfnür, and B. Lippert, J. Am. Chem. Soc., 115, 5538 (1993) and references cited.
< previous page
page_174
next page >
< previous page
page_175
next page > Page 175
113. I. Dieter-Wurm, M. Sabat, and B. Lippert, J. Am. Chem. Soc., 114, 357 (1992). 114. A. Schreiber, E. C. Hillgeris, and B. Lippert, Z. Naturforsch., 48b, 1603 (1993). 115. R. Dalbiès, D. Payet, and M. Leng, Proc. Natl. Acad. Sci. USA, 91, 8147 (1994). 116. N. Farrell, Y. Qu, and M. P. Hacker, J. Med. Chem., 33, 2179 (1990). 117. R. Alul, M. B. Cleaver, and J.-S. Taylor, Inorg. Chem., 31, 3636 (1992). 118. J. Kozelka, E. Segal, and C. Bois, J. Inorg. Biochem., 47, 67 (1992). 119. E. C. H. Ling, G. W. Allen, and T. W. Hambley, J. Am. Chem. Soc., 116, 2673 (1994). 120. A. M. Pyle and J. K. Barton, Progr. Inorg. Chem., 38, 413 (1990). 121. P. E. Nielsen, J. Mol. Recogn., 3, 1 (1990). 122. T. D. Tullius in Metal-DNA Chemistry, ACS Symposium Series 402 (T. D. Tullius, ed.), American Chemical Society, Washington, D.C., 1989, p. 1 ff. 123. S. A. Strobel and P. B. Dervan in Methods in Enzymology, Vol. 216 (R. Wu, ed.). Academic Press, San Diego, 1992, p. 309 ff. 124. H. E. Moser and P. B. Dervan, Science, 238, 645 (1987). 125. S. A. Strobel, H. E. Moser, and P. B. Dervan, J. Am. Chem. Soc., 110, 7927 (1988). 126. M. E. A. Churchill, T. D. Tullius, N. R. Kallenbach, and N. C. Seeman, Proc. Natl. Acad. Sci. USA, 85, 4653 (1988). 127. C. Walling, Acc. Chem. Res., 8, 125 (1975). 128. H. Wang and N. C. Seeman, Biochemistry, 34, 920 (1995). 129. D. S. Sigman, D. R. Graham, V. D'Aurora, and A. M. Stern, J. Biol. Chem., 254, 12269 (1979). 130. D. S. Sigman, A. Mazumder, and D. M. Perrin, Chem. Rev., 93, 2295 (1993). 131. A. G. Papavassiliou, Biochem. J., 305, 345 (1995). 132. J.-C. Francois, T. Saison-Behmoaras, and C. Hélène, Nucl. Acids Res., 16, 11431 (1988). 133. E. Palecek, P. Boublikova, F. Jelen, A. Krejcova, E. Makaturova, K. Nejedly, P. Pecinka, and M. Vojtiskova in Structure and
< previous page
page_175
next page >
< previous page
page_176
next page > Page 176
Methods, Vol. 3, DNA and RNA (R. H. Sarma and M. H. Sarma, eds.), Adenine Press, Schenectady, NY, 1990, p. 237 ff. 134. P. Mejzlik, J. Biomol. Struct. Dynam., 12, 327 (1994). 135. G. C. Glikin, M. Vojtiskova, L. Rena-Descalzi, and E. Palecek, Nucl. Acids Res., 12, 1725 (1984). 136. G. Galazka, E. Palecek, R. D. Wells, and J. Klysik, J. Biol. Chem., 261, 7093 (1986). 137. H. Htun and J. E. Dahlberg, Science, 241, 1791 (1988). 138. B. H. Johnston, Science, 241, 1800 (1988). 139. P. Karlovsky, P. Pecinka, M. Vojtiskova, E. Makaturova, and E. Palecek, FEBS Lett., 274, 39 (1990). 140. D. W. Ussery and R. R. Sinden, Biochemistry, 32, 6206 (1993). 141. D. R. Duckett, A. I. H. Murchie, and D. M. J. Lilley, EMBO J., 9, 583 (1990). 142. I. G. Panyutin and R. D. Wells, J. Biol. Chem., 267, 5495 (1992). 143. D. Sprous and S. C. Harvey, J. Biol. Chem., 267, 5502 (1992). 144. P. E. Nielsen, Nucl. Acids Res., 20, 2735 (1992). 145. J. R. Williamson, M. K. Raghuraman, and T. R. Cech, Cell, 59, 871 (1989). 146. R. Kiyama and R. D. Camerini-Otero, Proc. Natl. Acad. Sci. USA, 88, 10450 (1991). 147. C. D. McDonald and L. J. Maher III, Nucl. Acids Res., 23, 500 (1995). 148. D. M. J. Lilley and R. M. Clegg, Quart. Rev. Biophys., 26, 131 (1993).
< previous page
page_176
next page >
< previous page
page_177
next page > Page 177
7 DNA Interactions with Substitution-Inert Transition Metal Ion Complexes Bengt Nordén, Per Lincoln, Björn Åkerman, and Eimer Tuite Department of Physical Chemistry, Chalmers University of Technology, S-41296 Göteborg, Sweden
179
1. Introduction
2. Tools to Characterize Transition Metal Interactions with DNA
181
2.1. Optical Spectroscopy
2.1.1. Steady-State Isotropic Absorption and Emission
2.1.2. Steady-State Polarized Absorption and Emission
2.1.3. Transient Absorption and Time-Resolved Emission
181
188
190
190
2.1.4. Infrared and Resonance Raman
191
2.2. NMR Spectroscopy
191
2.3. Electrochemical Methods
192
2.4. Separation Methods
192
2.5. Binding Curves
195
2.6. Kinetics
< previous page
181
page_177
next page >
< previous page
page_178
next page > Page 178
2.7. Thermodynamic Methods
2.7.1. Salt Effects on Binding Constants
2.7.2. Enthalpic and Entropic Contributions to Binding Constants
2.8. Hydrodynamic Methods
2.8.1. Lengthening of the DNA Helix
2.8.2. Unwinding of Supercoiled DNA
2.8.3. Conductivity and Electrophoresis
2.9. X-Ray Diffraction
3. Interactions of Transition Metal Complexes with DNA
3.1. Binding and Photophysical Properties of Bound Transition Metal Complexes
3.1.1. Enantioselectivity
3.1.2. Influence of Metal, Charge, and Ligands on Binding Affinity
3.1.3. Intercalation vs. Groove Binding
3.1.4. Sequence Selectivity and Sequence-Related Enantioselectivity
3.2. Interactions with Z-DNA and Other Unusual DNA Conformations
3.3. Sequence-Specific Binding: Oligonucleotide-Tethered Complexes
3.4. Photoinduced Electron Transfer Reactions in the Presence of DNA
4. Diastereomeric Binding Geometries Studied with Polarized Spectroscopy
195
195
197
198
198
200
202
204
204
204
205
209
220
225
227
230
231
232
4.1. Polarized Spectroscopy of Ruthenium Complexes
233
4.2. Spectroscopy of Ru-phen Complexes
4.3. Binding Geometries of [Ru(phen)2(DPPZ)]2+ and [Ru(phen)2(BDPPZ)]2+
234
235
4.4. Roll Angle
239
4.5. Binding Geometry of [Ru(phen)3]2+
240
5. Concluding Remarks
241
Abbreviations
242
References
< previous page
233
page_178
next page >
< previous page
page_179
next page > Page 179
1 Introduction Transition metal coordination and organometallic complexes* can interact with nucleic acids by intercalation, groove binding, or external electrostatic binding (all ''outer-sphere" interactions) as well as by coordination of the metal directly to DNA [13]. The design and construction of small complexes with polypyridyl ligands for use as structural probes or artificial nucleases has been an active area of research during the last 10 years or so [1,2,4,5]. It can easily be envisaged how metal complexes might be designed to interact with DNA in defined ways, given the richness and breadth of coordination chemistry. For instance, by coordination of a large extended aromatic ligand it should be possible to construct a metal complex which can, at least partially, intercalate and thereby bring the central metal atom into a position very close to the bases or the backbone without having any physical contact with them. This clearly presents many possibilities whereby the metal can be externally activated to react with the DNA after it is bound, perhaps to covalently crosslink at a prescribed site or to damage the DNA with a view to achieving specific cleavage of the backbone. Furthermore, suitable design of the ancillary ligands could allow modifications to strengthen binding, to induce sequence- or conformation-specific binding, or to attach an activating group. A particular attraction of metal complexes is their three-dimensional nature, which makes them excellent candidates for spatial probes of DNA; design of organic molecules with comparable structural properties would present serious synthetic difficulties. The most intriguing possibility has probably been that of stereoselective/stereospecific binding for chiral metal complexes. This was first noted for binding of the [Fe(bpy)3]2+ complex to B-DNA by Nordén and Tjerneld [6]. Subsequently, others have investigated stereoselective binding of *The terms ligand and complex are used in different ways by coordination chemists and biophysical chemists. In this chapter we use both terms in the inorganic chemistry sense, i.e., a ligand is a group that is coordinated to the metal ion and a complex is the metal ion and ligand ensemble (coordination complex). Thus phenanthroline is a ligand in the [Ru(phen)3]2+ complex. The terms dye and drug are used to describe compounds that bind to DNA.
< previous page
page_179
next page >
< previous page
page_180
next page > Page 180
metal complexes with different structural motifs [5] and such interactions represent one of the main themes of this chapter. To date, the design possibility has remained largely unexploited in studies of small metal complexes with DNA. However, as we begin to gain a better understanding of which features are required of a complex to result in a desired mode of binding, it is sure to become a more important concept. Hence, the eventual design of metal complexes which act like sequence-specific proteins, recognizing the structural texture of the DNA as well as the electrostatic potential and H-bonding patterns in the grooves, may be achieved. In addition to changing the DNA binding properties, the chemical, photophysical, and photochemical properties of metal complexes can be radically altered by modification of ligands or by changing of the metal ion (or even just its oxidation state) [79]. Fortunately, there already exists an extensive literature dealing with the physiochemical effects of modifications, thus facilitating the design of modifications to achieve the properties required of the metal complex (so-called tuning of the properties) [8]. Indeed, before metal complexes were considered as DNA probes, there was great interest in the behavior of these complexes in microheterogeneous media (micelles, vesicles, polyelectrolytes, zeolites, clays, glasses, etc.) [10]. Much of this work was driven by attempts to find suitable media to facilitate harnessing of solar energy, but it was also found that coordination complexes could be used to probe the nature of microenvironments (e.g., hydrophobicity, microviscosity) by varying their ligands and thus their physical and photophysical properties [10]. In fact, these studies are extremely useful when interpreting data for the binding of metal complexes to DNA since they demonstrate how the properties of the metal complex, especially absorption and emission, vary in response to different microenvironments. In addition, there have been many studies on the effects of polyelectrolytes and micelles on rates of electron transfer involving metal complexes [10]. Since interest in using DNA to mediate electron transfer has grown recently, it is worthwhile making comparisons with these older studies which demonstrate the effects of localization of reactants on ''apparent" rate constants. This chapter attempts to take a broad view of the interactions of metal complexes with DNA. However, the emphasis is on the inter-
< previous page
page_180
next page >
< previous page
page_181
next page > Page 181
action of ruthenium complexes with DNA, with particular reference to the ongoing controversy about the modes of interaction of [Ru(phen)3]2+ and to recent studies with the complex [Ru(phen)2(DPPZ)]2+, which can truly be considered an intercalator. Similar polypyridyl complexes of other transition metals are discussed where comparisons are obvious and help to put the studies on ruthenium complexes into context. 2 Tools to Characterize Transition Metal Interactions with DNA Many experimental techniques have been applied to study the interactions of metal complexes with DNA (Table 1 collects information about use of the more unusual techniques, and Fig. 1 shows the ligands and their abbreviations as used in this chapter). In this section, we consider less the actual techniques than the information output each provides; the reader is referred to [11] for the more technical details. Here we focus on both the potential and the limitations of each technique and attempt to highlight dangers of overinterpretation. No technique can be used alone to determine the binding mode, but a judicious combination of optical and hydrodynamic techniques can be seen to provide a strong basis for characterizing the interactions of metal complexes with nucleic acids. 2.1 Optical Spectroscopy 2.1.1 Steady-State Isotropic Absorption and Emission Absorption and emission spectroscopy are routinely used to monitor interactions of dyes with nucleic acids because these optical properties tend to be quite sensitive to environment and are easily measured. It is a general observation that intercalation is accompanied by a redshift (usually attributed to a change of solvent polarity) and hypochromism (due to π-π stacking interactions with the nucleobases) in the dye absorption spectra. However, caution is advised in order to avoid overinterpretation since similar effects are observed, for example, when
< previous page
page_181
next page >
< previous page
page_182
next page > Page 182
Fig. 1. Part I.
< previous page
page_182
next page >
< previous page
page_183
next page > Page 183
Fig. 1. Part II. Structural diagrams of the different ligands which are mentioned in this chapter and the abbreviations used for them in the text (both in capital and lowercase letters).
< previous page
page_183
next page >
< previous page
page_184
next page > Page 184
TABLE 1 Examples of the More Uncommon Techniques Used to Study the Interactions of Metal Complexes with DNAa Technique used
Complexes studied
Ref.
X-ray crystal structure [Pt(tpy)Cl]+dAMP
X-ray fiber structure
148
[Pt(tpy)(HET)]+/CpG
147
[Ru(NH3)6]3+ with [d(CGCGCG)]2 (Z form)
142
[cis-Ru(bpy)2(9egua-κN7)Cl]Cl
143
[Pt(tpy)(HET)]2+ with CT-DNA
144
[Pt(py)2(en)]2+ [Pt(phen)(en)]2+ 1-D 1H & 13C NMR
[Pt(bpy)(en)2+ with CT-DNA
145
[cis-Rh(phen)2(dG-N1)(OH2)]3+
157
[cis-Rh(phen)2(dA-N3)(OH2)]3+ [cis-Ru(bpy)2(9egua-κN7)Cl]Cl [cis-Ru(bpy)2(9egua-κN7)(OH2)]PF6 [cis-Ru(bpy)2(9mhyp-κN7)Cl]Cl
143
[cis-Ru(bpy)2(9mhyp-κN7)(OH2)]PF6 1-D 1H NMR
∆- and Λ-[Ru(phen)3]2+, ∆- and Λ-[Rh(phen]3]3+ & [Co(phen)3]3+ with [d(GTGCAC)]2 & [d(CGCGCG)]2
57
∆- and Λ-[Ni(phen)3]3+, ∆- and Λ-[Ni(phen)3]3+ with [d(GTGCAC)]2
58
(table continued on next page)
< previous page
page_184
next page >
< previous page
page_185
next page > Page 185
(table continued from previous page) Technique used 2-D 1H NMR
Electric linear dichroism
Complexes studied ∆- and Λ-[Ru(phen)3]2+/[d(CGCGATCGCG)]2
60
∆-[Rh(phen)2(phi)]3+/[d(GTCGAC)]2
205
∆- and Λ-[Ru(phen)3)2+/[d(CGCGATCGCG)]2
61
∆-[Ru(phen)2(dppz)]2+/[d(GTCGAC)]2
62
[Rh(NH3)4(phi)]3+/[d(TGGCCA)]2
206
∆- and Λ-[Ru(phen)3]2+ ∆- and Λ[Ru(phen)3]2+, [Fe(phen)3]2+
37
∆- and Λ-[Ru(bpy)3]2+, [Cu(phen)2]+
38
∆-, Λ-[Ru(bpy)2(phi)]2+, [Ru(bpy)2(phi)]2+ Flow linear dichroism
Ref.
39,207
[Fe(bpy)3]2+, [Cu(bpy)2]2+
6
[Pt(bpy)(en)]2+, [Cu(bpy)2]2+
40
[Pt(bpy)(en)]2+
41
rac-[Fe(bpy)3]2+, rac-[Fe(phen)3]2+
35
rac-[Ru(bpy)3]2+
Viscometry
∆- and Λ-[Ru(phen)3]2+
42
∆- and Λ-[Ru(phen)2(dppz)]2+
21
[Pt(tpy)(HET)]+, [Pt(tpy)Cl]+
208
[Pt(tpy)(SPh)]+
77
[Pt(tpy)(SC4H9)]+
135
[Pt(tpy)(S(CH2)nS)(typ)Pt]2+
135
rac, ∆- and Λ-[Ru(phen)3]2+
87
[Ru(tpy)(dppz)(OH2)]2+
88
rac, ∆- and Λ-[Ru(phen)2(dppz)]2+
86
(table continued on next page)
< previous page
page_185
next page >
< previous page
next page >
page_186
Page 186 Continued TABLE 1 Technique used Topoisomerase unwinding
Complexes studied
Ref.
rac-[Ru(phen)3]2+ (22°) rac-[Ru(phen)3]2+ (19°)
124
rac-[Ru(phen)2(bpy)]2+ (18°) rac-[Ru(phen)2(phi)]2+ (26°)
54
rac-[Ru(bpy)2(phi)]2+ (17°) rac-[Ru(bpy)2(dppz)]2+ (30°)
125
rac-[Ru(bpy)2(dppz)]2+ (30°)
126
rac-[Ru(phi)2(bpy)]3+ (18°) rac-[Rh(phi)(phen)2]3+ (21°)
127
rac-[Ru(tpy)(dppz)(OH2)]2+ (17°) Unwinding by ccc-DNA electrophoretic mobility
Unwinding by ccc-DNA sedimentation
[Pt(py)2(en)]2+, [Pt(phen)(en)]2+ [Pt(bpy)(en)]2+ with CT-DNA rac-[Zn(phen)3]2+, [Zn(phen)2]2+, Zn(phen)Cl2 ∆- and Λ-[Ru(phen)3]2+
88 145 133 134
[Pt(tpy)(HET)]+ [{Pt(tpy)(HET)}2Pt]4+
208 209
aB form unless otherwise stated.
< previous page
page_186
next page >
< previous page
page_187
next page > Page 187
the minor groove binder 4',6-diamidino-2-phenylindole (DAPI) interacts with AT-rich regions [12]. Other nonintercalators also show redshifts accompanied by hypo- or hyperchromism on binding to DNA: di-tertbutylproflavine, modified to be too sterically bulky to allow intercalation [13], triphenylmethyl dyes such as crystal violet and methyl green [1416], and methylene blue with poly(dA)·poly(dT) [17,18]. It often appears that the extent of spectral change is related to the strength of binding and that intercalator spectra are more perturbed than those of groove binders; however, binding affinities cannot be assessed simply by comparing hypochromicities or spectral shifts. Comparison of DAPI spectra with different polynucleotides do indeed indicate that when minor groove-bound, with AT sequences, the red shift (20 nm) and hypochromicity (15%) are smaller than when intercalated, with GC sequences (29 nm and 41%) [19]. Likewise, DNA has larger effects on the metal-toligand charge transfer (MLCT) absorption of the higher affinity [Ru(phen)3]2+ (24% hypochromicity) than [Ru(bpy)3]2+ (14% hypochromicity) [20]. However, for [Ru(phen)2(DPPZ)]2+, which has a considerably greater affinity for DNA than the other two complexes, hypochromicity in the MLCT band remains low (17%) while that in the ILDPPZ band is greater (40%) [21], even though the former transition also involves the intercalated DPPZ ligand [22]. Changes in the emission properties of a dye such as spectral shifts and intensity changes can also accompany different types of binding interaction: ethidium (intercalator) and DAPI (minor groove binder) both experience large fluorescence enhancements on binding to DNA [23]. Luminescence quenching on binding to DNA tends to occur due to electron transfer reactions with the nucleotides, as seen with thionine dyes [24]. However, electron transfer quenching on binding cannot be taken to imply close association with the bases via intercalation since it is also observed for [Ru(bpz)3]2+ [25] which, like [Ru(bpy)3]2+, is not expected to intercalate. It is generally considered that efficient energy transfer from the bases to a dye is the hallmark of intercalation [26] but, surprisingly, no such studies have been reported with metal complexes. Some studies have investigated the effects of quenchers on the emission of DNA-bound metal complexes in attempts to differentiate between intercalated and surface-bound species. However, the current consensus seems to be that negatively charged quenchers do not react with DNA-bound fluorophores whereas positively charged quenchers
< previous page
page_187
next page >
< previous page
page_188
next page > Page 188
react efficiently with fluorophores bound to DNA in any mode, simply because the local concentrations are increased [27,28]. Spectroscopic data can be analyzed quantitatively to obtain binding affinities and site sizes. We define β as the fractional change of a spectroscopic property (A) as the binding ratio is varied between conditions corresponding to fully bound (Ab) and fully free dye (Af):
where Ct, Cf, and Cb represent total, free, and bound dye concentrations, respectively. It is crucial to determine Ab carefully by measuring at conditions where 100% of the dye is bound. Otherwise the fully bound value can be graphically estimated by measuring the changes under conditions where the nucleic acid is in large excess over the dye. Then, for example, for absorption, since Cb = Ct(εmεf)/(εbεf), the expression for Kapp from the mass action Eq. (3) (Sec. 2.5) can be written in the form [29]:
where Kapp is the apparent binding constant, Bapp is the binding site size, and P is the DNA concentration. A plot of 1/(εmεf) vs. 1/P should yield a linear plot with an intercept of 1/(εbεf). This expression holds as long as εb does not depend on ν (Cb/P) or on the concentration of complex, which is not true if there is more than one spectroscopically distinct bound species. Since the concentrations of free and bound dye are not determined independently here, the binding data are not as reliable as those obtained using separation techniques. However, optical techniques allow facile collection of data under many different conditions, particularly at low ionic strengths required for weakly binding compounds where dialysis cannot be applied (vide infra), and are thus widely used in this context. 2.1.2 Steady-State Polarized Absorption and Emission Some of the problems associated with measuring an average of properties of different bound forms and free dye also apply to polarized spectroscopies: linear dichroism (LD), circular dichroism (CD), and emission anisotropy (FA). Linear dichroism does have the advantage that only well-oriented material produces a signal and thus for studies of dye-
< previous page
page_188
next page >
< previous page
page_189
next page > Page 189
DNA interactions only bound dye molecules contribute to the spectrum in addition to the DNA chromophores [30,31]. Also, only dyes in a chiral environment produce a CD signal: unless the dye has an intrinsic chirality and a CD spectrum which may be perturbed upon binding to DNA, an induced CD is produced upon binding to the chiral DNA [31]. The emission of a dye in solution is usually isotropic but upon binding to DNA is likely to become polarized: the increase in anisotropy will depend on the lifetime of the dye as well as the degree to which motion is hindered. Changes in these spectroscopic properties can also be analyzed to provide quantitative binding data. However, these techniques are more powerful than isotropic spectroscopies since they provide additional qualitative information about the nature of the binding mode. Although CD spectroscopy is very sensitive, it is also very difficult to interpret CD changes in a qualitative way given our current level of understanding [31]. Calculations have shown that large induced-CD signals are expected for groove binding and smaller signals for intercalation. However, small induced-CD signals are also predicted for groove binders in certain positions [3234]. Furthermore, the sign of the induced-CD signal is very sensitive to the exact orientation of the bound dye, making it dangerous to use induced CD for more than diagnosis of interaction. However, for characterizing the interactions of chiral metal complexes with DNA, CD has proven extremely useful since it allows the enantioselectivity to be very sensitively monitored. For an inversion-stable complex, dialysis of the racemate against DNA results in the dialyzate being depleted of the favored enantiomer, resulting in a net CD spectrum; on the other hand, if the complex is inversion-labile, a net CD will result in a solution of complex and DNA resembling that of the favored enantiomer [35]. CD is also used to monitor the conformation of the DNA and is particularly useful when examining the binding of dyes to unusual conformations (Z-DNA or triplex) to determine whether the structure is perturbed by dye binding [36]. LD is more useful since it provides information about the orientation of the dyes with respect to the helix axis and, thus, the base pairs [30]. Both electric (ELD) and flow (FLD) LD techniques have been applied to binding of metal complex enantiomers to DNA [6,21,35,3742]. The limitation is that the transition moment directions in the metal complex must be well characterized to allow a definitive analysis
< previous page
page_189
next page >
< previous page
next page >
page_190
Page 190 of the binding geometry. Use of LD, a specialization of our group, is dealt with in greater detail in Sec. 4. Emission anisotropy (FA) reports how rigidly the dye is held in its binding site within the duration of its excited state lifetime. It is considered to be an important tool for differentiating between intercalated and either groove-bound or surface-bound dyes, and has been applied in several transition metal (TM) binding studies [4346]. FA can also be used to monitor macromolecular motions, if a probe with a lifetime longer than the motion time scale can be attached to the macromolecule. In this context, ruthenium complexes are now being suggested as probes of protein folding [47]. 2.1.3 Transient Absorption and Time-Resolved Emission Transient absorption spectroscopy probes dye excited states and, while it provides no information about the binding mode, it is a useful adjunct to emission studies since it can help clarify which processes give rise to emission enhancement or quenching upon binding. It has been used, for instance, to characterize the bifunctional DNA binder [Ru(TAP)2-(POQ)]2+ [48], to demonstrate electron transfer in the quenching of *[Ru(TAP)3]2+ by DNA [49,50], and to study the mechanism of DNA damage by the
system [51,52].
Measurements of emission lifetimes can provide a great deal more information about the behavior of the emitting excited state than steady-state measurements can. For instance, if there are two binding sites causing opposing effects on the emission, this could only be seen by monitoring the decay of the excited state. Studies of lifetimes also allow effects of quenchers or oxygen on different species, e.g., free and bound, to be monitored. 2.1.4 Infrared and Resonance Raman Although vibrational spectroscopy has been applied to investigate nucleic acid structures, it is rarely used to probe the binding of drugs to nucleic acids [53]. However, in a few cases resonance Raman spectroscopy (RR) has been applied in studies of metal complex-DNA interactions [46,5456]. RR reports how binding to DNA affects the energy levels of metal complexes. It has been suggested that changes in spectra on binding to DNA can be used to define an interaction. Thus, it has
< previous page
page_190
next page >
< previous page
page_191
next page > Page 191
been suggested that DIP interacts more strongly than bpy in [Ru(bpy)2-(DIP)]2+ [54] and phen more than tpy in [Ru(tpy)(phen)(OH2)]2+ [56], but interpretation is not trivial. Little work has been carried out in this field and it is best to exercise caution in using this technique quantitatively. 2.2 NMR Spectroscopy After crystallography, nuclear magnetic resonance (NMR) is the technique which can potentially provide most direct structural details about drug-DNA complexes. The advent of 2-D NMR has been revolutionary; it is now possible to pinpoint contacts between the DNA and the drug. Previously, using proton shifts, one could say which parts of the drug or the DNA were affected by the binding, but it was impossible to say whether this was due to direct interactions or structural distortions induced by the binding. Proton shifts as a function of binding ratio can be analyzed like absorption and emission changes to provide binding constants; if binding is at a specific site, the binding constant for that site is obtained in this way provided the free ligand concentration can be determined simultaneously. The first studies with octahedral metal complexes employed 1-D NMR [57,58], and since large downfield minor groove shifts and small upfield major groove shifts were observed, it was suggested that for both enantiomers binding in both grooves occurred, with intercalation in the major groove favoring the ∆ enantiomer and ''surface binding" in the minor groove favoring Λ. However, the upfield shifts in the major groove are consistent with binding of the dye solely in the minor groove, since similar shifts are observed when DAPI binds to AT sequences [59]. More recently, 2-D NMR has shown minor groove contacts for both [Ru(phen)3]2+ enantiomers with a decamer [60,61] and also major groove contacts for ∆-[Ru(phen)2(DPPZ)]2+ with a hexamer [62]. 2.3 Electrochemical Methods As with other physical properties, the bulk electrochemical properties of a dye solution are altered when the dye associates with DNA. Hence, upon interaction with DNA, the amplitude of a metal complex cyclic
< previous page
page_191
next page >
< previous page
page_192
next page > Page 192
voltammogram is reduced in the presence of the polymer, in a manner which depends on the binding ratio [6367]. As long as the limiting diffusion coefficient of the bound dye can be measured with certainty, the change in current can be analyzed to obtain binding constants and site sizes. However, the technique is nonstandard and, excepting unusual cases, is unlikely to yield better data than absorption or emission titrations, which are easy to perform. Its principal advantage is the possibility of acquiring binding parameters for different redox species of a complex from the same experiment. 2.4 Separation Methods For quantitative purposes, techniques which separate bound and free dye provide the most reliable data for determination of binding parameters. These include equilibrium dialysis, DNA affinity chromatography, sedimentation, and filtration methods. The concentrations of free and bound dye are determined independently, which removes the greatest limitation imposed by use of indirect methods; and there is no requirement that complete binding be attained. However, these techniques are time and material consuming; hence fewer data are usually collected than with spectroscopic titrations. Dialysis is the most widely applied separation technique, but one problem is that moderately high salt concentrations (~ 50 mM Na+) must be used to avoid Gibbs-Donnan effects [68]; lower ionic strengths are generally required to favor binding of low-affinity species. Thus, for a species like [Ru(bpy)3]2+, which barely binds in 50 mM NaCl, dialysis is less useful than absorption or emission, which can be used at low ionic strengths. 2.5 Binding Curves Various techniques provide quantitative binding data in terms of free (Cf) and bound (Cb) dye, and DNA (P) concentrations. We describe here several models which have been used to analyze binding data for metal complexes with DNA. It is common practice to present binding data as a Scatchard plot, i.e., ν/Cf vs. ν, where ν≡ Cb/P which is superlative in its ability to enhance nonlinearities due to binding site overlap and multiple binding
< previous page
page_192
next page >
< previous page
page_193
next page > Page 193
sites. When cooperative binding effects occur, the binding isotherm, ν vs. log Cf, is also an appropriate plot. The basic binding model of Scatchard [69] derives from a simple consideration of the law of mass action, where Bapp represents the apparent number of binding sites per P (which we define in this chapter as a base pair*) and Kapp is the apparent binding constant:
Thus, the binding data should yield a linear plot with slope = Kapp and Y intercept = KappBapp. Nonlinearity of the binding plot may arise from several effects which are difficult to distinguish: overlap of binding sites, cooperative effects, existence of two different binding sites, or dye aggregation in solution. When dimerization occurs the initial slope of the curve becomes (Kapp + 2Kd) [29]. The presence of two different binding modes should not be inferred from a curved plot unless there is independent physical evidence (e.g., absence of isosbestic point) that the nature of the bound complex changes with ν. Equation (4) can also be expressed in terms of the concentration of bound complex; this form of the equation has been used in several studies [56,6367,70,71]:
However, despite being widely used in studies of metal complexes with DNA, the Scatchard formalism is not generally appropriate for analysis of small-molecule binding to DNA, since it only applies when the binding sites are independent, i.e., so well separated that they can be treated as if they were on separate molecules. *Definition of the DNA concentration in terms of base pairs facilitates later descriptions of more sophisticated models in a physically significant manner. Of course, the equation still applies for P defined as base concentration, as long as Bapp is described in the same units.
< previous page
page_193
next page >
< previous page
page_194
next page > Page 194
A more rigorous model developed by McGhee and von Hippel [72] which can be applied to binding to homogeneous lattices (i.e., sequence-neutral binding to DNA, or binding to homopolynucleotides) describes, for noncooperative binding, the Scatchard plot as
Here n is the number of base pairs included in a binding site (1/B).* The main feature of this model is that saturation of all binding sites becomes difficult because of the buildup of sites smaller than the binding site at high binding ratios. Hence, the plots are generally nonlinear, even when there is only one type of binding site. The Y intercept of this plot reads off directly as K. The intercept of the X axis is 1/n but can be difficult to evaluate;n can also be obtained by extrapolation of the low ν slope to the abscissa, which then yields an intercept of 1/(2n 1). This equation reduces to the Scatchard equation when n = 1, i.e., there is no possible overlap of potential sites. It is important to realize, when analyzing experimental data, that n and K are interrelated. Also, if data are collected only in the lowν region, fitting may yield unrealistically high values for K and n. The McGhee/von Hippel treatment can be further extended to include cooperative effects [72], but since this model has not been used in studies of metal complexes with DNA we do not describe it here. For more complicated binding systems, the excluded site model of Crothers [73] deals with binding of small molecules to a heterogeneous lattice, i.e., for sequence-specific complexes with CT-DNA [74]:
where α (0 ≤α≤ 1) is the fraction of the DNA molecule which contributes to binding and K(0) is the intrinsic binding constant for an isolated potential binding site. Hence, the Y intercept gives an apparent binding constant αK, and K(0) cannot be evaluated unless the nature of the binding site and, thus, α is known. This model has been used by Lippard to analyze the GC dependency of the binding of Pt complexes to heterogeneous DNAs [75]. When α = 1, so that all base pairs represent potential binding sites, this equation reduces to Eq. (6). Equations describing *Note that in some papers (e.g., [54]), P is given in base concentration and n in base pairs, thus necessitating the inclusion of factors of 2 with the ν term.
< previous page
page_194
next page >
< previous page
page_195
next page > Page 195
cooperative binding to a heterogeneous lattice have also been described [76] and applied to analyze binding of [Pt(tpy)(HET)]+ to RNA. 2.6 Kinetics Very few studies of binding kinetics of substitution-inert metal complexes with DNA have been made. The [Pt(tpy)(SPh)]+ complex [77] associates too rapidly with DNA to be followed by stopped flow, but the dissociation is found to occur via at least three species en route from the bound to free drug states. Ionic strength effects on association and dissociation constants have proved to be good discriminators between intercalative and external binding [59,78], but the potential of this approach has yet to be realized within the field of metal complexes. 2.7 Thermodynamic Methods 2.7.1 Salt Effects on Binding Constants For a monovalent inert salt (and neglecting, e.g., anion release from the drug [79]), various models for the ionic strength dependence of the binding constant for drugs with DNA (modeled as a uniformly charged cylinder) [7982] can be combined to yield the equation:
where Kobs is the observed binding constant at a concentration C of inert salt. KIM is the binding constant at 1 M inert salt, commonly used as an estimate of the nonelectrostatic contribution to the binding affinity. The slope ℜ in a plot of In Kobs vs. In C represents the salt effects in terms of counterion release due to the binding of the drug charge (ZΨ) and/or as a response to an altered charge density of the DNA due to dye-induced changes of the helix structure (B). Z is the drug charge and Ψ is the fraction of phosphate charges neutralized by the counterions. In the Record-Wilson theory, Ψ = 0.88 for B-DNA [79] but 0.82 when the helix is extended by intercalation [82], whereas the Manning theory amounts to setting Ψ = 1 in both cases [81]. In both theories a helix extension of 3.4 Å per bound drug corresponds to B = 0.24, and if the drug does not alter the DNA structure in a way that affects the charge density, then B = 0.
< previous page
page_195
next page >
< previous page
page_196
next page > Page 196
Neutral intercalators have ℜ = 0.20.4 [81,83,84] indicating ionic effects (despite neutrality), in fair agreement with theory. Monovalent [78,83,85] and divalent [59,78,82,85] intercalators have ℜ in the range 1.01.2 and 1.92.3, respectively, again in fair agreement with both the Record-Wilson (1.06 and 1.88) and the Manning theories (1.12 and 2.12). Slopes close to 2 have been reported for the divalent [Ru(phen)2-(DPPZ)]2+ [86] and [Ru(phen)3]2+ [44]. Considerably lower slopes for [Ru(phen)3]2+ (1.4 for ∆ and 1.2 for Λ) have been observed by Satyanarayana et al. [87]. Since [Ru(phen)3]2+ is believed not to intercalate [27,42,87], a smaller slope than for [Ru(phen)2(DPPZ)]2+ is indeed expected in the absence of helix extension, but according to experiments on neutral intercalators and theory by only ~0.2 (vide supra). Furthermore, [Ru(tpy)(DPPZ)(OH2)]2+ also has an unusually low value for ℜ (1.3) [71], despite being intercalated [88]. We shall return below to this apparent tendency for the octahedral Ru complexes to violate the counterion release theory. In contrast, the planar intercalating complexes [Pt(tpy)(SPh)]+ [77] and [(en)Pt(phen)]2+ [75] have values for ℜ of 1.2 and 2.2, respectively, again agreeing with theory. Studies of salt effects on the binding constant also provide information about the nonelectrostatic contribution to the binding in terms of the KIM values emerging from such an analysis. K1M values typical for classical intercalators (104105 M [78,83]) are observed for DPPZ-containing Ru complexes [86,71], whereas [Ru(phen)3]2+ [87] has at least a 100-fold lower K1M and thus is almost exclusively electrostatically bound. Using this approach to resolve charge effects, similar values of K1M for the metallointercalators [(phen)Pt(en)]2+ and [(tpy)Pt(HET)]+ were used by Howe-Grant and Lippard [75] to infer that the tpy and phen ligand ring systems stack in similar ways to the DNA bases. Salt effects on the binding constant can be used to separate ionic and nonionic contributions but only to the free energy of binding. This is usually a serious limitation since similar values of ∆G° sometimes hide great diversity in thermodynamic driving force for drug binding to receptors in general [89], and to DNA in particular [9093], due to enthalpy-entropy compensation. A full thermodynamic analysis in terms of ∆H° and ∆S° is therefore required to give a proper understanding of the source of the affinity of a certain drug for DNA.
< previous page
page_196
next page >
< previous page
page_197
next page > Page 197
2.7.2 Enthalpic and Entropic Contributions to Binding Constants It is usual to either (1) obtain ∆S° and ∆H° from the temperature dependence of the binding constant (van't Hoff analysis) or (2) obtain ∆G° from the binding constant, ∆H° through calorimetric measurements, and finally evaluate ∆S° = (∆H° ∆G°)/T. Breslauer and coworkers [94] stressed that the latter procedure is more reliable [95]. The binding of ethidium monocation to double-stranded DNA (ds-DNA) is characterized by a large negative ∆H° which is independent of salt concentration [96,97], whereas the likewise negative ∆S° becomes more negative as salt is added [96,97]. There is thus a strong enthalpic driving force for binding, usually interpreted as being due to stacking interactions with the bases. The electrostatic (salt-sensitive) contribution is mainly of entropic origin, and it becomes more unfavorable the higher the ionic strength. This is in agreement with a counterion release mechanism, since the entropy gained by releasing counterions from the ion cloud around the helix into the bulk is lower at higher background ion concentrations [79,80], but the binding cannot be said to be driven by counterion release since the overall entropy is unfavorable. A recent theoretical calculation of the free energy of binding [98] using more detailed, atomic models for various DNA-drug complexes essentially reproduces the predictions for ℜ of the models mentioned earlier. However, not only the entropy corresponding to the organization of the counterion cloud but also the other free energy contributions to binding (ion-ion, ion-drug, and ion-DNA interaction) were found to depend on log[Na+] in a linear fashion in this model. Several of these interactions can be expected to be at least partly enthalpic in nature and, thus, enthalpic effects may contribute to the slope ℜ [98]. Indeed the enthalpy of binding of daunomycin to DNA [90] decreases with increasing salt, in contrast to the saltinsensitive enthalpy for ethidium. It is thus clear that the source of the salt dependence of ∆G° must be different for ethidium and daunomycin, despite both dyes having a slope ℜ [for Eq. (8)] in agreement with counterion release predictions. The good agreement for ℜ with the cylinder models may thus be fortuitous due to enthalpy-entropy compensation [90,99]. Detailed studies of salt effects on enthalpy and entropy for the Ru com-
< previous page
page_197
next page >
< previous page
page_198
next page > Page 198
plexes are therefore motivated in order to investigate if the unusually low values of ℜ are due to a lack of such compensation effects. Calorimetry shows that at 50 mM NaCl the binding of [Ru(phen)2(DPPZ)]2+ is endothermic [86] and that the binding thus is driven by entropy. A similar picture emerges for [Ru(phen)3]2+ [86], but in neither case have salt effects on ∆S° and ∆H° been studied. Van't Hoff analysis shows that the planar drug [Pt(tpy)(PhS)]+ [77] also exhibits a strong favorable entropy of binding, which decreases with increasing ionic strength. This system therefore constitutes one of the most convincing cases of a binding driven by counterion release. For this intercalator [77] the enthalpy is favorable too, and to an extent that suggests a stacking interaction similar to that of classical intercalators. However, adding salt increases the enthalpy, in strong contrast to classical intercalators where ∆H° is salt-insensitive (ethidium bromide, EB [96]) or decreases with increasing salt (daunomycin [90]). Thus in terms of the enthalpy neither [Pt(tpy)(PhS)]+ nor [Ru(phen)2(DPPZ)]2+ (where ∆H° is essentially zero) conforms to the thermodynamic picture of the classical intercalators, although both metal complexes have been convincingly shown to intercalate. Strongly exothermic binding reactions are not limited to intercalative binding. The proven minor groove binders DAPI [100] and netropsin [101] both bind to DNA with a strong favorable ∆H°. 2.8 Hydrodynamic Methods 2.8.1 Lengthening of the DNA Helix The extension of the DNA helix due to insertion of a planar drug such as EB into the intercalation pocket can be directly observed by electron [102,103] as well as optical microscopy [104]. Usually simpler to apply as routine techniques, hydrodynamic methods are very useful for measuring lengthening effects, but in order to obtain quantitative data models have to be used. For short DNA hydrodynamic theory for rods can be applied to extract effective DNA lengths from data obtained by viscosity and sedimentation [105] or from the rotational diffusion constant [106]. For quantitative binding the lengthening is proportional to added dye, and an extension of 3.4 Å per bound drug corresponds to a length increase of ~10% (and a 33% increase in specific viscosity or rotational
< previous page
page_198
next page >
< previous page
page_199
next page > Page 199
diffusion coefficient) at a drug/phosphate ratio of 0.05. The classical intercalators EB and proflavine increase the DNA length by about 3 Å [105,106], whereas there is no increase in DNA length upon binding of the groove-bound derivative di-tert-butylproflavine [13]. For DNA molecules longer than 90 bp [107] DNA flexibility becomes important, and attempts to include effects of drug binding on helix stiffness in the theories have been made [108]. For comparative studies, however, the rod theory has proven applicable for mildly heterogeneous DNA samples centered around 200 bp, which have the advantage of being relatively easily obtained by sonication [109,110]. The degree of orientation of DNA in shear flow is sensitive to DNA length [30], as evidenced by the increase of the degree of flow orientation of linear DNA induced by ethidium bromide [111], Interpretation is hampered by a lack of quantitative models for extension and stiffness effects, but valuable qualitative information may be gained. The Λ enantiomer of [Ru(phen)2(DPPZ)]2+ increases DNA orientation by a factor of 2, compared to only a 25% increase for ∆ [21]. This suggests that Λ stiffens the DNA more than ∆ since viscosity of rod-like DNA indicates that the enantiomers extend the helix to the same extent [86]. However, the apparent length increases by as much as 25% at a drug/phosphate ratio of 0.05 [86]. This is in clear excess of extensions obtained with classical intercalators, and further studies of the mechanism underlying the increase in viscosity and orientation by [Ru(phen)2(DPPZ)]2+ are required. Extension measurements have proved useful for elucidating interesting aspects of the binding of Cu complexes to DNA. Viscosity of rodlike DNA shows that [Cu(phen)2]+ intercalates whereas [Cu(dmp)2]+ does not [112]. The lack of intercalation is not due to steric interference in the intercalation pocket from the methylation, since the dmp ligand (like phenanthroline) can itself intercalate in DNA [110]. Sigman [1] instead attributed the lack of intercalation to the increased dihedral angle between the ligands resulting from methylation. Veal and Rill [112], using 10% ethanol, failed to observe any extension by the phen ligand itself, in contrast to the findings of Graham and Sigman [110] in aqueous buffer, which may simply be due to weak phen binding in the less polar solvent. A similar decrease in ethidium affinity for DNA in ethanol-containing solvent may explain the results of Liu et al. [113]. From the
< previous page
page_199
next page >
< previous page
page_200
next page > Page 200
finding that (in 33% ethanol) [Cu(bcp)2]+ increases viscosity about twice as much as observed for EB, they suggest that this complex may increase viscosity by bridging between different DNA molecules instead of helix extension. However, the viscosity increase observed by Liu et al. is very close to the value corresponding to a 3.4-Å helix extension per drug, and it is rather the ethidium extension which is unusually low in these solvents [113]. Even in 10% ethanol EB causes a 50% smaller extension [112] than in aqueous solution [106], when comparison is based on the ratio of added drug to phosphate. Intermolecular bridging is thus not necessary to explain the viscosity data of Liu et al. [113]. In fact, viscosity increases corresponding to helix extensions of even more than 3.4 Å per complex have been observed repeatedly, e.g., for actinomycin D [114], ihremycin [115], and 9-aminoacridine [116], but also in these cases there is ample evidence for pure helix extension, enhanced by tilting of the intercalated drug in the pocket. 2.8.2 Unwinding of Supercoiled DNA Accommodation of an intercalating drug requires unwinding of the helix [117], which can be monitored by hydrodynamic techniques by use of supercoiled circular DNA, acting as an amplifier by transforming the change of secondary structure into changes in the global shape of the molecule. The decrease in degree of supercoiling that results from unwinding is observed as a decrease in sedimentation velocity [118], in electrophoretic mobility in gels [119,120], or in the orientation of the DNA in flow gradients [111], or as an increase in viscosity [13]. Importantly, nicked circular DNA provides a good control for drug-induced effects on the DNA structure not related to unwinding. The unwinding angle of the drug can be inferred from the drug/phosphate ratio where the hydrodynamic properties of the supercoiled and nicked circle coincide, if the fraction of dye bound and the initial degree of supercoiling is known. The latter can be calibrated with a drug of known unwinding angle [121], and usually ethidium (26°) is used [117]. In the mobility-based method free dye must be present in the gel in order to maintain the binding ratio during the electrophoretic analysis [120,122,123]. This experimental limitation is absent in an alternative assay, where supercoiled DNA is treated with topoisomerase in the presence of the dye at various drug/DNA ratios [54,88,124127]. Elec-
< previous page
page_200
next page >
< previous page
page_201
next page > Page 201
trophoretic analysis after removal of the drug allows the unwinding angle to be calculated from the drug-induced shift of the ladder of bands corresponding to different degrees of supercoiling. The applicability of this ''band counting" method [119] was considerably improved by the development of an Mg2+-free buffer for the topoisomerase by Kelly et al. [124], who also point out the importance of including controls for inhibitory effects on the topoisomerase by the drug under investigation. Unwinding is commonly used to demonstrate intercalation. The classical intercalator EB has been shown to unwind DNA by all methods available [111,118120], and Waring [117] summarized examples of many other intercalators, such as proflavine. From observations such as the lack of unwinding by the groove-binding derivative di-tertbutylproflavine [13], unwinding has sometimes come to be regarded as equivalent to intercalation. However, several observations have indicated that nonintercalators may also affect the helicity of DNA. Crystal violet does not intercalate according to viscosity of rod-like DNA [14,128] but still unwinds DNA [14], which is not inconsistent with suggested groove binding [15,16], since minor groove binding at least can affect helix winding [129]. Unwinding by nonintercalators has also been observed for irehdiamine A [118,130], which kinks DNA [131], and for cis-[Pt(NH3)2Cl2], which is believed to cause localized helix denaturation [132]. These observations demonstrate that many perturbations of the helix other than intercalation have to be considered in the interpretation of unwinding data. The ideal approach is to combine unwinding and extension measurements, as nicely illustrated by the elucidation of the different intercalation geometries for anthracycline drugs [114,115]. Indeed, interesting results have emerged for those metal complexes that have been studied with both techniques. The DNA binding of numerous racemic metal complexes have been studied by unwinding, in most cases by the topoisomerase approach. Pure bpy complexes ([Ru(bpy)3]2+) do not unwind DNA [124], but complexes containing the phen-(Ru [54,124], Zn [113]), phi- (Ru [54], Rh [127]), and DPPZ- (Ru [88,126]) ligands do. For Ru(II) these data suggest a slight trend in unwinding efficiency in the order DPPZ (30°) > phen (1922°) > bpy (0°). In the only study of enantiomeric effects on unwinding of which we are aware, a lower concentration of added ∆-[Ru(phen)3]2+ than of Λ wasrequiredtoinducefullunwindingtothe relaxed state [134]. However, in the absence of binding corrections it is
< previous page
page_201
next page >
< previous page
page_202
next page > Page 202
not possible to determine if these observations are really due to differences in unwinding angle or are only an effect of different binding affinities. Quantification would be worthwhile since the two [Ru(phen)3]2+ enantiomers have different effects on the viscosity of rod-like DNA [27,87]. In neither case are the viscosity data consistent with helix lengthening, however, so [Ru(phen)3]2+ is another example showing that unwinding can have other causes than extension, in this case suggested by Satyanarayana et al. [27] to be helix bending by a partial insertion of one phen ligand. Pt-tpy complexes with phenyl [77] or butanethiol [135] as the second ligand intercalate in B-DNA according to the viscosity assay on rod-like DNA. When linked by αω-dithiols of various chain lengths to form binuclear Ptterpyridine complexes [135], bis-intercalation occurs if the linker chain contains five carbon atoms or more (the corresponding minimum linker length of 8.2 Å requires the nearest-neighbor exclusion principle to be violated, as was also found for bisacridines [136]). As expected, the mononuclear Pt-terpyridine complexes unwind DNA, by 17°, and binuclear complexes unwind by twice as much, 3236° [135]. Interestingly, unwinding occurs even if the linker between the two Pt-terpyridines contains as few as four carbon atoms. The butane linker hence allows binding of the complex that results in severe helix distortion (double unwinding compared to the monomer) without full intercalation (extension comparable to monomer). The results of Kelly et al. [124] and Jones et al. [137] suggest an interesting classification based on effects of Mg2+ ions on the unwinding angle for different dyes [124]. Classical intercalators (e.g., EB) have an unwinding angle which is insensitive to salt, whereas drugs which unwind but do not lengthen DNA, such as [Ru(phen)3]2+, tend to have a dramatic reduction in the unwinding angle (per bound dye) as the salt concentration is raised. 2.8.3 Conductivity and Electrophoresis The release of counterions which accompanies binding of drugs can be monitored by changes in conductivity upon binding of the complex to DNA. Stradowski et al. [138] showed that the binding of [Ru(bpy)3]2+ leads to the release of about one monovalent counterion per drug from the DNA. This is significantly smaller than the value of approximately
< previous page
page_202
next page >
< previous page
page_203
next page > Page 203
two expected from the drug charge, and [Ru(bpy)3]2+ can thus be added to the list of Ru complexes with an unusually low degree of counterion release upon binding. In addition to the unwinding assay based on gel electrophoresis of circular DNA, gel mobilities of linear DNA have potential to provide information about drug-induced effects on the DNA structure. An obvious example is to probe lengthening of the DNA helix by intercalators, since gel electrophoresis is routinely used to discriminate between linear DNA molecules of different sizes, and also stiffening or kinking of the helix by drugs. Pore sizes of polyacrylamide and agarose gels approach molecular dimensions in DNA, which makes electrophoretic mobilities sensitive to changes in the DNA structure, as demonstrated by the unwinding assays or the discovery of bent DNA in AT tracts [139] using this technique. In fact, gel electrophoresis has the advantage over conventional hydrodynamic techniques [108] in that stiffness and bending effects can be resolved from helix extension effects experimentally by using long DNA (over 50 kbp in agarose gels, smaller in polyacrylamide gels), for which the mobility is insensitive to DNA length [140]. An illustrative example is given by the effects of binding of the enantiomers of [Ru(phen)3]2+ on the gel mobility of monodisperse DNA samples between 2 and 170 kbp [K. Gisselfält et al., to be published]. In 1% agarose, binding of the ∆ enantiomer reduces the mobility of a 170-kbp DNA (compared to native DNA) more than twice as much as the Λ, an effect which cannot be due to difference in helix extension since mobility is independent of DNA length in this size range. The Λ-induced mobility decrease for long DNA can be explained by the reduced DNA charge. The additional reduction of mobility by ∆ is likely to reflect drug-induced bends in the DNA [Gisselfält], as suggested by a flow orientation and viscosity decrease induced by the ∆ enantiomer [27,42]. Finally, gel electrophoresis can be efficiently used as an analytical technique for detection of complex formation by mobility shift assays. Formation of a complex usually affects the DNA mobility in terms of changes in charge, flexibility, or simply mass. In this vein, Feeney et al. [141] report the photo-induced formation of covalent complexes between [Ru(TAP)3]2+ and ss- and ds-DNA. For detailed structural analysis of the adduct spectroscopic methods were employed, but the mobility shift assay proved efficient in screening for effects of buffer concentration and composition on adduct formation.
< previous page
page_203
next page >
< previous page
page_204
next page > Page 204
2.9 X-Ray Diffraction This is the technique which can provide the most accurate information about the location of the dye in its DNA complex within fibers and single crystals, although questions remain as to whether a static crystal structure limited by packing considerations can really be equated with a dynamic solution structure. To date, there are no crystal structures for the octahedral metal complexes on which we are focusing. A structure of a Z-form oligonucleotide induced by [Ru(NH3)6]3+ was reported and compared with the same form induced by the more commonly studied [Co(NH3)6]3+ [142]. A crystal structure was also recently reported for a covalent ruthenium-bipyridyl complex with ethylguanine [143]. Early fiber diffraction results for platinum-bpy, -tpy, and -phen complexes were consistent with intercalation of these planar complexes [144,145], as suggested by other physical measurements [40,41,146]. Subsequently, [Pt(tpy)(HET)]+ was crystallized with a CpG minihelix in a stacked structure supporting intercalation [147] and [Pt(tpy)Cl]+ was crystallized with dAMP in an unusual stacked structure where the adenine residues formed Hbonded base pairs [148]. Other crystal structures have been solved, however, for both intercalators and minor groove binders [149]. We pay particular attention to the structure of the chiral intercalator actinomycin D, which has been crystallized with an oligonucleotide [150] and shows a structure very similar to that postulated from earlier structures obtained with deoxyguanosine [151,152]. Thus we consider the actinomycin structure as a model for the structure of a bulky intercalated compound and compare it with [Ru(phen)2(DPPZ)]2+ in Sec. 4. 3 Interactions of Transition Metal Complexes with DNA 3.1 Binding and Photophysical Properties of Bound Transition Metal Complexes The first substitution-inert metal complexes to be studied with DNA were analogs of the platinum anticancer drugs [146]. Since the ensembles formed by these drugs with DNA and dinucleotides were amenable
< previous page
page_204
next page >
< previous page
page_205
next page > Page 205
to X-ray diffraction [144147], it could be shown that these square planar metal complexes were intercalated in DNA. This discovery prompted others to extend the range of metal complexes studied. By variation of the central metal, it is possible to change not only the physicochemical properties of the complex (photophysics, electrochemistry), but also the geometry (tetrahedral, octahedral). Furthermore, subtle changes in the attached ligands allow the physical properties to be further tuned and alter the global shape of the complex. 3.1.1 Enantioselectivity From the earliest studies it has been clear that chiral metal complexes show some enantiospecificity in their interactions with DNA (Table 2). In the most extreme case of [Ru(DIP)3]2+, it was reported that the enantiomers bind with different specificities, ∆ only to B-DNA and Λ only to Z-DNA [45]. The concept of shape selectivity is addressed later; the present section concentrates on interactions with the normal B-form helix. In an early study, Nordén and Tjerneld [6] first observed the enantioselective interaction of [Fe(bpy)3]2+ with DNA. Since this complex is inversion-labile, a differential CD spectrum resembling that of the preferentially bound enantiomer is produced in the presence of DNA (Pfeiffer effect). It was possible to attribute this CD to the ∆ enantiomer. Subsequent studies with other complexes, using the sensitive method of removing the DNA-bound metal complex and measuring the CD spectra of the inversion-inert supernatant, have shown that a small enantioselectivity is, in fact, a common feature of the interaction of octahedral metal complexes with DNA (Table 2). There have been reports that there is no enantioselectivity for binding of [Ru(bpy)3]2+ to ds-DNA [54,138], but some of these experiments were carried out in high salt where little complex is bound [54]. However, other studies have clearly shown that there is indeed enantioselectivity for binding of both ∆-[Ru(bpy)3]2+ and ∆-[Fe(bpy)3]2+ to calf thymus DNA (CT-DNA) [35]. By comparison with these small complexes, the enantioselectivity observed for bulky, hydrophobic tris(DIP) metal complexes is quite remarkable: it appears that Λ-[Ru(DIP)3]2+ cannot interact with B-DNA at all, while ∆[Ru(DIP)3]2+ binds at least as strongly as
< previous page
page_205
next page >
< previous page
next page >
page_206
Page 206 TABLE 2 Reports of Enantioselectivity in Binding of Metal Complexes to DNA Complex
Technique
Selectivity
Ref.
[Fe(bpy)3]2+
CD: Pfeiffer effect
∆ preferred with B-DNA
[Zn(phen)3]2+
Dialysis & CD
Enantioselectivity:∆? with B-DNA
133
[Ru(phen)3]2+
Dialysis & CD
∆ preferred with B-DNA
134
[Ru(DIP)3]2+
Absorption changes
Only ∆ binds to B-DNA
153
[Ru(TMP)3]2+
Dialysis and CD
6
Λ preferred with A-RNA (no binding to B-DNA)
[Fe(bpy)3]2+, [Fe(phen)3]2+ [Ru(bpy)3]2+ → CD: Pfeiffer effect ∆ preferred with B-DNA for all complexes → Dialysis & CD
35
(a) ∆ preferred with B-DNA
a) [Fe(phen)3]2+ b) [Ru(phen)3]2+
159
CD
(b) ∆ preferred with Z-DNA, due to Z→B conversion
36
a) [Ru(bpy)2(phen)]2+ [Ru(phen)2(bpy)]2+ [Ru(5-NO2-phen)3]2+ (a) ∆ preferred with B-DNA
[Ru(bpy)2(DIP)]2+ [Ru(phen)2(DIP)]2+
Dialysis & CD
54
[Ru(DIP)2(phen)]2+ b) [Ru(bpy)2(phi)]2+
(b) Enantioselectivity, but absolute configurations not assigned
[Ru(phen)2(phi)]2+
(table continued on next page)
< previous page
page_206
next page >
< previous page
next page >
page_207
Page 207 (table continued from previous page) Complex
Technique
Selectivity
Ref.
[Ru(bpy)2(ppz)]2+ DNA chromatography & CD
∆ preferred with B-DNA
155
[Ru(bpy)2(qpyMe2)]2+
Dialysis & CD
∆ preferred with B-DNA
164
[Ru(phen)3]2+ [Ru(DIP)3]2+
∆ preferred with B-DNA Fluorescence intensities and polarization ∆ preferred with Z-DNA
45
[Ru(phen)3]2+
Binding constants
∆ preferred with B-DNA
87
[Ru(bpy)2(ppz)]2+
Binding constants
∆ preferred with B-DNA
46
[Ru(bpy)2(phi)]2+
Binding constants
∆ slightly preferred with B-DNA
39
[Ru(phen)2(dppz)]2+
Binding constants
∆ preferred with B-DNA
86
[Ru(phen)3]2+ [Ru(bpy)2(ppz)]2+
a) cis-[Rh(phen)2Cl2]+ cis-[Rh(phen)2(OH2)2]3+
(a) Λ preferred covalently and noncovalently
b) cis-[Rh(phen)2(OH2)Cl]2+ Dialysis & CD
(b) Λ preferred covalently
c) cis-[Rh(bpy)2Cl2]+
(c) No enantioselectivity
157
[Ru(phen)2(OH2)2]2+ [Ru(bpy)2(OH2)2]2+ [Ru(phen)2(OH2)(py)]2+
Ultrafiltration & CD
Λ preferred for covalent binding to B-DNA
158
[Ru(bpy)2(OH2)(py)]2+ Ru(phen)2Cl2
< previous page
EtOH precipitation & CD
Covalent enantioselectivity; Λ with B-DNA
page_207
156
next page >
< previous page
page_208
next page > Page 208
[Ru(phen)3]2+ [45,153]. Since the Λ enantiomer does appear to bind to alternative DNA conformations such as the Z form, it has been suggested as a probe for non-B-form DNA [1,5,154] (vide infra). More recently, complexes with an extended ligandpyridophenazine (PPZ)have shown similar selective binding of the ∆ enantiomer to B-DNA [155]. Binding of complexes with even more extended ligands such as dipyridophenazine (DPPZ) is too strong to allow use of such a separation technique under reasonable salt conditions. However, a preference of the ∆ enantiomer for B-DNA is apparent from a comparison of the recently determined binding constants of the enantiomers [86]. Similarly, in the few cases where binding constants have been determined for both enantiomers, some degree of preferential binding of the ∆ enantiomer is predicted for B-DNA: [Ru(phen)3]2+ [87], [Ru(phen)2(PPZ)]2+ [46], and [Ru(bpy)2(phi)]2+ [39]. It should be noted, however, that the degree of enantiospecificity will depend on the ionic strength used in such studies, since different salt dependencies of binding constant have been found for ∆ and Λ enantiomers of [Ru(phen)3]2+ [87] and [Ru(bpy)2(DPPZ)]2+ [86]. For complexes of the series [Ru(X)2(DIP)]2+ and [Ru(X)2(phi)]2+, it has been reported that the enantioselectivity increases when the ancillary ligand X is changed from bpy to phen [54], suggesting that the ancillary ligands are responsible for the enantioselectivity. It is interesting that the opposite enantioselectivity is observed in two cases. The covalent binding to CT-DNA of ruthenium and rhodium complexes of the type cis-[M(L)2X2]n+, where L is a diimmine ligand and X is a labile ligand, is selective for the Λ enantiomer [156158]. Moreover, the Λ enantiomer of the complex cis-[Rh(phen)2Cl2]+ is preferentially bound even before the covalent reaction occurs [157]. Additionally, the Λ enantiomers of metal complexes seem to be favored for binding to non-B nucleic acid conformations: Λ-[Ru(TMP)3]2+ with A-form RNA duplexes [159], and Λ-[Ru(DIP)3]2+ and Λ-[Ru(phen)3]2+ with [Co(NH3)6]3+-induced Z-form [poly(dG-dC)]2 [45], although rac-[Ru(phen)3]2+ and [Fe(phen)3]2+ have been shown to induce a Z→B transition in NaCl-induced Z-form [poly(dG-dme5C)]2 [36]. Thus far the enantioselectivities described have been with heterogeneous DNA. The same enantioselectivities are not necessarily observed with different sequences of B-DNA and such results are discussed further below (Sec. 3.1.4).
< previous page
page_208
next page >
< previous page
page_209
next page > Page 209
3.1.2 Influence of Metal, Charge, and Ligands on Binding Affinity For a small fraction of the metal complexes studied with DNA, binding affinity data are available. Generally, binding constants are determined for racemates although data are also available for certain enantiomers. Before considering specifically the mode of interaction of these complexes with DNA, it is worthwhile considering the influence that the coordination ligands, metal ion, and charge have on the affinities of the metal complexes for heterogeneous duplex DNA (Table 3). Numerous binding constants have been reported for homoleptic complexes {[M(L)x]n+} involving small ligands {L = bpy, phen, 5-NO2-phen}. With larger ligands {TMP, DIP, phi} the complexes become less soluble, making such determinations impossible. However, in some cases, the relative affinities of different complexes have been compared without the binding constants being absolutely determined, which provides some information to fill the gaps. Consulting Table 3, it is clear that tris(phenanthroline) complexes bind stronger than tris(bipyridyl) complexes. However, the differences in binding constants are not more than an order of magnitude and the tris(phenanthroline) complexes cannot be considered to be ''strong" binders. An interesting observation is that [Ru(TAP)3]2+ binds less strongly than [Ru(phen)3]2+ [160]. Since the dimensions of the complexes should be the same, this undoubtedly reflects the significant role that ligand hydrophobicity plays in stabilizing binding of these complexes. Also, when the phen ligand is functionalized with nitro groups, which could be involved in H bonding, the binding weakens to the level of [Ru(bpy)3]2+, showing again the importance of ligand hydrophobicity, i.e., if phen was already partially intercalated, the extension of the π system and the enhanced stacking ability conferred by NO2 would be expected to increase K. In the same vein of argument, binding of [Ru(phen)2(flone)]2+, the third ligand being a ketone, is of the same order as that of [Ru(phen)2(bpy)]2+ [54]. Although the binding constant for [Ru(DIP)3]2+ has not been determined, it seems to bind somewhat more strongly than [Ru(phen)3]2+, as judged by absorption, emission enhancement, polarization, and quenching experiments [43]. This may simply be due to the increased hydrophobicity of the DIP ligand rather than enhanced overlap of an
< previous page
page_209
next page >
< previous page
next page >
page_210
Page 210 TABLE 3 Binding Constants Reported for Metal Complexes with CT-DNA Dye
K (M1)
n (bp)
Buffer
Method
Analysis
Ref.
[M(L)x]n+ [Fe(bpy)3]2+
22 × 103 (13 × 105)a 1.11.4 × 103
[Co(bpy)3]2+ [Ru(bpy)3]2+
5.48.4 × 103 2 × 105 22 × 103 (14 × 105)a 6.8 × 102 0.7 × 103
[Os(bpy)3]2+
57.3 × 103
6.5 × 103 [Fe(bpy)3]3+
[Co(bpy)3]3+ [Os(bpy)3]3+
56.6 × 103
9.414 × 103 1725 × 103
nr (50) 34
3
1 mM cacodylate/10 mM NaCl (pH 7)
10 mM Tris/10 mM NaCl (pH 7.1) 5 mM Tris/50 mM NaCl (pH 7.1)
310
H2O (pH 5.26.5)
nr (25)
1 mM cacodylate/10 mM NaCl (pH 7)
nr 612 3
50 mM phosphate (pH 7) 5 mM Tris/50 mM NaCl (pH 7.5) 10 mM Tris/10 mM NaCl (pH 7)
3 10 mM phos/10 mM 34
3 3
McGhee (Scatchard)
35
CV
Scatchard
64
CV
Scatchard
64
Filtration
nr
Dialysis
McGhee (Scatchard)
35
Quenching
Scatchard
71
Dialysis
McGhee
54
CV
Scatchard
66
ECL
Scatchard
66
CV
Scatchard
64
CV
Scatchard
64
CV
Scatchard
66
138
(pH 7)
10 mM Tris/10 mM NaCl (pH 7.1) 5 mM Tris/50 mM NaCl (pH 7.1) 10 mM Tris/10 mM NaCl (pH 7)
Dialysis
(table continued on next page)
< previous page
page_210
next page >
< previous page
next page >
page_211
Page 211 (table continued from previous page) Dye [Fe(phen)3]2+
K (M1) 33 × 103 (1.8 × 105)a 715 × 103
[Co(phen)3]2+
2.8 × 104 3051 × 103
[Ru(phen)3]2+
∆: 4.9 × 104 ZΨ = 1.38 Λ: 2.8 × 104 ZΨ = 1.24 23 × 104 (4 × 104)a 8.1 × 103 (9 × 104)b 4.8 × 103 6.2 × 103 6.2 × 103 ZΨ = 2.2 3.1 × 103 5.5 × 103
n (bp) nr (3) 45 6.7 56 3.7
3.4
810
4 (10) nr 4 4c
4 4
Buffer 1 mM cacodylate/10 mM NaCl (pH 7)
5 mM Tris/50 mM NaCl (pH 7.1) 5 mM Tris/50 mM NaCl (pH 7.1) 5 mM Tris/50 mM NaCl (pH 7.1)
Method
Analysis
Ref.
Dialysis
McGhee (Scatchard)
35
CV
Scatchard
64
CV
Scatchard
63
CV
Scatchard
64
McGhee
87
McGhee
87
Dialysis
Scatchard (McGhee)
42
ECL
McGhee
65
Quenching
Scatchard
71
Dialysis
McGhee
134
Dialysis
McGhee
44
Dialysis
McGhee
54
Absorption
McGhee
54
5 mM Tris/10 mM NaCl (pH 7.1)
Dialysis & emission
5 mM Tris/10 mM NaCl (pH 7.1)
Dialysis & emission
1 mM cacodylate/10 mM NaCl (pH 7)
2 mM phosphate/25 mM C2O42 (pH 5) 50 mM phosphate (pH 7) 5 mM Tris/50 mM NaCl (pH 7.1) 5 mM Tris/50 mM NaCl (pH 7.2)
5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5)
(table continued on next page)
< previous page
page_211
next page >
< previous page
next page >
page_212
Page 212 Continued TABLE 3 Dye [Fe(phen)3]3+ [Co(phen)3]3+
K (M1) 715 × 103 6.2 × 103 5.8 × 103 1626 × 103
[Rh(phen)3]3+ [Ru(5-NO2-phen)3]2+
2.9 × 103 1.0 × 103
n (bp) 45 nr 6.7 56 nr 57
Buffer 5 mM Tris/50 mM NaCl (pH 7.1) 25 mM phosphate (pH 7) 5 mM Tris/50 mM NaCl (pH 7.1) 5 mM Tris/50 mM NaCl (pH 7.1) 50 mM phosphate (pH 7) 5 mM Tris/50 mM NaCl (pH 7.5)
Method
Analysis
Ref.
CV
Scatchard
64
Quenching
Scatchard
71
CV
Scatchard
63
CV
Scatchard
64
Quenching
Scatchard
71
Dialysis
McGhee
54
Dialysis
McGhee
54
Dialysis
McGhee
54
Dialysis
McGhee
164
Dialysis
McGhee
164
Dialysis
McGhee
164
[M(L)x(L')y]n+ [Ru(bpy)2(phen)]2+ [Ru(bpy)2(DIP)]2+ [Ru(bpy)2(qpy)]2+ [Ru(bpy)2(Meqpy)2+ [Ru(bpy)2(Me2qpy)]2+
7 × 102 1.7 × 103 1.3 × 103 1.4 × 103 2.8 × 103
1014 1218 23 2 3
5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.4) 5 mM Tris/50 mM NaCl (pH 7.4) 5 mM Tris/50 mM NaCl (pH 7.4)
(table continued on next page)
< previous page
page_212
next page >
< previous page
next page >
page_213
Page 213 (table continued from previous page) Dye [Ru(bpy)2(ppz)]2+
K (M1) 5.5 × 103 ∆: 2.1 × 104 Λ: 6.1 × 103
[Ru(bpy)2(dppz)]2+
>106 4.9 × 106
[Ru(bpy)2(phi)]2+
∆: 1.00 × 104 Λ: 0.86 × 104 16.0 × 104 4.8 × 104
[Ru(phen)2(Cl)2]+
4.3 × 102 1.5 × 102
[Ru(phen)2(bpy)]2+
2.4 × 103 5.5 × 103
[Ru(phen)2(flone)]2+ [Ru(phen)2(DIP)]2+
2.1 × 103 2.5 × 103
n (bp) 34 34 34 nr
Buffer 5 mM Tris/50 mM NaCl (pH 7.4) 5 mM Tris/50 mM NaCl (pH 7.4) 5 mM Tris/50 mM NaCl (pH 7.4) 5 mM Tris/50 mM NaCl (pH 7.0)
Method
Analysis
Ref.
Dialysis
McGhee
164
Emission
McGhee
46
Emission
McGhee
46
Dialysis
McGhee
125
1.7
50 mM phosphate (pH 7)
Emission
Scatchard
56
6
10 mM cacodylate (pH 7)
Absorption
Scatchard
39
6
10 mM cacodylate (pH 7)
Absorption
Scatchard
39
Dialysis
McGhee
54
Absportion
McGhee
54
Dialysis
Scatchard
157
Dialysis
Scatchard
157
Dialysis
McGhee
54
Absorption
McGhee
54
Dialysis
McGhee
54
Dialysis
McGhee
54
4 4 1c 1c 57 57 57 coop
5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 1 mM phosphate/2 mM NaCl (pH 7) 1 mM phosphate/2 mM NaCl (pH 7) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5)
(table continued on next page)
< previous page
page_213
next page >
< previous page
next page >
page_214
Page 214 Continued TABLE 3 Dye [Ru(phen)2(DIP)]2+ [Ru(phen)2(dppz)]2+
K (M1) 11.2 × 103 ∆: 6 × 107 ∆: 3.2 × 106 ZΨ = 1.9 Λ: 1.7 × 106 ZΨ = 2.1
[Ru(phen)2(phi)]2+
110 × 103 46.8 × 103 1.0 × 106
[Rh(phen)2(phi)]3+
≥107 4.9 × 106
[Ru(DIP)2(phen)]2+
10.1 × 103 11.1 × 103
n (bp) coop 2 3
3
23 23 0.9 nr 4.5 coop coop
Buffer 5 mM Tris/50 mM NaCl (pH 7.5) 1 mM cacodylate/10 mM NaCl (pH 7) 5 mM Tris/50 mM NaCl (pH 7.1)
5 mM Tris/50 mM NaCl (pH 7.1)
5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5)
Method
Analysis
Ref.
Absorption
McGhee
54
Emission
Estimate
21
Emission
Scatchard
86
Emission
Scatchard
86
Dialysis
McGhee
54
Absorption
McGhee
54
Refit [54]
Scatchard
56
Dialysis
Estimate
127
Refit [127]
Scatchard
56
Dialysis
McGhee
54
Absorption
McGhee
54
(table continued on next page)
< previous page
page_214
next page >
< previous page
next page >
page_215
Page 215 (table continued from previous page) Dye [Ru(phi)2(bpy)]2+
K (M1)
n (bp)
17.6 × 103 24.4 × 103
[Rh(phi)2(bpy)]3+
≥107
coop coop nr
Buffer 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5) 5 mM Tris/50 mM NaCl (pH 7.5)
Method Dialysis
Analysis Ref. McGhee
54
Absorption McGhee
54
Dialysis
Estimate
127
[M(L)x(L')y(L'')2]n+ [Ru(bpy)(tpy)(OH2)]2+
6.6 × 102 4.0 × 102 15 × 103
[Os(bpy)(tpy)(OH2)]2+ [Ru(bpy)(tpy)(OH2)]3+ [Ru(phen)(tpy)(OH2)]2+
5.5 × 102 5.4 × 103 78 × 103 3.7 × 103 3.9 × 103
[Ru(dppz)(tpy)(OH2)]2+
7.0 × 105
nr
50 mM phosphate (pH 7)
Quenching Scatchard
71
nr
50 mM phosphate (pH 7)
Emission
Scatchard
56
3c
50 mM phosphate (pH 7)
CV
Scatchard
67
nr
50 mM phosphate (pH 7)
Quenching Scatchard
71
3c
50 mM phosphate (pH 7)
CV
Scatchard
67
3c
50 mM phosphate (pH 7)
CV
Scatchard
67
nr
50 mM phosphate (pH 7)
Quenching Scatchard
88
nr
50 mM phosphate (pH 7)
Quenching Scatchard
71
nr
50 mM phosphate (pH 7)
Quenching Scatchard
88
1.9
50 mM phosphate (pH 7)
Quenching Scatchard
71
2
50 mM phosphate (pH 7)
Emission
Scatchard
56
2
50 mM phosphate (pH 7)
Absorption Scatchard
56
73 × 104 ZΨ = 1.32 7.0 × 105 6.2 × 105
aOriginal values given in parentheses were calculated according to Scatchard: the other values have been reestimated here according to McGhee/von Hippel: some differences are very large while others are not. bThe authors carried out the analysis using both Scatchard equation, which they found inadequate, and McGhee/von Hippel equation. Note that there is an order-of-magnitude difference between the binding constants obtained using the two approaches. cThe binding site size was constrained at the value shown to allow calculation of K. Note: nr, not reported; coop, cooperative binding.
< previous page
page_215
next page >
< previous page
next page >
page_216
Page 216 intercalated DIP ligand compared to phen, since it is not clear that the bulky DIP ligand could intercalate even if phen could. Barton and coworkers [161] reported the crystal structure of [Ru(DIP)3]2+, which shows the phenyl groups to be skewed since rotation of the phenyl groups within 40% of the phenanthroline plane resulted in unfavorable contacts. However, they still proposed that the DIP ligand could intercalate, although it would require base pair opening since the width of the ligand would be increased by 23 Å over that of phenanthroline. A model was suggested whereby the ligand partially intercalated from the major groove such that one phenyl group resided in the minor groove while the other remained in the major groove [161]. On the other hand, another homoleptic complex with hydrophobic ligands[Ru(TMP)3]2+does not bind to B-form DNA, although it has been shown to interact with [poly(dG)·poly(dC)] (ostensibly in the A form), and with A-form DNA-RNA hybrids and RNA-RNA duplexes [159,162]. The lack of binding to B-DNA may be related to steric factors and suggests the absence of a hydrophobic pocket large enough to accommodate this complex. Notable features of the A-form duplex (Table 4) are the extremely narrow and deep major groove, the wide and shallow minor groove, and a strong inclination of the base pairs [163]. Binding constants determined mainly by equilibrium dialysis in 5 mM Tris/50 mM NaCl have been reported by the Barton group [54,125] and others [86,164] for several series or partial series of the form [Ru(L)x(L')y]2+ (x + y = 3). Although most series are incomplete, we can TABLE 4 Dimensions (Å) of Different DNA Helical Conformations Minor groove
Major groove
DNA form
Pitch
Helical turn (bp)
Rise (Å)
Width
Depth
Width
Depth
B
33.8°
10
3.40
6.0
8.2
11.6
8.5
A
28.2°
11
2.56
11.1
2.6
2.2
13.0
ZI
44.6°
12
3.70
2.0
13.8
8.8
3.7
Source: Adapted with permission from [163].
< previous page
page_216
next page >
< previous page
page_217
next page > Page 217
group them into three nominal categories (L/L'): (1) bpy/phen, phen/flone, and phen/DIP show no significant change of ∆G° on replacement of one smaller ligand; (2) bpy/DIP and bpy/PPZ show a moderate (45%) increase of ∆G° on replacement of one ligand. The variation of binding data is graphically presented in terms of binding energies in Fig. 2. The boundaries between the categories are not clear-cut, since there is a clear trend for ∆G° to increase gradually with increasing size of the second ligand, which may be connected to hydrophobic considerations (Fig. 2C). However, since the evidence indicates that complexes with the DPPZ and phi ligands can intercalate, category (3) represents intercalators. Also, since it is generally accepted that [Ru(bpy)3]2+ does not intercalate, we suggest that category (1) represents nonintercalators: if phen did intercalate, we expect that changing a bpy for a phen would lead to a significant increase in ∆G°. In the only complete series, the biggest change in ∆G° occurs when two bpy ligands are replaced by phen (Fig. 2A). Category (2) represents ambiguous series which might belong to either of the other two categories. Since bpy/DIP is in category (2) and phen/DIP in category (1), it must be questioned whether DIP can in fact intercalate. The relatively small increase in ∆G° when bpy is replaced by PPZ compared to DPPZ might be either because PPZ is not intercalated or because of its relative hydrophilicity: this remains to be checked by viscosity. Although few ligands have been used in studies of binding constants for homoleptic complexes, several metals have been investigated. However, given that many different conditions, techniques, and models of analysis have been used, even within single studies, it is extremely difficult to compare results from different studies (Table 3). Using electrochemical techniques, it is possible to determine binding constants for different redox states of a complex (vide supra). For tris(bpy) complexes the trivalent complexes all have higher binding constants than the divalent complexes [64]. For tris(phen) complexes, however, the iron complex has the same affinity in its di-and trivalent states, while [Co(phen)3]2+ binds more strongly than [Co(phen)3]3+ [64]. This suggests that hydrophobic interactions could be more important than electrostatics in stabilizing binding. It is also notable that [Ru(phen)3]2+ and [Rh(phen)3]3+ have similar binding constants [71].
< previous page
page_217
next page >
< previous page
page_218
next page > Page 218
Fig. 2. Comparison of binding energies for ruthenium complexes [Ru(L)x(L')y]2+ bound to CT-DNA in 5 mM Tris/50 mM NaCl (pH ≈ 7), determined from data presented in Table 3 (from equilibrium dialysis determinations) assuming T = 298 K. For each series (L/L'), L is a smaller ligand than L'. In these graphs, the X axis represents ''y," the number of L' ligands in the complex. Panels A and B shows series where only small changes are observed on replacement of one L ligand with a larger one L'. In panel A, L = bpy and in Panel B, L = phen. Panel C shows the variation of binding energy for all series when only one ligand is changed, including some data already shown in panels A and B, in order to facilitate comparison. The role of electrostatic interactions cannot be dismissed, however, since the neutral metal complexes [Ru(L)2(CN)2], L = phen/bpy, do not show any significant binding to DNA [51,160]. This is also borne out when the larger intercalating phi ligand is involved, since Rh(III) complexes then have affinities more than an order of magnitude higher than those of Ru(II) complexes [54,127]. We do not consider in detail the [Ru(tpy)(OH2)(L)]2+ series, L = bpy, phen, DPPZ studied by Thorp and coworkers [56,67,71,88]. However, it may be noted that changing bpy to phen has little effect on the
< previous page
page_218
next page >
< previous page
page_219
next page > Page 219
< previous page
page_219
next page >
< previous page
page_220
next page > Page 220
binding constant while changing to DPPZ increases the affinity by about two orders of magnitude. 3.1.3 Intercalation vs. Groove Binding The question of how octahedral, and also tetrahedral, metal complexes interact with DNA at the microscopic level remains somewhat controversial. It is not yet clear whether a metal complex can really intercalate one of its ligands, when it is as small as bpy, tpy, or phen, or in which groove the interaction occurs. These ligands can be part of an intercalating metal complex since diffraction studies have demonstrated such for the square planar platinum complexes [146]. With octahedral and tetrahedral complexes, an extra consideration is whether steric clashes of the ancillary ligands with the surrounding DNA structure so hinder approach of the complex that the small ligand cannot extend into the stacked base pair stack. The least steric interference is expected if the metal complex approaches from the major groove (the wider groove in B-DNA), but strong binding of bulky drugs (actinomycin [150], chromomycin [165]) which distort the minor groove is not unprecedented. Given the right-handed helical nature of B-DNA, it is easy to imagine how steric hindrance could give rise to enantioselectivity for certain binding modes, such as that where one ligand intercalates and the other two form left-or right-handed propellers in the groove. While much of the debate about binding of octahedral complexes has centered on [Ru(phen)3]2+, it is really quite instructive to take a step back and consider how the properties of the smallest and simplest of these complexes, [Ru(bpy)3]2+, are altered when it binds to DNA. This complex has been largely ignored because it binds weakly to DNA. However, since it is widely accepted that the bpy ligands of this complex cannot intercalate, it is an excellent model for an electrostatic/groovebound metal complex. At the other end of the scale, the platinum compounds are good models for intercalative metal complexes and there are now octahedral metal complexes with an intercalated ligand, e.g., [Ru(phen)2(DPPZ)]2+ [21]. We begin by comparing the DNA binding of [Ru(bpy)3]2+ and [Ru(phen)3]2+. The luminescence of both [Ru(bpy)3]2+ and [Ru(phen)3]2+ is enhanced on binding to nucleic acids, although the enhancement with [Ru(phen)3]2+ is always larger (Table 5). In addition, for both dyes,
< previous page
page_220
next page >
< previous page
page_221
next page > Page 221
quenching of the excited state by oxygen is reduced when the dye is bound to nucleic acids, the inhibiting effect again being slightly greater with [Ru(phen)3]2+. On the basis of the similarity of photophysical changes upon binding, Görner et al. [20] proposed that [Ru(bpy)3]2+ and [Ru(phen)3]2+ bind in the same way, which they suggested was partial intercalation. However, it is now apparent from the work of Hiort et al. [42] and Satyanarayana et al. [27,87] that [Ru(phen)3]2+ is probably not intercalated (vide infra) and the similarity may reflect both complexes being groove-bound. The two complexes behave similarly in other respects also. For instance, recent results in our lab indicate that the counterion release parameter ℜ from Eq. (8) is approximately the same for the two complexes [Lincoln and Nordén, to be published]. Furthermore, for each enantiomer very similar average binding geometries are found but with a somewhat less ordered structure for [Ru(bpy)3]2+. Finally, for both complexes, the ∆ enantiomers reduce the persistence length of DNA [Lincoln and Nordén, to be published]. These results appear to contradict those of Kelly and coworkers, who reported that [Ru(bpy)3]2+ does not unwind supercoiled DNA when [Ru(phen)3]2+ does, and that [Ru(phen)3]2+ affects the melting curve of [poly(dA-dT)]2 in a manner reminiscent of ethidium, while the effect of [Ru(bpy)3]2+ more closely resembles that of Mg2+ [124]. However, these differences might be explained if much less [Ru(bpy)3]2+ than [Ru(phen)3]2+ was bound under the conditions used. These complexes can also interact with ss-DNA under low ionic strength conditions in a manner which appears to be purely electrostatic [20]. The effects of ss-DNA on the absorption and emission properties of the complexes are smaller than those of ds-DNA [20,138] and they are very sensitive to salt [138]. In addition, compared with polyelectrolytes, the effects of ss-DNA are quite like those of NaPA, where binding is purely electrostatic, whereas the effects of ds-DNA are more like those of NaPSS, where hydrophobic as well as electrostatic interactions contribute to the stabilization of binding [166]. This indicates that hydrophobic interactions are important for binding of these complexes to ds-DNA, as already concluded from comparisons of binding constants (vide supra). Furthermore, the similarity between the properties of the dyes bound to ds-DNA and NaPSS shows that it is not necessary to invoke stacking interactions with nucleobases to explain the effects of DNA on the photophysical properties of the complexes. Thus, having established that the properties of [Ru(bpy)3]2+ and
< previous page
page_221
next page >
< previous page
next page >
page_222
Page 222 TABLE 5 Effects of Nucleic Acids on the Emission Lifetimes (nsec) and on the Rate Constant for Quenching of the Excited States of Simple Ruthenium Complexes by Oxygen Complex
Polymer
[Ru(bpy)3]2+ Nonea
τoxygen (ns)
τair (ns)
τargon (ns)
kq (O2) M1 s1
150
370
580
3.8 × 109
580
700
CT-DNA [poly(dA-dT)]2
450
580
700
0.6 × 109
[poly(dG-dC)]2
330
530
700
1.3 × 109
Noneb
115
370
590
3.9 × 109
CT-DNA
460
600
700
Nonec
140
360
550
3.85 × 109
NaPSS
380
665
760
0.9 × 109
Noned
180
420
640
2.9 × 109
NaPA
220
450
600
2.2 × 109
NaPSS
440
680
850
0.8 × 109
(table continued on next page)
< previous page
page_222
next page >
< previous page
next page >
page_223
Page 223 (table continued from previous page) τoxygen (ns)
τair (ns)
τargon (ns)
kq (O2) M1 s1
Nonea
150
470
960
4.3 × 109
CT-DNA
700
1300
1450
[poly(dA-dT)]2
740
1350
1450
0.5 × 109
[poly(dG-dC)]2
450
800
1000
0.9 × 109
Noneb
160
430
900
4.1 × 109
CT-DNA
680
100
1300
Complex [Ru(phen)3]2+
Polymer
ss-DNA
460
Nonee
525 (530)
CT-DNA
733/2645 (630/2300)
Nonef,g
520f
905g
CT-DNA
650/1800f
880/1980g
Noned
180
530
1190
3.5 × 109
NaPA
230
600
1020
2.4 × 109
NaPSS
350/1050
550/1560
770/1690
1.0/0.2 × 109
a3mM phosphate buffer (pH 6.9): [160] does not state O2 concentrations. bH2O: [20] does not state O2 concentrations. cH2O: [E. Tuite et al., to be published] 0, 0.3, 1.4 mM O2. dH2O: [165] 0, 0.28, 1.35 mM O2. e5 mM Tris/50 mM NaCl (pH 7.2): [43] (similar τ values from [54], in parentheses). f5 mM Tris/50 mM NaCl (pH 7): [25]. e5 mM Tris/50 mM NaCl (pH 7): [210].
< previous page
page_223
next page >
< previous page
page_224
next page > Page 224
[Ru(phen)3]2+ are very similar when bound to DNA, we now deliberate the mode of binding of [Ru(phen)3]2+. Several models have already been proposed, as recently summarized by Satyanarayana et al. [27]. Until very recently, it was widely accepted that [Ru(phen)3]2+ partially intercalates. Barton and coworkers postulated a dual binding mode model where the complexes can both partially intercalate in the major groove and ''surface-bind" in the minor groove: surface-bound in this context does not mean loose electrostatic attraction but a specific interaction with the groove. On the basis of polarization [44] and NMR data [57], it was suggested that the ∆ enantiomer preferred the intercalative mode and the Λ enantiomer preferred the surface-bound mode. Early ELD studies had previously suggested that only the ∆ enantiomer was partially intercalated [37,38], but partial intercalation was first seriously questioned by FLD studies presented by Hiort et al. [42]. These results suggested a model in which neither enantiomer is intercalated. In addition, isodichroic points as a function of binding ratio and ionic strength indicated that two binding modes for each enantiomer were highly unlikely. The model suggested different binding geometries for ∆ and Λ, but recent reanalysis (Sec. 4) suggests that both enantiomers bind quite similarly. Subsequent studies by Satyanarayana et al. [27,87] showed how much of the work of Barton and coworkers could be explained in terms of free dye and only one form of bound dye. Their viscosity data showed that DNA was not lengthened by [Ru(phen)3]2+, which excluded intercalation [27,87]. DNA unwinding by [Ru(phen)3]2+ [54,124] presumably results from some structural distortion of DNA (Sec. 2.8). The best evidence for a second surface-bound mode comes from the quenching data of Orellana et al. [25,28], but this is considered to be electrostatic rather than groove binding. Indeed, since the photophysical properties of Barton's proposed "surface-bound" [Ru(phen)3]2+ are very similar to those of free complex, then this must be loose binding since the properties of [Ru(bpy)3]2+, presumably groove-bound, are significantly altered by DNA. However, such a binding mode is expected to be weak and favored only at high binding ratios and low ionic strength. Since [Ru(phen)3]2+ was found to bind only very weakly to T4 DNA which is glucosylated in the major groove, Barton and coworkers proposed major groove binding for this complex [44]. Molecular dynamic simulations also suggested a preference for both enantiomers to bind by
< previous page
page_224
next page >
< previous page
page_225
next page > Page 225
partial insertion in the major groove [167]. However, 2-D NMR has revealed contacts between both enantiomers of [Ru(phen)3]2+ and the minor groove of DNA [60,61]. In view of the effects of [Ru(phen)3]2+ binding on DNA, observed by FLD [42] and viscosity [27,28], binding like actinomycin [150] in the minor groove with distortion is possible. Thus, it appears that T4 may block not only major groove binders but also minor groove binders, perhaps by preventing necessary opening of the groove. However, a recent 2-D NMR study of ∆-[Ru(phen)2(DPPZ)]2+ showed a contact in the major groove of a hexamer [62]. Hence, since we believe that all of these complexes have similar binding geometries (Sec. 4), major groove interactions cannot be discounted for any of them. In fact, it is possible, given the small binding sites (34 bp) for these complexes, that binding interactions of approximately equal strength could occur in both grooves. 3.1.4 Sequence Selectivity and Sequence-Related Enantioselectivity It is a general observation that intercalators prefer GC sequences and groove binders AT sequences, although opposite selectivities are also known. Consistent with this, the intercalating Pt complexes are GC-specific [75,77] as are N-alkyl-1,10-phenanthrolines [168]. On the other hand, judged by cleavage efficiency, the ''tetrahedral" [Cu(phen)2]+ is AT-selective with particular preference for 3'-AT-5' sites and also 3'-TG-5' sites but low reactivity at 3'-CG-5' sites. However, the planar [Cu(phen)]+ shows more GC selectivity of cleavage [169,170]. Since these complexes have been shown to be intercalated by viscosity [112], this indicates that the ancillary nonintercalated ligand is responsible for the change of sequence selectivity. In [Cu(phen)2]+ the dihedral angle between the rings can vary between 50° and 90°, while in [Cu(dmp)2]+, which does not intercalate, it is constrained at 90° [112]. It is not known unequivocally in which groove the complexes bind. Veal and Rill [112] suggest minor groove while Tamilarasan and McMillin suggest major groove, principally on the basis that there is no binding with T4 DNA [171]. However, a similar result was misleading for ruthenium complexes (vide supra), so we believe that this question remains to be answered. Veal and Rill suggest that the AT selectivity has two causes, viz., unfavorable steric hindrance with the guanine N2-
< previous page
page_225
next page >
< previous page
page_226
next page > Page 226
amino groups and favorable van der Waals interactions between partial charges on the phenanthroline and adenine moieties [112]. They further suggest that the cooperativity they observe could be explained if [Cu(phen)2]+ binds in the minor groove, thus distorting the DNA and facilitating binding of a second complex. The sequence selectivity of [Ru(phen)3]2+ is more complicated since it is different for the two enantiomers and also depends on ionic strength. An AT selectivity for the racemate with CT-DNA was apparent for [Ru(phen)3]2+ in 12 mM phosphate/3 mM NaCl buffer since emission enhancement is greater with [poly(dA-dT)]2 than [poly(dG-dC)]2 [124]. However, Barton et al. [44] found no apparent selectivity of the racemate with CT-DNA in 5 mM Tris/50 mM NaCl, although selectivities were observed for binding of the enantiomers. Dialysis showed that the ∆ enantiomer was preferred with DNA having more than 40% GC while the Λ enantiomer was preferred at lower GC contents. Calculation of selectivity on the basis of polarization data gave different results and suggested ∆ selectivity at all GC contents except 0%, where no enantioselectivity was apparent. Λ had lower polarization than ∆, which was rationalized in terms of the Λ enantiomer preferring the ''surface-bound" and not contributing significantly to polarization. However, if the enantiomers are bound with essentially the same geometry, as we propose (Sec. 4), a lower polarization for Λ could result from this enantiomer being able to shift more in its binding site during its long lifetime. If the data were corrected for the difference in enantiomeric polarizations, then the selectivity by polarization would probably resemble that determined by dialysis. The salt dependence of the enantioselectivity was determined by Barton et al. [44] for CT-DNA (42% GC), Clostridium perfringens DNA (26% GC), and Micrococcus lysodeikticus (74% GC). With 26% GC, the Λ enantiomer is always preferred but there was a slight decrease in Λ selectivity with increasing salt. With 74% GC, the ∆ enantiomer was always preferred, with the selectivity increasing slightly with salt. With CT-DNA, the selectivity showed a marked variation with salt, from Λ selectivity at very low salt to strong ∆ selectivity at high salt.K Similar experiments were carried out by Hiort et al. [42] who also examined the effect of binding ratio on the enantioselectivity. The ∆ enantioselectivity at high GC content and Λ enantioselectivity at low GC content were reproduced, albeit with different magnitudes, as
< previous page
page_226
next page >
< previous page
page_227
next page > Page 227
was the tendency for increased GC selectivity with increasing ionic strength. The respective preferentialities are more marked at high compared to low binding ratios. At low binding ratios (≈0.05), the order of binding affinities was:
At higher binding ratios, ∆/DNA became weaker and Λ/DNA stronger. The racemate showed a markedly weaker binding to CT-DNA than either of the enantiomers and it was suggested that the Λ enantiomer bound more weakly to DNA perturbed by simultaneous binding of the ∆ enantiomer, i.e., homochiral cooperativity. Each of the enantiomers had essentially identical LD spectra with DNA, [poly(dA-dT)]2, and [poly(dG-dC)]2, indicating similar binding geometries in all cases. Finally, the sequence-related enantioselectivity has also been investigated by Satyanarayana et al. [27] using competition dialysis: ∆-[Ru(phen)3]2+ interacted preferentially with GC-rich DNA and Λ-[Ru(phen)3]2+ with ATrich DNA. For both enantiomers, the selectivity was quantified in terms of a requirement for a single preferred base pair at the binding site. The sequence selectivity of each enantiomer was judged to be low compared with sequencespecific binders such as daunomycin and actinomycin [27]. A requirement for intercalation 5'-to a G base has been reported for the intercalator [Pt(tpy)-(HET)]+ [75]. Molecular dynamics modeling [172] suggested a preference of Λ[Ru(phen)3]2+ for pur-3',5'-pyr sites with A-3',5'-T favored over G-3',5'-C (for partial insertion from the major groove), but detailed experimental studies have not been carried out for comparison. Many studies have established the same pattern of sequence-related enantioselectivity and it should be possible to use this information to elucidate the binding of [Ru(phen)3]2+ enantiomers. However, it is still not clear in which groove the complexes bind. Perhaps modeling of the sequence selectivities could provide some insight on this matter. 3.2 Interactions with Z-DNA and Other Unusual DNA Conformations Z-DNA is a left-handed conformation of DNA, adopted most readily by GC sequences, which can be induced in the presence of Na+, Mg2+,
< previous page
page_227
next page >
< previous page
page_228
next page > Page 228
spermine, or tricationic metal hexammine complexes, and is stabilized in underwound, negatively supercoiled, closedcircular DNAs. Clearly, Z-DNA is not a left-handed mirror image of B-DNA (Table 4). It has a narrow, deep, minor groove lined with phosphate groups and a wide, shallow, major groove; the edges of the base pairs form the major groove, which has a slightly convex surface. Unlike A or B helices, the repeat unit is a dinucleotide. Electrophilic addition at C8 of guanine stabilizes Z-DNA and promotes a B → Z structural transition. By contrast, the binding of intercalators and minor groove binders to Z-DNA generally promotes a Z → B transition. However, exceptions have been reported by the Barton group who suggest that tris(phen) and tris(DIP) metal complexes bind enantioselectively to Z-DNA without inducing transitions to B-DNA [45]. In fact, there have been several contradictory reports about the binding of metal complexes to Z-DNA. The earliest studies [134] indicated a slight preference of ∆-[Ru(phen)3]2+ for B-form CT-DNA (vide supra), but no enantiopreferentiality for binding to Z-form [poly(dG-dC)]2. More dramatically, it was found that ∆-but not Λ[Ru(DIP)3]2+ bound to B-form CT-DNA while both enantiomers bound similarly to Z-form [poly(dG-dC)]2 [153]. It was reasonably suggested that the major groove of Z-DNA should easily accommodate both enantiomers since the backbone is directed to the minor groove. However, later studies [45] suggested that, for both [Ru(phen)3]2+ and [Ru(DIP)3]2+, the Λ enantiomers were preferred with Z-[poly(dG-dC)]2. In fact, it was found that ∆[Ru(DIP)3]2+ did not interact significantly with Z[poly(dG-dC)]2 and Λ-[Ru(DIP)3]2+ did not interact with B-[poly(dG-dC)]2. The low concentration of [Co(NH3)6]3+ used to induce the Z form in these experiments might underlie differences with previous work (perhaps high cation concentrations induce Z-DNA aggregation and the poorly soluble, hydrophobic complexes condense with aggregates in a nonselective manner). However, considering the Z-DNA structure and previous arguments, it is difficult to see why only Λ-[Ru(DIP)3]2+ is bound to Z-DNA. The behavior of Λ-[Ru(DIP)3]2+ suggested that it could be used as a probe for Z-DNA structures; however, subsequent studies with circular DNA indicated that this dye also had interactions with non-Z-form structures (vide infra). Given the absence of a Z → B transition, comparison was made
< previous page
page_228
next page >
< previous page
page_229
next page > Page 229
[45] with chromomycin A3 for which this is also the case [173]. Chromomycin had earlier been reported to bind in the major groove [174] but subsequent work had demonstrated binding as a dimer in the minor groove [165,175]. Thus, to turn the original argument around, this supports minorrather than majorgroove binding of the metal complexes. In direct contrast to the results of Barton and coworkers, it has been clearly shown, using CD spectroscopy, that [Ru(phen)3]2+ and [Fe(phen)3]2+ do induce a Z → B conformational transition when they bind to [poly(dGdme5C)]2 under conditions where the Z conformation is induced by NaCl [36]. This is consistent with the behavior of most groove binders and intercalators. Using competition dialysis, Satyanarayana et al. compared the binding of [Ru(phen)3]2+ enantiomers to [poly(dGdme5C)]2 and [poly(dG-dC)]2 [27]. It was found that in both B-and Z-forming buffers, the complex had a higher affinity for the methylated polymer, but there was no significant selectivity of either enantiomer in different buffers. These later studies strongly suggest that the behavior observed by Barton and coworkers is associated with the use of [Co(NH3)6]3+ to induce the Z form rather than being a special property of ruthenium polypyridyl complexes. It is now apparent that DNA can adopt a wide range of conformations, and it has been reported by the Barton group that tris(DIP) metal complexes can be used to recognize unusual secondary structures in natural DNAs [2,5]. The ability of these metal complexes to cleave DNA in situ has been exploited to map the sites of their interaction. Λ[Co(DIP)3]3+ cleaves at GC inserts in plasmids, but also at other sites found to be close to some but not all alternating pur/pyr sequences [176178], with a tendency to cleave at junctions between pur/pyr and homopurine sequences [178]. [Rh(DIP)3]3+ was also found to cleave near cruciform sites [179] and within introns [180]; it appears that functionally important sites are specifically targeted. However, it is not certain that the metal complexes recognize unusual conformations. Lower levels of cleavage were also observed in some relaxed and linearized plasmids [177,178], and the pattern of S1 digestion of pBr322 was significantly different in the absence and presence of [Rh(DIP)3]3+ [179]. In fact, these results suggest that the metal complexes might bind
< previous page
page_229
next page >
< previous page
page_230
next page > Page 230
to sites in the plasmids which are easily distorted into unusual structures rather than recognizing such conformations. 3.3 Sequence-Specific Binding: Oligonucleotide-Tethered Complexes Covalent attachment of a dye to an oligonucleotide allows specific binding to DNA with high affinity. A modified oligonucleotide can target ss-RNA or ss-DNA (antisense) and/or ds-DNA (antigene). In this way, sequence-specific fluorescent probes and cleaving agents (''artificial nucleases") can be assembled. With such applications in mind, common fluorophores such as fluorescein and rhodamine, and cleaving agents such as [Cu(phen)2]+ and Fe-EDTA complexes, have been attached to oligonucleotides [181]. Binding of intercalators (ethidium, acridines) increases the stability of DNA-oligonucleotide hybrids and the effect is enhanced by covalently attaching the dye to the oligonucleotide [181]. Additional long-term stability of the hybrids can be induced by crosslinking the substrate and oligonucleotide strands, e.g., with covalently attached psoralens and platinum complexes [181]. These approaches are also valid with modified oligonucleotides (methylphosphonates, phosphorothioates). However, it was recently found that intercalators, including [Ru(phen)2(DPPZ)]2+, bound only weakly to conjugates of DNA with the promising analog peptide nucleic acid (PNA) [182]. Thus, it is of interest to covalently attach metal complexes to oligonucleotides. Complexes containing DPPZ are emissive probes [21,56,125]; other complexes are cleaving agents, e.g., [Ru(phen)3]2+ [183,184] and [Ru(bpz)2(DPPZ)]2+ [185]; and complexes which form covalent complexes [141] have potential as crosslinking agents. Several approaches have been made to the attachment of ruthenium complexes to oligonucleotides, and condensed and solid phase synthetic pathways for attachment of phenanthroline and analogous ligands to both the 3' and 5' ends and to bases have been published [181,186193]. Hence it has been possible to use oligonucleotidetethered ruthenium complexes as luminescence reporters [186188,192], in fluorescence resonance energy transfer (FRET) studies [189,190], and as sequence-specific cleaving agents [194].
< previous page
page_230
next page >
< previous page
page_231
next page > Page 231
3.4 Photoinduced Electron Transfer Reactions in the Presence of DNA The question of whether DNA can mediate long-range electron transfer has been addressed inconclusively by radiation chemists for decades with conclusions varying from 23 base pairs to >100 base pairs [195]. Such experiments are generally carried out on frozen samples and are likely incomparable with solution work. However, it seems probable that either the frozen water shell around the DNA or the sugar-phosphate backbone might act as electron ''conductor," rather than the π system of the stacked bases. DNA has been reported to enhance the emission quenching of *[Ru(phen)3]2+ and *[Ru(bpy)3]2+ enantiomers by rac-[Co(bpy)3]3+, rac-[Co(phen)3]3+, rac-[Cr(phen)3]3+, and rac-[Rh(phen)3]3+ [196,197]. Estimates of rate constants were made using the bulk quencher concentration, unadjusted for condensation by the polymer. Hence the increase of the apparent quenching rate when the compounds are bound to DNA is not surprising since other polyanions induce such effects, which are generally attributed to the high local concentrations of reactants [10]. It was suggested that the reduced dimensionality of the DNA could be partially responsible for the enhanced rates, and also that the π framework might be responsible for mediation of long-range electron transfer, since quenching was also observed at 253 K in 90% glycerol [197], which should eliminate diffusion. Subsequent work by Orellana et al. [28] clarified that the apparent enhancement of quenching by DNA in these systems is probably due to concentrating effects. Indeed, when local concentrating effects were corrected for, it was concluded that the actual rate of the electron transfer step was slower in the presence of DNA than in aqueous solution [28], as was also concluded for the apparently enhanced quenching of intercalated ethidium by methylviologen [198]. Furthermore, decreased mobility of the metal complexes bound to DNA was apparent [28] (restricted diffusion of methylviologen (MV2+) along DNA has also been reported [198]), and the interpretation of the work in viscous solution [197] was left open [28] until experiments to test the reliability of the data were executed. Newer work has made claims that the stacked π system of the DNA base pairs ("π way") can mediate long-range ultrafast (
< previous page
page_232
next page > Page 232
electron transfer between intercalated ligands of metal complexes [199,200]. These studies used *rac[Ru(phen)2(DPPZ)]2+ and rac-[Rh(phi)2(phen)]3+, where each reactant should have one ligand intercalated, in contrast to previous work using tris(phen) complexes, which do not have fully intercalated ligands. However, these reports provide no spectroscopic evidence for electron transfer products and conclusions are based on quenching of the emission of *[Ru(phen)2(DPPZ)]2+; but this complex is also nonemissive in solution and there is thus some question as to whether it is truly being quenched or simply being displaced from DNA. Experimental evidence to exclude the latter possibility has not been presented, although recent work with a similar system has revealed transients on the nsec and µsec time scales when the Rh(III) quencher is in excess [201]. Evidence from another lab [202] indicated that DNA can also mediate electron transfer between covalently attached nonintercalated metal complexes with a rate (on the nsec time scale) comparable to the most efficient protein systems. Recent data from our laboratory investigating the quenching of *[Ru(phen)2(DPPZ)]2+ enantiomers by intercalating and nonintercalating viologens have not shown any ultrafast or particularly long-range electron transfer mediated by the π system. In agreement with other studies [203], we find that electron transfer might take place over a separation of perhaps 36 base pairs, but not over large distances. Using LD spectroscopy, we have been able to show that the reactants bind independently and that the observed quenching is unlikely to arise from displacement of the ruthenium complex. 4 Diastereomeric Binding Geometries Studied with Polarized Spectroscopy Although no detailed three-dimensional structure has yet been solved for any chiral metal complex bound to a DNA fragment, the direct and indirect structural evidence from spectroscopic and thermodynamic studies allows several important conclusions to be drawn about binding geometries and differences in behavior between enantiomeric forms. The clearest picture has been obtained for the [Ru(phen)2(DPPZ)]2+ complex, which we shall discuss in some detail in this section.
< previous page
page_232
next page >
< previous page
page_233
next page > Page 233
4.1 Polarized Spectroscopy of Ruthenium Complexes Of any electronic spectroscopic technique, LD spectroscopy of DNA ensembles with enantiomeric pairs of ruthenium complexes most distinctly shows up diastereomeric effects [21,42]. Λ enantiomers invariably show a much more positive LD at the MLCT band center (420 nm) than their ∆ counterparts when bound to flow-oriented DNA. Binding to DNA also induces noticeable changes in the CD spectra of ruthenium complex enantiomers, but the structural implications here are less clear-cut due to the strong inherent CD which is not yet completely understood. Differences in normal (isotropic) absorption spectra between pairs of DNA-bound enantiomers are in general quite small, although there is a clear trend for slightly more hypochromism in the MLCT band of ∆ complexes. The LD spectrum, normalized to perfect orientation, is related to the angles α(i) that the individual transition moments i (with absorbance εi) of the complex make with the helix (orientation) axis as
where Aiso is the isotropic absorption of the unoriented sample. Differences in the enantiomeric LD spectra are thus direct consequences of the diastereomeric binding geometries of the complexes. However, in order to extract geometrical information, both the polarizations of the transitions within the complex and the extent of overlap of differently polarized absorption bands must be known. The latter may in part be accomplished by measuring the photoselection emission anisotropy (FA) as a function of excitation wavelength for Ru complexes bound to DNA in very viscous buffer (saturated sucrose, 20°C); the former leads us now to a short excursion through the properties of the MLCT chromophore. 4.2 Spectroscopy of Ru-phen Complexes The characteristic orange color of Ru complexes with bipyridyl-type ligands derives from intense d-π* MLCT transitions in the 400to 500- nm region [8,9]. For [Ru(bipy)3]2+, most of the intensity has been shown to be (degenerately) polarized in the plane perpendicular to the C3 axis
< previous page
page_233
next page >
< previous page
page_234
next page > Page 234
(E transitions) with a smaller fraction of intensity polarized parallel to the C3 axis (A2 transitions). The ultraviolet (UV) region is dominated by intraligand π-π* transitions which give rise to a strong exciton component with A2 symmetry occurring near 260 nm. The MLCT photophysics is characterized by a long-lived (msec) excited state exhibiting an E-polarized luminescence around 600 nm [204]. [Ru(phen)2(DPPZ)]2+ has received particular attention since upon binding to DNA it recovers emission which is completely quenched in aqueous solution [125]. The quenching mechanism is reported to involve protonation of one of the pyrazine nitrogens of the excited state DPPZ ligand anion radical. Hence, the increased luminescence of [Ru(phen)2(DPPZ)]2+ when bound to DNA is probably due to a more or less efficient shielding from solvent at the binding site of the DPPZ ligand. Upon an uniaxial perturbation along one of the C2 axes of [Ru(phen)3]2+, either by substitution (phen to DPPZ ligand) or by interaction with DNA, the degenerate E transitions will split into two sets of transitionsA and B(E)which are mutually orthogonal but still effectively polarized in the original E plane. These two major polarizations of the visible absorption band provide the main spectroscopic handles for the geometrical determination. In addition, we observe transitions of A2 origin, (B(A2)), and, for the DPPZ complex, additional DPPZ ligand short-axis polarized transitions (B(sh)). Figure 3 depicts the polarizations in a model of the proposed binding geometries of ∆- and Λ-[Ru(phen)2(DPPZ)]2+. The assignment of the polarizations is supported by intermediate neglect of differential overlap/spectroscopic (INDO/S) calculations in our laboratory (Broo et al., to be published). 4.3 Binding Geometries of [Ru(phen)2(DPPZ)]2+ and [Ru(phen)2(BDPPZ)]2+ [Ru(phen)2(BDPPZ)]2+ shows a strongly negative LD at 320 nm, where an intense π-π* transition polarized along the long axis of the BDPPZ ligand absorbs. The magnitude of the LDr shows that the long axis of the BDPPZ ligand is oriented perpendicular to the DNA helix axis. A strong hypochromicity, estimated at more than 40% for this transition, indicates that the BDPPZ chromophore is in intimate contact with the π system of the DNA bases. The DPPZ complex, with a slightly shorter ligand, also binds to
< previous page
page_234
next page >
< previous page
page_235
next page > Page 235
Fig. 3. Schematic diagram of the DNA binding geometries for the [Ru(phen)2-(DPPZ)]2+ enantiomers having the roll angles determined from linear dichroism spectra. The complex is shown bound in the minor groove. White lines indicate the orientation of the plane of the intercalating DPPZ ligand. The B-polarized transitions are depicted and the A polarization is normal to the page. DNA with the C2 axis perpendicular to the helix axis, which is concluded from comparisons of the FA and LDr curves (Fig. 4A). The perpendicular orientation of the C2 axis makes it convenient to define the other angular parameters determinable by LD as a rotation around this axis. Such rotation is described by the angle β between the DNA helix axis and the normal of the plane of the ligand containing the twofold axis, viewed from ligand to DNA. The value of β will thus be zero for an ''ideal" intercalation geometry. 4.4 Roll Angle If the roll angle β had been exactly zero for [Ru(phen)2(DPPZ)]2+, the angles between, for instance, a B(E)polarized transition moment and the DNA axis would have been +35° and 35° for the ∆ and Λ enan-
< previous page
page_235
next page >
< previous page
page_236
next page > Page 236
Fig. 4. Emission anisotropy {FA = (Iv Ih)/(Iv + 2Ih)} and reduced linear dichroism (LDr) for ∆-[Ru(phen)2(DPPZ)]2+ (upper panel, A) and ∆-[Ru(phen)3]2+ (lower panel, B). Since the FA is very similar for the ∆ and Λ enantiomers, only the curves for ∆ are shown. FA was measured on samples immobilized in saturated sucrose solution at low temperature. The FA values have been scaled by a factor of 7.5 to give them the same theoretical limits (+3 and 1.5) as the reduced LD.
< previous page
page_236
next page >
< previous page
page_237
next page > Page 237
tiomer, respectively. This would have resulted in identical LD spectra for the two enantiomers. Hence, the observation of substantially different LD spectra for ∆- and Λ-[Ru(phen)2(DPPZ)]2+ implies a nonzero roll for at least one of the enantiomers. An approximate ''reflection symmetry" between the LDr and FA profiles (Fig. 4A) of ∆-[Ru(phen)2(DPPZ)]2+ must be due to the helical axis of DNA roughly bisecting the angle between directions of the two major transition moments, B(E) and B(A2), that are perpendicular to the C2 axis. This rather fortuitous circumstance enables accurate determination of the roll angle for the ∆-[Ru(phen)2(DPPZ)]2+ complex (Table 6). The sign of the roll is unambiguously determined due to the identification of a B(sh) transition with strongly negative LD. For Λ[Ru(phen)2(DPPZ)]2+, the roll angle was determined assuming a negligible A2 intensity in the low-energy part of the MLCT absorption region, which is supported by polarized crystal spectra [204] and INDO/S calculations referred to above. As seen in Table 6, the roll angles are of similar magnitude and all are positive. An apparent paradox with regard to the large differences in the LD spectra mentioned above is presented by the very similar roll TABLE 6 Angular (degrees) Parameters for [Ru(phen)2(L)]2+ Enantiomers Bound to CT-DNA ∆ L phen
Λ
αa
βa
α
β
+47
+12
26
+9
(47)
(82)
(+26)
(+61)
DPPZ
+42
+7
22
+13
BDPPZ
+40
+5
28
+7
aα = angle between helix axis and B(E) transitions; β = roll angle as described in text. Uncertainties in angles are approximately ±3° for ∆ and ±5° for Λ enantiomers.
< previous page
page_237
next page >
< previous page
page_238
next page > Page 238
angles for the two enantiomers. The explanation is that a clockwise roll will correspond to directions of the B(E)polarized intensity at, respectively, the angles (35° +β) and (35° β) relative to the helix axis. This difference of ±β is found to account for the observed LD difference. Within experimental error, the roll geometries of the two enantiomers are thus the same; the different LD spectra are merely a consequence of the spectroscopic diastereomerism due to the different zorientations of the dissymmetrical part, the two phenanthroline propeller blades, relative to the DNA helix axis. The fact that we obtain angles around +10° for all [Ru(phen)2-(DPPZ)]2+-like derivatives studied, with the plus sign representing a clockwise roll when viewing toward the DNA, indicates a significant deviation in this direction, due either to the geometry of the intercalator complex itself or to a roll which is intrinsic to the DNA structure. It is instructive to consider the metal complex as composed of one symmetrical and one dissymmetrical part, viz., the ligand containing the C2-axis (i.e., DPPZ) and the two phenanthroline ''propeller blades." Only the latter, dissymmetrical part gives rise to an intrinsically diastereomeric interaction. However, the roll would also modulate this interaction so that, for example, for ∆-[Ru(phen)2(DPPZ)]2+, β = +10° would permit the propeller blades to have their longest dimension parallel to the groove and thereby minimize the radial distance to the helix center (maximum penetration). For the Λ enantiomer, β would have to be +80° for such an arrangement and since this is, of course, impossible with DPPZ intercalated, this represents one determinant of chiral discrimination. However, as we have seen, neither the geometries nor the thermodynamic stabilities indicate any drastic structural or thermodynamic differences between the DNA complexes with the two enantiomers of [Ru(phen)2(DPPZ)]2+. Differences are mainly manifest in the optical properties, on the one hand as a consequence of diastereomeric orientation effects on the optical tensors (seen in LD and CD) and on the other as a consequence of different penetrations (seen in emission and hypochromicity). At the same time, clear differences in fitting suggest that although the actual variations in binding thermodynamics appear to be somewhat subtle in this particular case, the machinery for enantiomeric selection does exist.
< previous page
page_238
next page >
< previous page
page_239
next page > Page 239
We have noticed a striking similarity between the binding geometry of [Ru(phen)2(DPPZ)]2+ and that of actinomycin D as found in the first crystal structure of a larger intercalator in a long DNA sequence [150]. Actinomycin D is present in the enantiomeric form that corresponds to the ∆ enantiomer of [Ru(phen)2(DPPZ)]2+. Furthermore, this drug has one planar aromatic three-ring system which is intercalated and two peptide propeller wings that correspond, respectively, to the symmetrical (DPPZ) and dissymmetrical (two phens) moieties of [Ru(phen)2(DPPZ)]2+. Actinomycin binds from the minor groove and steric interference of the peptide chelate wings gives rise to a distinct bend of the helix. 4.5 Binding Geometry of [Ru(phen)3]2+ Figure 4B shows the FA and LDr spectra for the ∆ enantiomer of [Ru(phen)3]2+. The first thing to be noted is that the metal complex chromophore, owing to the perturbation by DNA, no longer has D3 symmetry since in that case no FA value larger than 0.1 would have been possible (E-polarized emission). Analysis of these spectra along similar lines as those of the DPPZ complex provides the roll angles in Table 6. For [Ru(phen)3]2+ complexes with DNA the existence of three identical ligands complicates the situation and four binding geometry solutions are formally all compatible with the LD data for each enantiomer. To begin with, we cannot immediately exclude a geometry in which the C2-phen ligand to which the excitation is localized is pointing away from DNA (i.e., two phens are symmetrically directed inward). However, the observation that DNA provides some shielding from quenching by oxygen supports a geometry having the C2-phenanthroline pointing into the DNA groove. Of the remaining two possible geometries for each enantiomer, one roll angle is very similar to the one concluded for [Ru(phen)2(DPPZ)]2+. For ∆-[Ru(phen)3]2+, this is clearly the most probable orientation as judged from simple steric arguments, while the Λ enantiomer would (just as for the [Ru(phen)2(DPPZ)]2+ complex) have two phen wings more or less touching the walls of the groove. The general resemblance between [Ru(phen)2(DPPZ)]2+ and [Ru(phen)3]2+, regarding perturbation of absorption and CD spectra
< previous page
page_239
next page >
< previous page
page_240
next page > Page 240
and also with respect to gross features of their LDr spectra, is needed suggestive of similar geometrical solutions (as evidenced by Table 6). One may discuss the origins of the similarity in roll angles for the intercalating [Ru(phen)2(DPPZ)]2+ and nonintercalating [Ru(phen)3]2+ complexes. A potential aligning force could arise from the relatively rapid lateral shift motions of the base pairs, which would favor parallel orientation of a phenanthroline ligand in close contact with the stack of bases. Another variant of this concept is a more permanently shifted base pair leaving room for an inserted phen ligand (without opening the DNA). These two mechanisms, to explain the alignment of a planar ligand moiety parallel to the DNA base planes without being intercalated, could be likened to books in a bookshelf. The first one corresponds to the step formed by a book being pulled partly out of the stack of books; an approaching (new) book would tend to align itself parallel to the protruding step. The second mechanism would correspond to one book being partly pushed into the stack of books; the slot formed would allow partial insertion of a new book. Both kinds of interactions sterically provide an alignment force but are also anticipated to have an attractive component by stacking overlap between the π systems of ligand and bases. 5 Concluding Remarks From the information in this chapter, it is clear that despite intensive effort the interactions of substitution-inert metal complexes with DNA are not yet entirely understood. Although our understanding has improved enormously in recent years, the only complexes whose binding mode is unambiguous are the square planar platinum complexes, where X-ray diffraction augmented physical data to demonstrate intercalation. More recently, attention has focused on ''tetrahedral" copper complexes and octahedral ruthenium group complexes. While much progress has been made in defining their DNA binding properties, it is still unclear whether partial intercalation is possible for some complexes such as [Ru(phen)3]2+. A major point of controversy is which groove the complexes bind in, and questions remain about the origin of their enan-
< previous page
page_240
next page >
< previous page
page_241
next page > Page 241
tio-and sequence selectivities and their interactions with Z-DNA. Since LD has proven invaluable in demonstrating the nonclassical behavior of [Ru(phen)3]2+, and encouraged by our recent analyses pointing to similar binding geometries for all octahedral metal complexes, we hope to provide at least some answers in future studies. The 3-D nature of the metal complexes discussed here and the scope for making required modifications present attractive prospects. However, if these complexes are to find applications as DNA probes or as ''artificial nucleases," their interactions with DNA must be more completely understood. Only then will it be possible to rationally design new complexes to act in a designated manner. Abbreviations The abbreviations for most ligands, together with their structures, are given in Fig. 1. Additional abbreviations are as follows:
CD
circular dichroism
CT-DNA
calf thymus DNA
CV
cyclic voltammetry
DAPI
4',6-diamidino-2-phenylindole
dAMP
2' -deoxyadenosine 5' -monophosphate
ds-DNA
double-stranded DNA
EB
ethidium bromide
ECL
electrogenerated chemiluminescence
EDTA
ethylenediamine-N,N,N', N'-tetraacetate
ELD
electric linear dichroism
en
ethylenediamine
EOR
electric orientation relaxation
FA
emission anisotropy, fluorescence anisotropy
FLD
flow linear dichroism
FRET
fluorescence resonance energy transfer
IL (transition)
intraligand
INDO/S (calculations)
intermediate neglect of differential overlap/ spectroscopic
LD
linear dichroism
LDr
reduced linear dichroism
< previous page
page_241
next page >
< previous page
page_242
next page > Page 242
MLCT (transition)
metal-to-ligand charge transfer
MV2+
methylviologen
NaPA
sodium poly(acrylate)
NaPSS
sodium poly(styrenesulfonate)
NMR
nuclear magnetic resonance
PNA
peptide nucleic acid
[poly(dA-dT)]2
polydeoxy(adenylic-thymidylic) acid
[poly(dA)]·[poly(dT)] polydeoxyadenylic acid-polydeoxythymidylic acid [poly(dG-dC)]2
polydeoxy(guanylic-cytidylic) acid
[poly(dG-dme5C)]2
polydeoxy(guanylic-5-methylcytidylic) acid
[poly(dG)]·[poly(dC)] polydeoxyguanylic acid-polydeoxycytidylic acid POQ
5-[4-[(7-chloroquinolin-4-yl)amino]-2thiabutylcarboxaamido]phenanthroline
pur
purine
pyr
pyrimidine
RR (spectroscopy)
resonance Raman
ss-DNA
single-stranded DNA
TM
transition metal
T4-DNA
Coliphage T4 DNA
UV
ultraviolet
References 1. D. S. Sigman, A. Mazumder, and D. M. Perrin, Chem. Rev., 93, 2295 (1993).
2. A. M. Pyle and J. K. Barton, Progr. Inorg. Chem., 38, 413 (1990). 3. W. I. Sundquist and S. J. Lippard, Coord. Chem. Rev., 100, 293 (1990). 4. L. A. Basile and J. K. Barton, Metal Ions in Biological Systems, 25, 31 (1989). 5. C. S. Chow and J. K. Barton, Meth. Enzymol., 212, 219 (1992). 6. B. Nordén and F. Tjerneld, FEBS Lett., 67, 368 (1976). 7. V. Balzani and R. Ballardini, Photochem. Photobiol., 52, 409 (1990).
< previous page
page_242
next page >
< previous page
page_243
next page > Page 243
8. A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, and A. von Zelewsky, Coord. Chem. Rev., 84, 85 (1988). 9. K. Kalyanasundaram, Photochemistry of Polypyridine and Porphyrin Complexes, Academic Press, San Diego, 1992. 10. K. Kalyanasundaram, Photochemistry in Microheterogenous Systems, Academic Press, San Diego, 1987. 11. C. R. Cantor and P. R. Schimmel, Biophysical Chemistry, Part II, Techniques for the Study of Biological Structure and Function, W. H. Freeman, San Francisco, 1980. 12. M. Kubista, B. Åkerman, and B. Nordén, Biochemistry, 26, 4545 (1987). 13. W. Müller, D. M. Crothers, and M. J. Waring, Eur. J. Biochem., 39, 223 (1973). 14. L. P. G. Wakelin, A. Adams, C. Hunter, and M. J. Waring, Biochemistry, 20, 5779 (1981). 15. B. Nordén and F. Tjerneld, Chem. Phys. Lett., 50, 508 (1977). 16. S. K. Kim and B. Nordén, FEBS Lett., 315, 61 (1993). 17. E. Tuite and B. Nordén, J. Am. Chem. Soc., 116, 7548 (1994). 18. E. Tuite and B. Nordén, J. Chem. Soc., Chem. Commun., 53 (1995). 19. B. Nordén, S. Eriksson, S. K. Kim, M. Kubista, R. Lyng, and B. Åkerman in Molecular Basis of Specificity in Nucleic Acid-Drug Interactions (B. Pullman and J. Jortner, eds.), Kluwer, Dordrecht, 1990, pp. 2341. 20. H. Görner, A. Tossi, C. Stradowski, and D. Schulte-Frohlinde, J. Photochem. Photobiol., B: Biol., 2, 67 (1988). 21. C. Hiort, P. Lincoln, and B. Nordén, J. Am. Chem. Soc., 115, 3448 (1993). 22. E. Amouyal, A. Homsi, J.-C. Chambron, and J.-P. Sauvage, J. Chem. Soc., Dalton Trans., 1841 (1990). 23. R. F. Steiner and Y. Kubota in Excited States of Biopolymers (R. F. Steiner, ed.), Plenum Press, New York, 1983, pp. 203254. 24. E. Tuite, J. M. Kelly, G. D. Beddard, and G. S. Reid, Chem. Phys. Lett., 226, 517 (1994). 25. A. Kirsch-De Mesmaeker, G. Orellana, J. K. Barton, and N. J. Turro, Photochem. Photobiol., 52, 461 (1990). 26. J.-B. LePecq and C. Paoletti, J. Mol. Biol., 27, 87 (1967).
< previous page
page_243
next page >
< previous page
page_244
next page > Page 244
27. S. Satyanarayana, J. C. Dabrowiak, and J. B. Chaires, Biochemistry, 32, 2573 (1993). 28. G. Orellana, A. Kirsch-De Mesmaeker, J. K. Barton, and N. J. Turro, Photochem. Photobiol., 54, 499 (1991). 29. V. A. Bloomfield, D. M. Crothers, and I. Tinoco, Jr., Physical Chemistry of Nucleic Acids, Harper and Row, New York, 1974. 30. B. Nordén, M. Kubista, and T. Kurucsev, Quart. Rev. Biophys., 25, 51 (1992). 31. B. Nordén and T. Kurucsev, J. Mol. Recognition, 7, 141 (1994). 32. R. Lyng, T. Härd, and B. Nordén, Biopolymers, 26, 1327 (1987). 33. R. Lyng, A. Rodger, and B. Nordén, Biopolymers, 31, 1709 (1991). 34. R. Lyng, A. Rodger, and B. Nordén, Biopolymers, 32, 1201 (1992). 35. T. Härd and B. Nordén, Biopolymers, 25, 1209 (1986). 36. T. Härd, C. Hiort, and B. Nordén, J. Biomol. Struct. Dynam., 5, 89 (1987). 37. A. Yamagishi, J. Chem. Soc., Chem. Commun., 572 (1983). 38. A. Yamagishi, J. Phys. Chem., 88, 5709 (1984). 39. K. Naing, M. Takahashi, M. Taniguchi, and A. Yamagishi, J. Chem. Soc., Chem. Commun., 402 (1993). 40. B. Nordén, Inorg. Chim. Acta, 30, 83 (1978). 41. B. Nordén, FEBS Lett., 94, 204 (1978). 42. C. Hiort, B. Nordén, and A. Rodger, J. Am. Chem. Soc., 112, 1971 (1990). 43. C. V. Kumar, J. K. Barton, and N. J. Turro, J. Am. Chem. Soc., 107, 5518 (1985). 44. J. K. Barton, J. M. Goldberg, C. V. Kumar, and N. J. Turro, J. Am. Chem. Soc., 108, 2081 (1986). 45. A. E. Friedman, C. V. Kumar, N. J. Turro, and J. K. Barton, Nucl. Acids Res., 19, 2595 (1991). 46. S. A. Tysoe, R. J. Morgan, A. D. Baker, and T. C. Strekas, J. Phys. Chem., 97, 1707 (1993). 47. E. Terpetschnig, H. Szmacinski, H. Malak, and J. R. Lakowicz, Biophys. J., 68, 342 (1995). 48. J.-P. Lecomte, A. Kirsch-De Mesmaeker, M. Demeunynck, and J. Lhomme, J. Chem. Soc., Faraday Trans., 89, 3261 (1993).
< previous page
page_244
next page >
< previous page
page_245
next page > Page 245
49. J. M. Kelly, M. M. Feeney, A. B. Tossi, J.-P. Lecomte, and A. Kirsch-De Mesmaeker, Anti-Cancer Drug Design, 5, 69 (1990). 50. J.-P. Lecomte, A. Kirsch-De Mesmaeker, J. M. Kelly, A. B. Tossi, and H. Görner, Photochem. Photobiol., 55, 681 (1992). 51. A. B. Tossi. H. Görner, and D. Schulte-Frohlinde, Photochem. Photobiol., 50, 585 (1989). 52. A. B. Tossi and H. Görner, J. Photochem. Photobiol. B: Biol., 17, 115 (1993). 53. E. Taillandier and J. Liquier, Meth. Enzymol., 211, 307 (1992). 54. A. M. Pyle, J. P. Rehmann, R. Meshoyrer, C. V. Kumar, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 111, 3051 (1989). 55. C. Turro, S. H. Bossmann, G. E. Leroi, J. K. Barton, and N. J. Turro, Inorg. Chem., 33, 1344 (1994). 56. S. R. Smith, G. A. Neyhart, W. A. Kalsbeck, and H. H. Thorp, N. J. Chem., 18, 397 (1994). 57. J. P. Rehmann and J. K. Barton, Biochemistry, 29, 1701 (1990). 58. J. P. Rehmann and J. K. Barton, Biochemistry, 29, 1710 (1990). 59. W. D. Wilson, F. A. Tanious, H. J. Barton, R. L. Jones, K. Fox, R. L. Wydra, and L. Strekowski, Biochemistry, 29, 8452 (1990). 60. M. Eriksson, M. Leijon, C. Hiort, B. Nordén, and A. Gräslund, J. Am. Chem. Soc., 114, 4933 (1992). 61. M. Eriksson, M. Leijon, C. Hiort, B. Nordén, and A. Gräslund, Biochemistry, 33, 5031 (1994). 62. C. M. Dupureur and J. K. Barton, J. Am. Chem. Soc., 116, 10286 (1994). 63. M. T. Carter and A. J. Bard, J. Am. Chem. Soc., 109, 7528 (1987). 64. M. T. Carter, M. Rodriguez, and A. J. Bard, J. Am. Chem. Soc., 111, 8901 (1989). 65. M. T. Carter and A. J. Bard, Bioconj. Chem., 1, 257 (1990). 66. M. Rodriguez and A. J. Bard, Anal. Chem., 62, 2658 (1990). 67. N. Grover, N. Gupta, P. Singh, and H. H. Thorp, Inorg. Chem., 31, 2014 (1992). 68. K. E. van Holde, Physical Biochemistry, Prentice-Hall, New York, 1971. 69. G. Scatchard, Ann. N.Y. Acad. Sci., 51, 660 (1949).
< previous page
page_245
next page >
< previous page
page_246
next page > Page 246
70. D. L. Carlson, D. H. Huchital, E. J. Mantilla, R. D. Sheardy, and W. R. Murphy, Jr., J. Am. Chem. Soc., 115, 6424 (1993). 71. W. A. Kalsbeck and H. H. Thorp, J. Am. Chem. Soc., 115, 7146 (1993). 72. J. D. McGhee and P. H. von Hippel, J. Mol. Biol., 86, 469 (1974). Erratum: J. Mol. Biol., 103, 679 (1976). 73. D. M. Crothers, Biopolymers, 6, 575 (1968). 74. J.-J. Lawrence and M. Daune, Biochemistry, 15, 3301 (1976). 75. M. Howe-Grant and S. J. Lippard, Biochemistry, 18, 5762 (1979). 76. J. K. Barton and S. J. Lippard, Biochemistry, 18, 2661 (1979). 77. L. P. G. Wakelin, W. D. McFadyen, A. Walpole, and I. A. G. Roos, Biochem. J., 222, 203 (1984). 78. W. D. Wilson, C. R. Krishnamoorthy, Y. Wang, and J. C. Smith, Biopolymers, 24, 1941 (1985). 79. M. T. Record, C. F. Anderson, and T. M. Lohman, Quart. Rev. Biophys., 11, 103 (1978). 80. G. S. Manning, Quart. Rev. Biophys., 11, 179 (1978). 81. R. A. G. Friedmann and G. S. Manning, Biopolymers, 23, 2671 (1984). 82. W. D. Wilson and I. G. Lopp, Biopolymers, 18, 3025 (1979). 83. J. B. Chaires, W. Priebe, D. E. Graves, and T. G. Burke, J. Am. Chem. Soc., 115, 5360 (1993). 84. H.-C. Becker, MSc thesis, Chalmers University of Technology, Göteborg, 1994. 85. W. D. Wilson, F. A. Tanious, R. A. Watson, H. J. Barton, A. Strekowska, D. B. Harden, and L. Strekowski, Biochemistry, 28, 1984 (1989). 86. I. Haq, P. Lincoln, D. Suh, B. Nordén, B. Z. Chowdhry, and J. B. Chaires, J. Am. Chem. Soc., 117, 4788 (1995). 87. S. Satyanarayana, J. C. Dabrowiak, and J. B. Chaires, Biochemistry, 31, 9319 (1992). 88. G. A. Neyhart, N. Grover, S. R. Smith, W. A. Kalsbeck, T. A. Fairley, M. Cory, and H. H. Thorp, J. Am. Chem. Soc., 115, 4423 (1993). 89. P. Gilli, V. Ferretti, G. Gilli, and P. A. Borea, J. Phys. Chem., 98, 1515 (1994).
< previous page
page_246
next page >
< previous page
page_247
next page > Page 247
90. J. B. Chaires, Biopolymers, 24, 403 (1985). 91. K. J. Breslauer, D. P. Remeta, W. Chou, R. Ferrante, J. Curry, D. Zaunczkowski, J. G. Snyder, and L. A. Marky, Proc. Natl. Acad. Sci. USA, 84, 8922 (1987). 92. S. Chakraborty, R. Nandi, and M. Maiti, Biochem. Pharmacol., 39, 1181 (1990). 93. D. P. Remeta, C. P. Mudd, R. L. Berger, and K. J. Breslauer, Biochemistry, 32, 5064 (1993). 94. D. P. Remeta, C. P. Mudd, R. L. Berger, and K. J. Breslauer, Biochemistry, 30, 9799 (1991). 95. K. J. Breslauer, E. Freire, and M. Straume, Meth. Enzymol., 211, 533 (1992). 96. H. P. Hopkins and W. D. Wilson, Biopolymers, 26, 1347 (1987). 97. H. W. Zimmerman, Angew. Chem., Int. Ed. Engl., 25, 115 (1986). 98. V. K. Misra, K. A. Sharp, R. A. Friedman, and B. Honig, J. Mol. Biol., 238, 245 (1994). 99. F. Barcelo, J. Martorell, F. Gavilanes, and J. M. Gonzales-Ros, Biochem. Pharmacol., 37, 2133 (1988). 100. G. Manzini, M. L. Barcellona, M. Avitabile, and F. Quadrifoglio, Nucl. Acids Res., 11, 8861 (1983). 101. L. A. Marky, K. S. Blumenfeld, and K. J. Breslauer, Nucl. Acids Res., 11, 2857 (1983). 102. D. Freifelder, J. Mol. Biol., 60, 401 (1971). 103. J. L. Butour, E. Delain, D. Coulaud, J. B. LePecq, J. Barbet, and B. P. Roques, Biopolymers, 17, 873 (1978). 104. S. B. Smith, L. Finzi, and C. Bustamante, Science, 258, 1122 (1992). 105. G. Cohen and H. Eisenberg, Biopolymers, 8, 45 (1969). 106. M. Hogan, N. Dattagupta, and D. M. Crothers, Biochemistry, 18, 280 (1979). 107. D. Porschke, Biophys. Chem., 40, 169 (1991). 108. K.-E. Reinert, J. Biomol. Struct. Dynam., 9, 331 (1991). 109. J. B. Chaires, N. Dattagupta, and D. M. Crothers, Biochemistry, 21, 3933 (1982). 110. D. R. Graham and D. S. Sigman, Inorg. Chem., 23, 4188 (1984).
< previous page
page_247
next page >
< previous page
page_248
next page > Page 248
111. H. Yoshida, C. E. Swenberg, and N. Geacintov, Biochemistry, 26, 1351 (1987). 112. J. M. Veal and R. L. Rill, Biochemistry, 30, 1132 (1991). 113. F. Liu, K. A. Meadows, and D. R. McMillin, J. Am. Chem. Soc., 115, 6699 (1993). 114. W. Müller and D. M. Crothers, J. Mol. Biol., 35, 251 (1968). 115. H. Fritzsche, H. Triebel, J. B. Chaires, N. Dattagupta, and D. M. Crothers, Biochemistry, 21, 3940 (1982). 116. M. Wirth, O. Buchardt, T. Koch, P. E. Nielsen, and B. Nordén, J. Am. Chem. Soc., 110, 932 (1988). 117. M. J. Waring, Ann. Rev. Biochem., 50, 159 (1981). 118. M. Waring, J. Mol. Biol., 54, 247 (1970). 119. W. Keller, Proc. Natl. Acad. Sci. USA, 72, 4876 (1975). 120. R. T. Espejo and J. Lebowitz, Anal. Biochem., 72, 95 (1976). 121. J.-M. Saucier, B. Festy, and J.-B. LePecq, Biochemie, 53, 973 (1971). 122. R. J. DeLeys and D. A. Jackson, Biophys. Biochem. Res. Commun., 69, 446 (1976). 123. S. K. Poddard and J. Maniloff, Electrophoresis, 5, 172 (1984). 124. J. M. Kelly, A. B. Tossi, D. J. McConnell, and C. OhUigín, Nucl. Acids Res., 13, 6017 (1985). 125. A. E. Friedman, J.-C. Chambron, J.-P. Sauvage, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 112, 4960 (1990). 126. Y. Jenkins, A. E. Friedman, N. J. Turro, and J. K. Barton, Biochemistry, 31, 10809 (1992). 127. A. Sitlani, E. C. Long, A. M. Pyle, and J. K. Barton, J. Am. Chem. Soc., 114, 2303 (1992). 128. W. Müller and F. Gautier, Eur. J. Biochem., 54, 385 (1975). 129. H. Triebel, H. Bär, A. Walter, G. Burckhardt, and C. Zimmer, J. Biomol. Struct. Dynam., 11, 1085 (1994). 130. J.-M. Saucier, Biochemistry, 16, 5879 (1977). 131. N. Dattagupta, M. Hogan, and D. M. Crothers, Proc. Natl. Acad. Sci. USA, 75, 4286 (1978). 132. G. L. Cohen, W. R. Bauer, J. K. Barton, and S. J. Lippard, Science, 203, 1014 (1979).
< previous page
page_248
next page >
< previous page
page_249
next page > Page 249
133. J. K. Barton, J. J. Dannenberg, and A. L. Raphael, J. Am. Chem. Soc., 104, 4967 (1982). 134. J. K. Barton, A. T. Danishefsky, and J. M. Goldberg, J. Am. Chem. Soc., 106, 2172 (1984). 135. W. D. McFadyen, L. P. G. Wakelin, I. A. G. Roos, and B. L. Hillcoat, Biochem. J., 238, 757 (1986). 136. L. P. G. Wakelin, M. Romanos, T. K. Chen, D. Glaubiger, E. S. Canellakis, and M. J. Waring, Biochemistry, 17, 5057 (1978). 137. R. L. Jones, A. C. Lanier, R. A. Keel, and W. D. Wilson, Nucl. Acids. Res., 8, 1613 (1980). 138. C. Stradowski, H. Görner, L. J. Currell, and D. Schulte-Frohlinde, Biopolymers, 26, 189 (1987). 139. D. M. Crothers and J. Drak, Meth. Enzymol., 212, 46 (1992). 140. B. Nordén, C. Elvingsson, M. Jonsson, and B. Åkerman, Quart. Rev. Biophys., 24, 103 (1991). 141. M. M. Feeney, J. M. Kelly, A. Kirsch-De Mesmaeker, J.-P. Le-comte, and A. B. Tossi, J. Photochem. Photobiol. B: Biol., 23, 69 (1994). 142. P. S. Ho, C. A. Frederick, D. Saal, A. H.-J. Wang, and A. Rich, J. Biomol. Struct. Dynam., 4, 521 (1987). 143. P. M. van Vliet, J. G. Haasnoot, and J. Reedijk, Inorg. Chem., 33, 1934 (1994). 144. P. J. Bond, R. Langridge, K. W. Jennette, and S. J. Lippard, Proc. Natl. Acad. Sci. USA, 12, 4825 (1975). 145. S. J. Lippard, P. J. Bond, K. C. Wu, and W. R. Bauer, Science, 194, 726 (1976). 146. S. J. Lippard, Acc. Chem. Res., 11, 211 (1978). 147. A. H. J. Wang, J. Nathans, G. van der Marel, J. H. van Boom, and A. Rich, Nature, 276, 471 (1978). 148. Y.-S. Wong and S. J. Lippard, J. Chem. Soc., Chem. Commun., 824 (1977). 149. O. Kennard and W. N. Hunter, Angew. Chem. Int. Ed. Engl., 30, 1254 (1991). 150. S. Kamitori and F. Takusagawa, J. Am. Chem. Soc., 116, 4154 (1994). 151. H. M. Sobell and S. C. Jain, J. Mol. Biol., 68, 21 (1972).
< previous page
page_249
next page >
< previous page
page_250
next page > Page 250
152. H. M. Sobell, C.-C. Tsai, C. Jain, and S. G. Gilbert, J. Mol. Biol., 114, 333 (1977). 153. J. K. Barton, L. A. Basile, A. Danishefsky, and A. Alexandrescu, Proc. Natl. Acad. Sci. USA, 81, 1961 (1984). 154. C. J. Murphy and J. K. Barton, Meth. Enzymol., 226, 576 (1993). 155. A. D. Baker, R. J. Morgan, and T. C. Strekas, J. Am. Chem. Soc., 113, 1411 (1991). 156. J. K. Barton and E. Lolis, J. Am. Chem. Soc., 107, 708 (1985). 157. R. E. Mahnken, M. A. Billadeau, E. P. Nikonowicz, and H. Morrison, J. Am. Chem. Soc., 114, 9253 (1992). 158. N. Grover, N. Gupta, and H. H. Thorp, J. Am. Chem. Soc., 114, 3390 (1992). 159. H.-Y. Mei and J. K. Barton, J. Am. Chem. Soc., 108, 7414 (1986). 160. A. B. Tossi and J. M. Kelly, Photochem. Photobiol., 49, 545 (1989). 161. B. M. Goldstein, J. K. Barton, and H. M. Berman, Inorg. Chem., 25, 842 (1986). 162. H.-Y. Mei and J. K. Barton, Proc. Natl. Acad. Sci. USA, 85, 1339 (1988). 163. S. Neidle, DNA Structure and Recognition, Oxford University Press, Oxford, 1994. 164. R. J. Morgan, S. Chatterjee, A. D. Baker, and T. C. Strekas, Inorg. Chem., 30, 2687 (1991). 165. X. Gao and D. J. Patel, Biochemistry, 28, 751 (1989). 166. G. L. Duveneck, C. V. Kumar, N. J. Turro, and J. K. Barton, J. Phys. Chem., 92, 2028 (1988). 167. I. S. Haworth, A. H. Elcock, J. Freeman, A. Rodger, and W. G. Richards, J. Biomol. Struct. Dynam., 9, 23 (1991). 168. E. J. Gabbay, R. E. Scofield, and C. S. Baxter, J. Am. Chem. Soc., 95, 7850 (1973). 169. J. M. Veal and R. L. Rill, Biochemistry, 28, 3243 (1989). 170. J. M. Veal, K. Merchant, and R. L. Rill, Nucl. Acids Res., 19, 3383 (1991). 171. R. Tamilarasan and D. R. McMillin, Inorg. Chem., 29, 2798 (1990). 172. I. S. Haworth, A. H. Elcock, A. Rodger, and W. G. Richards, J. Biomol. Struct. Dynam., 9, 553 (1991).
< previous page
page_250
next page >
< previous page
page_251
next page > Page 251
173. R. H. Shafer, B. P. Roques, J. B. LePecq, and M. Delepierre, Eur. J. Biochem., 173, 377 (1988). 174. M. A. Keniry, S. C. Brown, E. Berman, and R. H. Shafer, Biochemistry, 26, 1058 (1987). 175. D. L. Banville, M. A. Keniry, M. Kam, and R. H. Shafer, Biochemistry, 29, 6521 (1990). 176. J. K. Barton and A. L. Raphael, J. Am. Chem. Soc., 106, 2466 (1984). 177. J. K. Barton and A. L. Raphael, Proc. Natl. Acad. Sci. USA, 82, 6460 (1985). 178. B. Müller, A. Raphael, and J. K. Barton, Proc. Natl. Acad. Sci. USA, 84, 1764 (1987). 179. M. R. Kirshenbaum, R. Tribolet, and J. K. Barton, Nucl. Acids Res., 16, 7943 (1988). 180. I. Lee and J. K. Barton, Biochemistry, 32, 6121 (1993). 181. N. T. Thuong and C. Hélène, Angew. Chem., Int. Ed. Engl., 32, 666 (1993). 182. P. Wittung, S. K. Kim, O. Buchardt, P. Nielsen, and B. Nordén, Nucl. Acids Res., 22, 5371 (1994). 183. J. M. Kelly, D. J. McConnell, C. OhUigín, A. B. Tossi, A. Kirsch-De Mesmaeker, A. Masschelein, and J. Nasielski, J. Chem. Soc., Chem. Commun., 1821 (1987). 184. M. B. Fleisher, K. C. Waterman, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 25, 3549 (1986). 185. C. Sentagne, J.-C. Chambron, J.-P. Sauvage, and N. Paillous, J. Photochem. Photobiol. B: Biol., 26, 165 (1994). 186. W. Bannwarth, D. Schmidt, R. L. Stallard, C. Hornung, R. Knorr, and F. Müller, Helv. Chim. Acta, 71, 2085 (1988). 187. W. Bannwarth and D. Schmidt, Tetrahedron Lett., 30, 1513 (1989). 188. W. Bannwarth, Anal. Biochem., 181, 216 (1989). 189. W. Bannwarth, W. Pfleiderer, and F. Müller, Helv. Chim. Acta, 74, 1991 (1991). 190. W. Bannwarth and F. Müller, Helv. Chim. Acta, 74, 2000 (1991). 191. J. Telser, K. A. Cruickshank, K. S. Schanze, and T. L. Netzel, J. Am. Chem. Soc., 111, 7221 (1989).
< previous page
page_251
next page >
< previous page
page_252
next page > Page 252
192. Y. Jenkins and J. K. Barton, J. Am. Chem. Soc., 114, 8736 (1992). 193. D. E. Bergstrom and N. P. Gerry, J. Am. Chem. Soc., 116, 12067 (1994). 194. J. M. Kelly, A. B. Tossi, D. J. McConnell, C. OhUigin, C. Hélène, and T. Le Doan in Free Radicals, Metal Ions and Biopolymers (P. C. Beaumont, D. J. Deeble, B. J. Parsons, and C. Rice-Evans, eds.), Richelieu Press, London, 1989, pp. 143156. 195. K. F. Baverstock and R. B. Cundall, Rad. Phys. Chem., 32, 553 (1988). 196. J. K. Barton, C. V. Kumar, and N. J. Turro, J. Am. Chem. Soc., 108, 6391 (1986). 197. M. D. Purugganan, C. V. Kumar, N. J. Turro, and J. K. Barton, Science, 241, 1645 (1988). 198. P. Fromherz and B. Reiger, J. Am. Chem. Soc., 108, 5361 (1986). 199. C. J. Murphy, M. R. Arkin, Y. Jenkins, N. D. Ghatlia, S. H. Bossman, N. J. Turro, and J. K. Barton, Science, 262, 1025 (1993). 200. C. J. Murphy, M. R. Arkin, N. D. Ghatlia, S. Bossmann, N. J. Turro, and J. K. Barton, Proc. Natl. Acad. Sci. USA, 91, 5315 (1994). 201. E. D. A. Stemp, M. R. Arkin, and J. K. Barton, J. Am. Chem. Soc., 117, 2375 (1995). 202. T. J. Meade and J. F. Kayyem, Angew. Chem. Int. Ed. Engl., 34, 352 (1995). 203. A. M. Brun and A. Harriman, J. Am. Chem. Soc., 114, 3656 (1992). 204. E. Krausz and J. Ferguson, Progr. Inorg. Chem., 37, 293 (1989). 205. S. David and J. K. Barton, J. Am. Chem. Soc., 115, 2984 (1993). 206. J. G. Collins, T. P. Shields, and J. K. Barton, J. Am. Chem. Soc., 116, 9840 (1994). 207. K. Naing, M. Takahashi, M. Taniguchi, and A. Yamagishi, Bull. Chem. Soc. Jpn., 67, 2424 (1994). 208. K. W. Jennette, S. J. Lippard, G. A. Vassiliades, and W. R. Bauer, Proc. Natl. Acad. Sci. USA, 71, 3839 (1974). 209. J. C. Dewan, S. J. Lippard, and W. R. Bauer, J. Am. Chem. Soc., 102, 858 (1980). 210. J.-P. Lecomte, A. Kirsch-De Mesmaeker, and G. Orellana, J. Phys. Chem., 98, 5382 (1994).
< previous page
page_252
next page >
< previous page
page_253
next page > Page 253
8 Effect of Metal Ions on the Fluorescence of Dyes Bound to DNA Vasil G. Bregadze, Jemal G. Chkhaberidze, and Irine G. Khutsishvili Institute of Physics, Academy of Sciences, 6 Tamarashvili Street, Tbilisi 380077, Georgia
254
1. Introduction
1.1. Dynamic Properties and Interactions of DNA with Dyes and Metal Ions: Complex Stability and Lifetime
255
1.2. DNA Binding Sites for Metal Ions and Dyes
1.3. Radiationless Electron Excitation Energy Transfer between Dyes Intercalated in DNA
2. Fluorescence Excitation Difference Spectra of Dyes in Complexes with DNA. Estimation of the Amount of Free and Bound Dyes
3. Fluorescence Quenching by Transition Metal Ions of Ethidium Bromide, Acridine Orange, and Proflavine Intercalated in DNA
256
258
260
265
4. Conclusions
< previous page
254
page_253
next page >
< previous page
page_254
next page > Page 254 266
Acknowledgments
266
Abbreviations
266
References
1 Introduction Dyes are important tools for the investigation of DNA-metal ion complexes because dyes are so small that they can easily penetrate into the cell nucleus and be absorbed by the DNA macromolecule. Many metal ions are either cancerogenic (like chromium or nickel ions) or mutagenic (like cadmium or lead ions), and some may also stimulate cancer development (like copper ions) [1,2]. Most of the known dyes, such as ethidium bromide (EB), acridine orange (AO), and proflavine (PF), have mutagenic properties. Such dyes are also used to exert their photodynamic influence on DNA and are thus applied in medical treatment for certain malignant skin diseases. We have investigated the effects of some metal ions from the first transition series on the photodestruction of EB intercalated in DNA. It was found that the intercalation slows down the EB photodestruction process and among the ions used, i.e., Co2+, Ni2+, and Cu2+ the Cu2+ ion is most effective. The example shows that these ions can accept electronic excitation. On the other hand, metal ions can lead to DNA damage via depurinization, breaking of single chains, and so forth [35], and the probability for such damages increases if metal ions are in the excited state. In ternary dye-metal ion-DNA complexes upon dye (donor) excitation the energy of the electron excitation most probably migrates to the metal ion (acceptor), thus increasing its interaction with DNA and consequently cause defects in DNA. 1.1 Dynamic Properties and Interactions of DNA with Dyes and Metal Ions: Complex Stability and Lifetime Recently new results appeared which can only be explained by taking into account the dynamic properties of macromolecules, i.e., also of DNA [6]. Modeling of the DNA interaction with the divalent metal ions of the first transition series is not an exception [5]. Comparison of the
< previous page
page_254
next page >
< previous page
page_255
next page > Page 255
results of their movements in DNA with their energetic characteristics reveals certain correlations rather well [6]. Previously [5], we gave Eq. (1),
derived from Frenkel's model for gas and vapor adsorption by solids [7,8], which connects the lifetime of complexes τ with the equilibrium (stability) constant K and which can be successfully applied to DNA and proteins interacting with small species, like metal ions and intercalators. For aqueous solutions τ0 quantifies the duration of the relaxation of rotary and translation movements of solvent molecules, metal ions, solvated ions, and other low molecular weight substances; it is always in the order of 10111010 sec. Thus, studies of the thermodynamics of the interactions of metal ions and dyes with DNA in the equilibrium state permit one to evaluate the dynamic characteristics, i.e., the lifetime of the complexes. It is worth noting that another approach to estimate the lifetime of complexes is necessary if the complex formation process is more complicated, as is the case in DNA interactions with larger species such as proteins. Let us now consider the logarithms of the stability constants of DNA complexes formed with Co2+, Ni2+ Cu2+, or Zn2+ [9], and EB [10], AO [11], or PF [12]. At an ionic strength of I = 0.01 M (NaCl) they are equal to 4.43 (Co2+), 4.76 (Ni2+), 5.15 (Cu2+), 4.36 (Zn2+), and 6.75 (EB), 5.68 (AO), 5.80 (PF). It is evident that the values of log K for the four metal ions differ only slightly. Similarly, the values for the dyes used in our experiments can also be explained by the dynamic characteristics of DNA. The lifetime of the DNA complexes with Co2+, Ni2+, Cu2+, or Zn2+ and the dyes, estimated according to Eq. (1), results in τ values of at least 106 sec for the metal ions and 105104 sec for the dyes. These lifetimes correspond to the large amplitude movements of phosphates, sugars, and bases, connected with the double-helix transformation from one form to another, i.e., the large-amplitude untwisting of the double helix [opening of separate pairs of nucleobases, cross-shaped structures, and clips formation (107105 sec)] [6]. 1.2 DNA Binding Sites for Metal Ions and Dyes Many dyes characteristically interact with DNA by penetrating between neighboring base pairs (intercalation), while the nature of the
< previous page
page_255
next page >
< previous page
page_256
next page > Page 256
interaction of divalent metal ions with DNA is mainly of an electron donor-acceptor type. Possible binding sites in DNA are the atoms carrying a partial negative charge, i.e., the atoms able to act as donors for electron pairs. Such atoms are N1, N3, and N7 in adenine; N3, N7, and O6 in guanine; O2 and N3 in cytosine; O2 in thymine, and all the nonbridging oxygen atoms present in the phosphate groups (which have a charge of 0.83e). Despite the great difference in the nature of binding of intercalated dyes and metal ions, their interactions with DNA also have common characteristics, e.g., the nonlinear character of the adsorption isotherms of these species on DNA in Scatchard charts. Nonlinearity in the case of dyes is caused by the adsorption process, which occurs according to the principle that binding does not take place at neighboring binding sites; this is a border case of anticooperative binding [13]. For metal ions the nonlinearity of the adsorption isotherm is explained either by at least two different binding sites or by anticooperative binding at ionic strengths higher than 0.1 M [5,9]. In addition, the divalent metal ions of the first transition series act as clips (even Cu2+ at concentrations below 0.25 per nucleotide increases the thermostability of DNA [3]). Intercalators increase the DNA thermostability as well [14]. Recently we showed [15] that, e.g., binding of Ni2+ with DNA, having various nucleotide compositions, and with polynucleotides mainly occurs at alternating dimeric units of the and type. Despite the great variety of metal ions interacting with DNA [3,5], they basically interact via their hydrated shell, thus competing with intercalated dyes for binding at DNA. 1.3 Radiationless Electron Excitation Energy Transfer between Dyes Intercalated in DNA The influence of metal ions on the fluorescence (F) of dyes intercalated in DNA can be provoked by their competition for binding with DNA (Sec. 1.2), on the one hand, and F quenching according to the Förster mechanism, on the other [16]. Before considering F quenching by metal ions it was necessary to confirm that the Förster mechanism can be applied to ternary complexes. For this purpose we investigated the energy transfer from AO to EB in DNA complexes. For this case we
< previous page
page_256
next page >
< previous page
page_257
next page > Page 257
could theoretically evaluate the energy transfer radii and also estimate them experimentally by changing the distance between the dyes. The radius of energy transfer, R0, at which the efficiency of energy transfer is 50%, is expressed by the following equation according to the Förster mechanism [16]:
where , and where φD is the quantum efficiency of donor F in the presence of acceptor and n is the index of the environmental refraction. F(ν) is the fluorescence spectrum of the donor normalized to 1; ε(ν) is the coefficient of the extinction of the acceptor at the frequency ν and K2 is the orientational factor characterizing the donor and acceptor interlocation in space. According to Eq. (2), the radii of the energy transfer, R0, have been calculated for DNA-dye-M2+ complexes. For the coefficients in Eq. (2) the following values were used: K2 = 2/3 [17]; for DNA in aqueous solution n = 1.6 [18]; for AO φD = 0.75 [12]; for PF φD = 0.45 (fluorescence quantum yield of PF intercalated between adjacent A-T pairs). The latter value can be obtained by dividing the apparent quantum yield of PF bound to DNA (0.15) by the mole fraction of adjacent A-T pairs (0.3364) on the assumption that the PF molecules are randomly distributed [19]; for EB φD = 0.68 [10]. The energy transfer radii R0 obtained from Eq. (2) for AO-EB pairs, where AO is a donor and EB is an acceptor of electron excitation, proved to be 35 ± 3 Å. For the experimental estimation of R0 we applied the method suggested in [20], where it was shown that for most of the dyes the following correlation holds:
where qod is the quantum efficiency of donor F when the distance between energy donor and acceptor R →∞, qd at a given R, qoa is the quantum efficiency of sensibilized acceptor F at R → 0 and qa at a given R. For the calculation of fluorescence quenching at different distances between energy donor and acceptor we varied the concentration of the AO and EB dyes intercalated in DNA. For the determination of the distance between the energy donor and acceptor it is important to know the exact quantity of dyes bound to
< previous page
page_257
next page >
< previous page
page_258
next page > Page 258
DNA. AO and EB have the same center of binding with DNA; therefore we used the equation of McGhee and von Hippel for a competitive interaction [13] and calculated the number of dyes bound to DNA, i.e., the distances between energy donors and acceptors for all concentrations. In Fig. 1 F quenching of AO and F ignition of EB is shown depending on the distance between them. By using the least square method for Eq. (2) we plotted the curves of quenching (for AO) and ignition (for EB). On the basis of these data we obtained the experimental value R0 = 39 ± 3 Å, which is in satisfactory correlation with the theoretically evaluated radii value R0 = 35 ± 3 Å. 2 Fluorescence Excitation Difference Spectra of Dyes in Complexes with DNA. Estimation of the Amount of Free and Bound Dyes It is clear that the effect of M2+ on the intensity of the fluorescence spectra of dyes bound to DNA must be caused by at least two principal phenomena: the ejection of intercalated dye molecules into the solution and the electron excitation energy transfer from dye molecules to M2+ ions. Thus, for the estimation of the influence of M2+ ions on the fluorescence quenching of PF, AO, and EB in DNA-dye complexes it was necessary to evaluate the number of dye molecules remaining in the complex with DNA after the formation of ternary DNA-dye-M2+ complexes. One of the methods for estimating the number of free and bound dye molecules is by studying the dye fluorescence excitation spectra. Figure 2 shows the F excitation spectra of EB, AO, and PF, free and bound to DNA, in 0.01 M NaCl water solution. The spectra were recorded with a SDL-1 (Russia, LOMO) spectrometer with a doublelightgrasp Cherny-Terner monochromator, and also with an MDR-2 monochromator for F excitation. As fluorescence excitation source a 100-W iodic quartz bulb has been chosen. Despite the great sensibility of the F excitation difference spectra we also used a double-beam spectrophotometer SPECORD M40 (Carl Zeiss, Jena) with fluorescent adjust-
< previous page
page_258
next page >
< previous page
page_259
next page > Page 259
Fig. 1. Ignition of AO (donor) (1) fluorescence and decrease of the EB (acceptor) fluorescence sensibilized by the donor (2), in dependence on the distance between the intercalated AO and EB in DNA. The points correspond to the experimental data under the following conditions: [DNA] = 5 × 105M (per base pair), [NaCl] = 102M, AO and EB concentrations varied from 1.25 × 1.25 × 106M to 2 × 105M. Excitation wavelength 440 nm. The lines show the curves for quenching and ignition as calculated by Eq. (3). ment of the same company to increase the accuracy of our measurements. Fluorescence excitation difference spectra (FEDS) of PF, AO, and EB have been monitored in the transmission mode. Figure 3 presents FEDS of EB, AO and PF, both free and bound to DNA. FEDS of free dye, DNA-dye, and DNA-dye-M2+ complexes have been recorded relative to the DNA-dye complex. Evaluation of the spectra shown in Fig. 3 made it obvious that the difference spectra of the F excitation of the ternary DNA-dye-M2+ complexes are the simple superposition of the difference spectra of F excitation of the free dyes. From the analysis of the obtained spectra we could determine which part of the dyes converts from the intercalated state to the free state. As a result, the quantity of dyes bound to DNA is decreased, as is the F intensity of dyes intercalated in DNA. For the estimation of the F intensity change of each bound dye, we normalized the F intensity of the DNA-dye complex to unity (i.e., we divided the F intensity of the ternary DNA-dye-M2+ complex by the concentration of
< previous page
page_259
next page >
< previous page
page_260
next page > Page 260
Fig. 2. Fluorescence excitation spectra of PF, AO, and EB molecules free () and bound to DNA (---), normalized to the intensity of the fluorescence excitation spectra of the dyes bound to DNA. the intercalated dye in the binary DNA-dye complex, for which we assumed F to be 1). 3 Fluorescence Quenching by Transition Metal Ions of Ethidium Bromide, Acridine Orange, and Proflavine Intercalated in DNA In Sec. 1.3 was shown the theoretical and experimental basis for the application of the Förster mechanism (radiationless electron excitation energy transfer in a dipole-dipole approximation) to ternary complexes
< previous page
page_260
next page >
< previous page
page_261
next page > Page 261
Fig. 3. Fluorescence excitation difference spectra of EB, AO, and PF. The spectra of free dye (. . .), DNA-dye (---), and DNA-dye-M2+ () complexes have been recorded relative to the DNA-dye complex. (1) DNA-dye-0.1 Cu2+/base pair, (2) DNA-dye-0.4 Cu2+/base pair, (3) DNA-dye-2.0 Cu2+/base pair.
< previous page
page_261
next page >
< previous page
next page >
page_262
Page 262 of the type DNA-energy donor-energy acceptor with the examples of DNA-AO-EB complexes. The simple comparison of the absorption spectra of Co2+, Ni2+, and Cu2+, on the one hand, with the F spectra of EB, AO, and PF in the visible and near-IR regions, on the other, shows that they overlap. When both dye molecules and metal ions are present in the complexes with DNA, their F and absorption spectra overlap even more. For instance, the absorption spectra of Co2+ and Ni2+ become more intensive upon DNA binding, and the Cu2+ absorption spectrum along with the great intensity change also undergoes a shift to shorter wavelengths [5], which further increases overlapping with the EB, AO, and PF fluorescence spectra. Table 1 shows integrals of overlapping of the fluorescence spectra of EB, AO, and PF with the absorption spectra of Co2+, Ni2+, and Cu2+ in ternary DNA-dye-M2+ complexes. In the same table the radii of the radiationless electron excitation energy transfer for the above ternary complexes are also given; the R0 values have been calculated using Eq. (2) (see Sec. 1.3). Despite the small values for R0, which lie in the interval of 612 Å, the effectiveness of metal ions in the EB, AO, and PF fluorescence quenching is significant at metal ion concentrations generally used in studies of many physicochemical characteristics of DNA. Figure 4 shows the number of dye molecules ejected from DNA into the solution, as evaluated by means of the methods described in Sec. 2, and the dependence of the effectiveness of the F quenching of EB, AO, and PF intercalated in DNA on the concentration of Mn2+, Co2+, TABLE 1 Overlapping Integrals J and Radii of the Radiationless Energy Transfer of Electron Excitation R0 for Ternary DNA-Dye-M2+ Complexes Calculated with Eq. (2) Ions Dyes
Ni2+
Co2+
Cu2+
J
R0 (Å)
J
R0 (Å)
J
R0 (Å)
EB
6.1 × 1018
8.5
9.1 × 1018
9.1
2.8 × 1017
10.9
AO
2.1 × 1018
7.2
3.8 × 1017
11.7
2.0 × 1017
10.5
PF
1.2 × 1018
6.0
3.2 × 1017
10.4
2.0 × 1017
9.6
< previous page
page_262
next page >
< previous page
page_263
next page > Page 263
Fig. 4. The data shown on the right scale ordinate axis have been evaluated from FEDS of the dye molecules ejected from DNA into the solution. The fluorescence quenching efficiency of EB, AO, and PF in complexes with DNA is shown on the left scale. On the abscissa axis the concentrations of Mn2+, Co2+, Ni2+, Cu2+, and Zn2+ added to the DNA-dye solution are given, evaluated per mole of DNA base pairs. Dots denote ) Mn2+, the amount of dye molecules ejected by ions: ( ( ) Zn2+, ( ) Co2+, ( ) Ni2+, ( ) Cu2+; and the fluorescence quenching efficiency by ions: (♦) Mn2+, ( ) Zn2+, ( ) Co2+, ( ) Ni2+, ( ) Cu2+. Concentrations: [DNA] = 5 × 105M, [NaCl] = 102M, [dye] = 5 × 106M.
< previous page
page_263
next page >
< previous page
page_264
next page > Page 264
Ni2+, Cu2+, or Zn2+ added to the DNA-dye solution. The concentration of added M2+, as calculated per phosphate group of DNA, is given on the x-axis, and the concentration of the ejected dyes and the quenching coefficient of the dyes remaining in the complex, taking into account the dyes ejected by metal ions and the specific F, are given on the y axis. Let us consider separately the influence of each metal ion on the DNA-dye complex. As the absorption spectra of Zn2+ and Mn2+ are insignificant in the visible region, their R0 values are small, and in fact these metal ions cannot quench the dye F as a result of dipole-dipole interactions. It is known that the EB molecule has no preference regarding base pairs in the intercalation process. At the same time, the quantum efficiency of the intercalated dye F depends slightly on the neighboring base pair. This can explain the observation that in the case of Zn2+ and Mn2+ the specific F, i.e., the F intensity, is not changed if the ejected dyes are taken into account. As for Ni2+, Co2+, and Cu2+, they show different abilities in F quenching of EB, in particular, R0(Cu2+) > R0(Co2+) > R0(Ni2+), which is correspondingly evident from the experimental plots. AO is bound 1.3 times more strongly to regions rich in A-T compared to those rich in G-C [11]. At the same time, the F quantum efficiency for AO is larger from regions rich in A-T than from those rich in G-C [11]. As Zn2+ and Mn2+ are not able to quench F but they cause at the same time a redistribution of dye from the regions rich in G-C to the regions rich in A-T, F must increase; indeed, this has been proven by experimental data. Ions such as Ni2+, Cu2+, and Co2+ cause both redistribution of dyes and F quenching. However, the F quantum efficiency of dye for the regions rich in A-T and G-C differs in these cases only slightly, i.e., φAT/φGC = 1.03 [11]. Therefore, the curves due to the specific F of these metal ions do not indicate the redistribution of dyes along DNA. In the case of PF we observe a different pattern. Like AO, PF has a preference for A-T pairs, but in contrast to AO its quantum efficiency in the G-C neighborhood from the intercalated state is 0. Thus its redistribution from a region rich in G-C to a region rich in A-T must cause a significant increase of F. Indeed, in the case of PF not only Mn2+ and Zn2+, but also Ni2+ and Cu2+ cause an increase of the specific F at low concentration, i.e., the share of quenching is small, but at a high Cu2+ concentration, when the quenching process becomes significant, the corresponding specific F is decreased. Only in the case of PF does the
< previous page
page_264
next page >
< previous page
page_265
next page > Page 265
specific F of Co2+ decrease over the whole range of metal ion concentration; this is caused by a high ability of Co2+ in PF fluorescence quenching. As is seen in Table 1, the radii of energy transfer between the dye intercalated in DNA and the metal ion bound to DNA varies in the range 612 Å. Moreover, some metal ions (e.g., Cu2+ and Co2+) have a great ability of F quenching of the dye intercalated in DNA, i.e., one has to assume that the distance between the dye and the metal ion does not exceed 68 º. However, under high concentrations of metal ions redistribution of dye from regions rich in GC to regions rich in A-T takes place. This indicates that in spite of the specific charge change of DNA, due to dye intercalation, the adsorption of metal ions in this region of the double helix is still going on, showing a high ability of metal ions to interact with phosphate groups of the DNA double helix. 4 Conclusions Electron excitation energy transfer from AO to EB in DNA complexes has been investigated. Experimentally estimated energy transfer radii proved to be 39 ± 3 Å and this figure is in satisfactory agreement with the theoretically calculated value R0 = 35 ± 3 Å. For the analysis of the amount of free dye molecules and those intercalated in DNA, the new method of measuring difference spectra of the dye fluorescence excitation is recommended. Consideration of the stereochemistry and of the potential binding sites for divalent metal ions of the first transition series regarding DNA, together with the process of adsorption of intercalated dyes, i.e., EB, AO, or PF, on the one hand, and the application of a thermodynamic approach relating the lifetime of complexes with their stability constants, on the other, allowed connection of the dynamic characteristics of DNA with the stability constants of these species and their ability to compete in binding to DNA. The effectiveness of Co2+, Ni2+, and Cu2+ in ternary DNA-dye-M2+ complexes to quench the fluorescence of EB, AO, and PF has been estimated by taking into account the ability of metal ions and dyes to compete in their binding to DNA. Fluorescence ignition of AO bound to DNA upon interaction with
< previous page
page_265
next page >
< previous page
page_266
next page > Page 266
Mn2+ and Zn2+ undoubtedly indicates the redistribution of the dye from DNA regions rich in G-C pairs to those rich in A-T pairs. Acknowledgments The authors express their gratitude to Dr. D. Lominadze for his cooperation in the investigation of energy transfer in DNA-AO-EB complexes. This work was supported in part by a Soros Foundation Grant awarded by the American Physical Society. Abbreviations
AO
acridine orange
EB
ethidium bromide
F
fluorescence
FEDS
fluorescence excitation difference spectra
IR
infrared
M2+
divalent metal ion
PF
proflavine
References 1. C. P. Flessel, A. Furst, and S. B. Radding in Metal Ions in Biological Systems, Vol. 10 (H. Sigel, ed.), Marcel Dekker, New York, Chap. 2, 1980. 2. E. L. Andronikashvili, V. G. Bregadze, and J. R. Monaselidze in Metal Ions in Biological Systems, Vol. 23 (H. Sigel and A. Sigel, eds.), Marcel Dekker, New York, Chap. 9, 1988. 3. G. L. Eichhorn in Inorganic Biochemistry, Vol. 2 (G. L. Eichhorn, ed.), Mir, Moscow, Chap. 34, 1978. 4. K. E. Wetterhahn, B. Demple, M. Kulesz-Martin, and E. S. Copeland, Cancer Res., 52, 4058 (1992).
< previous page
page_266
next page >
< previous page
page_267
next page > Page 267
5. V. G. Bregadze in Metal Ions in Biological Systems, Vol. 32 (A. Sigel and H. Sigel, eds.), Marcel Dekker, New York, Chap. 12, 1996. 6. L. V. Yakushevich, Mol. Biol. (Russia), 23, 652 (1989). 7. J. Frenkel, Z. Phys., 26, 117 (1924); Statistical Physics, M-L, Acad. Sci. USSR, 1948. 8. A. I. Slutsker, A. L. Mikhailin, and I. A. Slutsker, Physics-Uspechi (Russia), 164, 357 (1994). 9. E. S. Gelagutashvili and V. G. Bregadze, Bull. Acad. Sci. Georgian SSR, 128, 113 (1987). 10. J. B. Le Pecq and C. Paoletti, J. Mol. Biol., 27, 87 (1967). 11. J. Kapuscinski and Z. Darzynkiewich, J. Biomol. Struct. Dynam., 5, 127 (1987). 12. Y. Kubuta and R. F. Steiner, Biophys. Chem., 6, 279 (1977). 13. C. R. Cantor and P. R. Schimmell, Biophysical Chemistry, Part 3, Mir, Moscow, 1985. 14. A. T. Karapetian, V. I. Permogorov, and M. D. Frank-Kamenetski in Biopolymers Conformational Changes in Solution (E. L. Andronikashvili, ed.), Nauka, Moscow, 1973. 15. V. G. Bregadze and I. G. Khutsishvili, Proc. Acad. Sci. Georgia, Chem. Ser., 3 (1995), in press. 16. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Mir, Moscow, 1986. 17. Th. Förster in Modern Quantum Chemistry (O. Sinanoglu, ed.), Academic Press, New York, 1965. 18. R. E. Harrington, J. Am. Chem. Soc., 92, 6957 (1970). 19. T. G. Beridze, Satellites DNA, Nauka, Moscow, 1982. 20. V. L. Ermolaev, E. N. Bodunov, E. B. Sveshnikova, and T. A. Shakhverdov, Radiationless Electron Excitation Energy Transfer (M. D. Galanin, ed.), Nauka, Leningrad, 1977.
< previous page
page_267
next page >
< previous page
next page >
page_269
Page 269
9 Photolytic Covalent Binding of Metal Complexes to DNA Mark A. Billadeau and Harry Morrison Department of Chemistry, Purdue University, 1393 Brown Building, West Lafayette, IN 47907-1393, USA
270
1. Introduction
270
1.1. Ground State Covalent Binding to DNA
271
1.2. Concept of ''Photochemical Cisplatin" Reagents
272
1.3. Criteria for Effective Photo Cisplatins
272
1.4. Methods of Characterizing Covalent DNA Binding
273
2. d6 Metal Complexes
2.1. Thermal and Photochemical Properties of Polypyridyl Rhodium(III) Complexes
2.2. cis-Dichlorobis(polypyridyl)rhodium(III) Complexes and DNA
273
274
275
2.2.1. Dark Interactions with DNA
276 2.2.2. Photochemical Interactions of
with DNA 277
2.2.3. Characterization of Nucleoside Adducts
280 Photochemistry
2.2.4. Deoxyguanosine Effects on
< previous page
page_269
next page >
< previous page
page_270
next page > Page 270 283
2.2.5. 284 3. Longer Wavelength Reagents: The d3 Metal Complex, 284 3.1. Thermolysis and Photolysis of 285
3.2. Dark Interactions with DNA
286
3.3. Photochemical Interactions with DNA
288 3.4. Deoxyguanosine Effects on of Chromium(III) to DNA
Photochemistry and the Mechanism of Binding
290
4. Future Directions
291
Acknowledgments
291
Abbreviations
292
References
1 Introduction Since the initial reports by Rosenberg et al. on the biological affects of some platinum complexes, in particular cisplatin [1,2], there has been a great deal of interest in the development of transition metal complexes for use in medicine. Most of the research has focused on the creation of chemotherapeutic agents but other biological applications include imaging agents, spectroscopic probes of the structure of DNA, and artificial nucleases [3]. The focus of this chapter will be research directed to the development of new, photoactivated chemotherapeutic agents, specifically those involving covalent binding of the metal to the target biological substrate. 1.1 Ground State Covalent Binding to DNA Most research on potentially chemotherapeutic transition metal complexes has involved thermal activation. Thus the reactivity of cisplatin derives from the thermally induced substitution of the chloride ligands by water followed by covalent binding of the metal to DNA bases. Analogs of cisplatin have been of particular interest, where the ligands,
< previous page
page_270
next page >
< previous page
page_271
next page > Page 271
the oxidation state of the platinum, and even alternative central metals have been varied in order to increase the therapeutic properties and decrease the toxicity of this class of reagents [4,5]. Other classes of thermally activated antitumor reagents include metallocene dihalides [6,7], ruthenium-dimethylsulfoxide complexes [813], and rutheniumpolypyridyl complexes [14,15]. While the modes of action of these drugs may differ, there are some general similarities: (1) the environment necessary for substitution of the labile ligands is localized at the sites where genetic information is replicated; (2) substitution of these ligands leads to disruption of the replication process; and (3) there is a preferential accumulation of the drug within unhealthy tissue. Specifically for cisplatin, key features include (1) substitution of the chloride ligand inside rather than outside the cell due to the lower intracellular vs. extracellular chloride concentration (~ 4 mM and 100 mM, respectively) and (2) electrostatic attraction between the aquated cisplatin and DNA which eventually leads to the formation of DNA lesions and the inhibition of DNA replication. There is no specifically preferential accumulation within unhealthy tissue other than that which occurs due to the increased consumption of bodily nutrients by the continuously growing tissue. 1.2 Concept of ''Photochemical Cisplatin" Reagents All of the complexes discussed above have a common characteristic, i.e., the presence of thermally labile ligands. This facile substitution is a source of toxicity for these complexes in that substitution can and does occur within healthy tissue. Photochemotherapy has the potential to minimize such toxicity because cell damage requires activation of the drug by light, and the site of irradiation can be highly controlled. Most work in this area has focused on the development of photosensitizing drugs which produce DNA-damaging oxygen species [16]. The involvement of diffusible reactive oxygen species can lead to the loss of site specificity. An additional disadvantage is the reduced activity to be expected in hypoxic cells. In response to these shortcomings we have been developing reagents which are photoactivated but damage DNA via the "cisplatin-like" substitution of labilized ligands by the nucleic acid rather than by involving oxygen.
< previous page
page_271
next page >
< previous page
page_272
next page > Page 272
1.3 Criteria for Effective Photo Cisplatins There are several features of the ideal photo cisplatin reagent. The primary requirement is thermal stability of the complex under physiological conditions. The reagent must be inert to substitution except when activated by the appropriate wavelength of light. Preassociation of the complex with the DNA can be expected to increase the efficiency of the DNA binding reaction. Such preassociation can occur through electrostatic interactions between, for example, a cation and the DNA, or via hydrophobic interactions (e.g., by intercalation). However, there should be no chemical consequence of such interaction without irradiation and the complex should have no biological activity in the dark. The requisite photochemical lability is typically found in those metal complexes in which the ligand field (LF) excited state is lowest lying. Ideally the photochemistry should proceed with high quantum efficiency and be driven with long-wavelength light, i.e., by wavelengths beyond those absorbed by biological constituents. The ligands to be released should be good leaving groups and biologically innocuous. Finally, the metal complex should form stable covalent bonds with DNA, which can disrupt the DNA replication process. 1.4 Methods of Characterizing Covalent DNA Binding A variety of techniques have been used for the characterization of metallated DNA. Some of these have been employed for the structural characterization of the metal-DNA complex, such as the use of NMR spectroscopy on platinated deoxyoligonucleotides [1722] or X-ray studies of platinated deoxydinucleotides [23,24], while others involve the characterization of the metallated-nucleoside products formed from enzymatically digested metallated DNA [2527]. The existence of a metal-DNA adduct can also be detected using shifts in the absorption maxima of DNA bases following reaction with a complex [28]. Size exclusion chromatography is a particularly useful technique for assessing the extent to which metal has become covalently bound to the nucleic acid. One can use absorption spectroscopy to monitor the eluting fractions and by monitoring the eluant at a wavelength absorbed by the DNA, and also at a wavelength absorbed only by metal-containing fractions, one can determine the fraction of metal in the mixture which
< previous page
page_272
next page >
< previous page
page_273
next page > Page 273
is firmly bound to the DNA [27,29]. We have found atomic absorption spectroscopy to be invaluable for quantifying the amount of bound metal [27]. Analysis follows exhaustive dialysis and multiple precipitations of the metallated DNA with the quantity of the DNA determined by any of several techniques, e.g., 260-nm absorbance, colorimetrically by the Burton assay [30], or fluorometrically [31]. 2 d6 Metal Complexes Among the transition metals, the properties of the complexes of the d6 metals have received a great deal of attention. This has provided a wealth of information on the synthesis, photophysics, and photochemistry of these compounds [3237] and allows one to choose a complex appropriate for a desired photochemical process. In particular, the polypyridyl d6 metal complexes of rhodium(III) have the thermal stability and photochemical activity deemed desirable for the photo cisplatin concept. 2.1 Thermal and Photochemical Properties of Polypyridyl Rhodium(III) Complexes There are two general classes of polypyridyl rhodium(III) complexes that have been studied to date; those which are tris- and those which are bis-chelated. The photochemistry of the tris(polypyridyl)rhodium(III) complexes is dictated by the lowest lying excited state in these complexes, 3IL [38]. Photochemistry from this excited state typically involves electron transfer since the excited state is a powerful oxidizing agent, i.e., E0≈ 2.0 V for . While this reaction can lead to ligand substitution, the oxidation of DNA is known to lead to DNA cleavage. The tris(polpyridyl)rhodium(III) complexes are therefore inappropriate if one wants to achieve the formation of DNA-metal lesions without such cleavage. On the other hand, the lowest lying excited state in the bis(polypyridyl)rhodium(III) complexes has been identified as the 3LF [38]. This excited state typically leads to photosubstitution, with the most labile ligand (weakest σ donor) being replaced [39]. Chlorine atoms are common ligands and complexes are readily prepared with two chlorides in
< previous page
page_273
next page >
< previous page
next page >
page_274
Page 274 the cis configuration, analogous to the configuration found in cisplatin. The thermal stability of a cisdichlorobis(polypyridyl)rhodium(III) complex, , was studied by heating an aqueous solution of the optically active complex while monitoring the formation of products and the percentage of the optical activity which remained [39]. No reaction was observed at ambient temperature; only at elevated temperatures (~ 80°C) in show reactivity.
the presence of a reducing agent or a strong base did
The cis-dihalobis(polypyridyl)rhodium(III) complexes have been shown to be photochemically reactive, with the reported quantum efficiencies for loss of halide, using 350 nm excitation, ranging from 0.02 to 0.14 [40]. The photochemistry involves stepwise aquation with initial formation of the haloaqua complex (a monoprotic acid) followed by formation of the diaqua complex (a diprotic acid). Since the photoaquation proceeds with displacement of the weakest ligand, pH affects the second photoaquation step, i.e., photoaquation of the acid form of the haloaqua complex is a degenerate reaction, while photoaquation of the base form yields the diaqua complex. Photosubstitution has been shown to be dissociative by the pressure dependency of the quantum yields and luminescence of the model complex,
[41]. The reaction has been shown to proceed in a nonstereospecific manner [39].
Photoisomerization of Rh(phen)2XYn+ has yet been reported.
to its trans isomer has been reported [42] whereas no trans isomer of
In addition to photosubstitution, photoredox processes have been reported for two cisdihalobis(polypyridyl)rhodium(III) complexes, and [43]. In this study the quantum efficiencies for the formation of Rh(II) via reductive quenching with 1,2,4-trimethoxybenzene were found to be 0.34 and 0.065 for higher efficiency for by the phenyl substituents.
and
, respectively [43]. The
was attributed to slower back electron transfer due to steric hindrance
2.2 cis-Dichlorobis(polypyridyl)rhodium(III) Complexes and DNA The reactivity of DNA with two complexes, some detail.
< previous page
and
page_274
, has been studied in
next page >
< previous page
next page >
page_275
Page 275 2.2.1 Dark Interactions with DNA As noted earlier, ''photochemical cisplatin" reagents should be biologically and chemically inert under physiological conditions. Both
and
were shown to be only slightly mutagenic by the
Ames test [44]. A 23% loss of was observed when a phosphate buffer (pH 7) solution of the bisphenanthroline complex was heated at reflux for 24 hr, but no DNA-Rh adducts were formed upon incubation of with DNA at 37°C [27]. There is a low level of association of , and its diaqua analog, with calf thymus DNA (at I = 0.022 Ka = 150 and 470, respectively), as evidenced by equilibrium dialysis data [27]. ,
The measured association constants for
,
, and
at three ionic strengths are shown in Table 1. The association constants increase as the ionic strength is diminished, as would be expected for an electrostatic interaction between DNA and cationic metal complexes. The association constants are also affected by the overall charge of the metal complex, i.e., the association constants decrease through the series , measured for the neutral complex Ru(phen)2(CN)2 [45].
, and
. No association could be
TABLE 1 DNA Association Constants, K (M1), for Various Metal-Phenanthroline Complexes Complex
I = 0.22
I = 0.022
I = 0.0042
0
150
430
450
2500
6400
200
2700
11000
0
470
1600
Source: Reprinted from [27] by permission of the copyright owner.
< previous page
page_275
next page >
< previous page
page_276
next page > Page 276
These data are consistent with the proposal of groove binding as the mode of DNA association for the phenanthroline complexes [46,47]. Groove binding is thought to be driven by a complex mixture of hydrophobic, electrostatic, and hydrogen bonding interactions [48]. Note that the bis aqua complex has an effective charge of 2+ at pH 7 [49,50]. 2.2.2 Photochemical Interactions of
with DNA
and bind covalently to DNA when they are photolyzed in Both the presence of the nucleic acid. The quantum efficiencies for the photometallation of native DNA, under argon and oxygen, are 1.1 and 0.91 × 103 for
and 0.88 and 0.91 × 103 for
can react [27]. The similarity of the dichloro and diaqua quantum efficiencies suggests that the directly with the DNA, an observation corroborated by a study in which the racemic complex was photolyzed with DNA and the starting material reisolated. High-performance liquid chromatography (HPLC)-purified material exhibited negative Cotton effects indicative of an enrichment in the ∆ isomer [27]. A similar observation was made for photolysis of racemic cis-Rh(phen)2Cl(OH2)2+ with DNA [27]. Photometallation is affected by the nature of the DNA and the ionic strength. Since metallation occurs at the electronrich sites of the DNA bases, one would expect higher metal incorporation when the DNA bases are more accessible as in denatured DNA. This is evident in the quantum efficiencies for photometallation, 5.9 × 103 (argon) and 6.9 × 103 (air) for and 2.2 × 103 (argon) and 2.1 × 103 (air) for [27]. We noted earlier that lowering the ionic strength increases the association of these complexes with DNA. As one might suspect, a diminished ionic strength also increases the photometallation, with both native and denatured DNA (cf. Table 2). To gain an understanding of the DNA base selectivity of , photolyses were performed with four polyribonucleotides (Poly G, Poly A, Poly U, and Poly C). A selectivity for the purines was observed, particularly for guanosine (cf. Table 3). Cisplatin exhibits a similar selectivity. The source of the selectivity for guanosine will be discussed below.
< previous page
page_276
next page >
< previous page
page_277
next page > Page 277
TABLE 2 Rhodium Incorporation Values (nmol Rh/mg DNA) for Photolyses of cis-Rh(phen)2Cl2+ with Native and Denatured DNA at Four Ionic Strengthsa Ionic strength
Native DNA
Denatured DNA
0.083
369
595
0.041
448
694
0.021
580
782
0.010
601
810
aAll photolysis times were 158 min. Source: Reprinted from [27] by permission of the copyright owner. 2.2.3 Characterization of Nucleoside Adducts Adducts containing the chromophore ''Rh(phen)2" have been detected following the aerobic photolyses of with native DNA, dGMP, and dG. The adducts appear in the HPLC as a set of three peaks, the ratio of which varies with the reactant. Photolyses with DNA and dGMP following enzymatic digestion yield more of adduct I than adducts II and III [51]. Photolysis with dG itself yields a larger proportion TABLE 3 Rhodium Incorporation Values (nmol Rh/mg DNA) for Photolyses of Polyribonucleotidesa
with
Poly G
265
Poly Y
68
Poly A
167
Poly C
12
aAll photolysis times were 2 hr. Source: Reprinted from [27] by permission of the copyright owner.
< previous page
page_277
next page >
< previous page
page_278
next page > Page 278
of Adduct II. Adduct I photochemically converts to adduct II whereas adduct II thermally converts to adduct I. The photochemical conversion is nearly quantitative but the thermal conversion yields a mixture. However, both reactions proceed without decomposition. Adduct II has been extensively characterized and has been assigned the structure cis-Rh(phen)2(N1-dG)(OH2)2+ (cf. Fig. 1). The site of metallation has been assigned to N1 based on the change in C13 chemical shifts observed for C2 and C6. Support is provided by the fact that the H8 resonance in the 1H nuclear magnetic resonance (NMR) spectrum varies with pH, indicating that the more common target site, N7, has been unaffected by metallation. Finally, the (diasteriotopic) exocyclic amine protons are distinguishable in the 1H NMR spectrum, further supportive of metallation on the six-membered ring. The magnetic environments of the phenanthroline protons (2 and 2') in the adducts are also useful in the assignment of structure. The 1H NMR signals for these protons in the various complexes are as follows: (10.02 ppm) , cis-Rh(phen)2Cl(OH2)2+ (10.05 and 9.52 ppm), (9.61 ppm), adduct II (9.60 and 8.91 ppm), and adduct I (9.69 and 9.62 ppm). It is clear from these data that the resonances for phenanthroline protons pointing at a chloride ligand appear at ~10 ppm, while those pointing at an aqua ligand appear near
Fig. 1. Proposed structure of dG adduct II.
< previous page
page_278
next page >
< previous page
page_279
next page > Page 279
9.6 ppm. In adduct II there is a further upfield shift of one proton to 8.91 ppm due to the ring current effects of the dG. Adduct I is similar to adduct II in most of its features but the two ~9.6 ppm resonances indicate the presence of an aqua ligand and no special effect of the dG. The latter would be explained if binding were to an exocyclic atom of the base, thus diminishing the ring current effects felt by the phenanthroline protons. The clean interconversion of the two adducts further supports their close similarity but no conclusive structural assignment has yet been made. In addition to the dG adducts discussed above, one dA adduct has been isolated from the photolysis of with dA. Fast atom bombardment-mass spectrometry (FAB-MS) analysis of this product gave a molecular ion of 732 amu, corresponding to [Rh(phen)2-(dA)(OH2)]. There is again no NMR evidence for N7 metallation but a downfield shift in C4 and upfield shifts in C2 and C6 are suggestive of a reaction of the purine at N3 [52]. Therefore, the structure of the dA-adduct has been tentatively assigned as cis-Rh(phen)2(N3-dA)(OH2)3+ (cf. Fig. 2) [27]. Interestingly, when the metal complex is irradiated with dG under argon, two adducts are formed which involve binding of the metal at N7 [53]. This is evidenced in the proton and carbon NMR spectra, wherein one observes a 0.6 ppm downfield shift of the dG H8 resonance and a 5 ppm downfield shift for C8 in each of the photoadducts. Confirmation is provided by acidifying a sample of one of the adducts with DCl and
Fig. 2. Proposed structure for dA adduct.
< previous page
page_279
next page >
< previous page
page_280
next page > Page 280
comparing the effect on the H8 resonance to the effect of acidification on dG. Whereas the chemical shift of the H8 proton in unbound dG shifted 0.8 ppm downfield the chemical shift for this proton in the adduct was unchanged. The downfield shift in dG is due to protonation at N7; the absence of such an effect supports the proposal of metallation of this nitrogen. A rationale for the differences in the sites of metallation under air and argon is presented below. 2.2.4 Deoxyguanosine Effects on The photochemistry of quenches the photoaquation of
Photochemistry can be affected by the presence of various species. Thus oxygen , presumably by energy transfer with formation of 1O2 [54].
By contrast, the quantum efficiency for photoaquation of is increased by the presence of dG (from 0.02 to 0.05) though photoaquation is essentially unaffected by the presence of dA [54]. There are three possible mechanisms by which dG can affect φdis: energy transfer, oxidative quenching of the rhodium excited state, and reductive quenching. Of these, reductive quenching is the most viable option, and the oxidation potentials of guanine (1.07 V vs. NHE) and adenine (1.30 V vs. NHE) [53], by comparison with an estimated 3LF reduction potential of 1.18 V vs. NHE [55], would be consistent with selective reductive quenching by dG [54]. To test this was studied in the presence of uric acid. Since uric acid has hypothesis, the photochemistry of a lower oxidation potential (0.67 V vs. NHE) [55], a higher φdis would be expected. In fact, a φdis of 1.0 was measured when uric acid was present [54]. Oxygen has a twofold effect on the photochemistry of in the presence of dG or uric acid. It quenches the binding reaction, as it does photoaquation in the absence of the base (see above), and it also changes the in the presence of oxygen primarily affords aquation product distribution. Photolysis of products and those adducts that have been isolated appear to involve binding to heteroatoms in the six-membered purine ring. In the absence of oxygen the major products are the rhodium-dG or rhodium-uric acid coupling products and, for dG, N7 has been assigned as the site of metallation (see above).
< previous page
page_280
next page >
< previous page
page_281
next page > Page 281
The oxidation chemistry of dG (and guanosine) is important in understanding the site of its metallation. Kasai et al. showed that excitation of riboflavin in the presence of DNA leads to the formation of a dG radical cation within the DNA [56]. This radical cation, which resides in the five-membered ring, hydrates and then deprotonates to yield the 8hydroxydeoxyguanosine (8-OH-dG) derivative. However, when riboflavin was irradiated with dG itself only trace amounts of 8-OH-dG were detected. This was attributed to deprotonation of the initial nucleoside radical cation, a pathway partially prevented in DNA by base stacking interactions [56]. The facility of dG+ deprotonation is corroborated by electrochemical, chemical, and radiological oxidation studies [5559]. Guanosine can be electrochemically oxidized in a one proton-one electron wave [57]. Small amounts of 8-OH-guanosine are formed during electrolysis, presumably concurrent with oxidation. However, the majority of the products were dimers and trimers of guanosine [57]. It is argued that these form by radical coupling of the initially formed radical at C8 or the secondary radicals at N1 and exocyclic NH. Studies of dG oxidation by indicate that the initially produced species is a radical cation with a pKa of 3.9, so that deprotonation should occur under physiological conditions. The exocyclic oxygen has been assigned as the radical site based on aromaticity arguments [58]. Electron spin resonance (ESR) and electron nuclear double resonance (ENDOR) studies indicate that the radical produced by deprotonation of the radiation-induced cation resides on the exocyclic amine [59] or N1 [60]. in the presence of dG can be explained by a With the above in mind, the photochemistry of competition between 3LF photosubstitution and electron transfer via reductive quenching of the 3LF. The 3d-d photosubstitution pathway is summarized in reactions 14 of Scheme 1:
< previous page
page_281
next page >
< previous page
page_282
next page > Page 282
Scheme 1. Proposed mechanism for dG adduct formation. In the scheme excitation is followed by either dissociative substitution of the weakest ligand (cf. reactions 2 and 3) or by quenching by oxygen via energy transfer (cf. reaction 4). When dG is present reductive quenching competes with this pathway and there is electron transfer from the ground state of dG to the 3LF (reaction 5), which yields rhodium(II) and dG+. Concomitant with reduction to rhodium(II), there is a reduction in coordination number to 5 by the loss of Cl. Under argon this ion pair can couple by oxidative addition (i.e., back electron transfer) to form a rhodium(III)-dG adduct in which the site of metallation is the site of the primary dG radical formed during reductive quenching, i.e., N7. Reductive quenching by dG also occurs in the presence of oxygen but the rhodium(II) complex is rapidly oxidized by oxygen
< previous page
page_282
next page >
< previous page
page_283
next page > Page 283
thus preventing the oxidative addition coupling reaction (cf. reaction 7). The net effect is a catalysis of the aquation of
.
Without rhodium(II) available for oxidative addition the dG+ (1) adds water to ultimately form 8-HO-dG (cf. reaction 8) or (2) deprotonates to form dG, for which there are several possible radical sites in the six-membered ring (cf. reactions 9 and 10). The dG could then form a rhodium(III)-(N1-dG) adduct through a reaction with excited state or ground state aquated rhodium phenanthroline complex (cf. reaction 11a and 11b) (note that the t1/2 of dG is estimated to be 1 msec) [61]. With the above reactions in mind it is interesting to note that oxygen has no affect on the φ of DNA metallation and that only six-membered ring adducts have been isolated from DNA photolyses. These results suggest these metalDNA reactions involve the minor groove. 2.2.5
While this complex exhibits photochemistry similar to that of , the molecule has a significantly different shape from the bisphenanthroline complex which would be expected to affect its interaction with DNA. In fact, in the absence of light the 2,2'-bipyridyl (bpy) complex shows no measurable association with DNA at ionic strengths from 0.22 to 0.0045 (some association of the bisphenanthroline complex is observed at the lower ionic strengths; see above) nor does it bind covalently to the nucleic acid in the dark [62]. However, it does react with the DNA upon photolysis, with the φ values for DNA metallation in air equal to 0.90 × 103 and 3.8 × 103 with native and denatured DNA, respectively [63]. By contrast with the bisphenanthroline complex, the reactions are approximately threefold more efficient under argon, i.e., φ values for native DNA metallation are 0.90 × 103 (air) and 2.45 × 103 (argon) [63]. The base selectivity for the bpy complex shows the same trend as for , i.e., high selectivity for the purines, especially guanosine [63]. These results suggest that
does not interact with DNA in the same manner as
. The high guanosine selectivity suggests that metallation of the DNA is still a consequence of a photoredox process but the greater oxygen dependency of φ for
< previous page
page_283
next page >
< previous page
page_284
next page > Page 284
native DNA metallation suggests a possibly greater role for N7-guanosine reactivity and interaction involving the major groove. 3 Longer Wavelength Reagents: The d3 Metal Complex, While the rhodium complexes discussed above have demonstrated the desired photochemical reactivity toward DNA, the potential clinical usefulness of these complexes is limited by the ultraviolet (UV) wavelengths needed to drive the covalent binding reactions. Such wavelengths are also strongly absorbed by biological constituents. We have therefore studied the thermal and photochemical interactions of has an LF absorbance at 558 nm.
with DNA since this complex
3.1 Thermolysis and Photolysis of This complex has been shown to be less thermally stable than . Hancock et al. reported equilibrium constants of 3 M and 0.11 M for the first and second aquations at 60°C and I = 0.11 M [64]. Since this equilibrium would be affected by various factors (temperature, counterion concentration, and pH), a thermolysis study was carried out under biologically relevant conditions, i.e., phosphate-buffered solution at pH 7, 37°C [29]. Under these conditions the rate of equilibration was slower than that observed by Hancock et al. Interestingly, the equilibrium was shifted more to the diaqua complex than had previously been observed. The relative enhancement of the second aquation is attributed to the pH since the pKa of the chloroaqua complex is likely to be such that the conjugate base, cis-Cr(phen)2Cl(OH)+, would predominate at pH 7 (the pKa values of the diaqua complex are known to be 3.4 and 6.0) [65]. The chloride would be preferentially replaced in this species whereas water would be the better leaving group when the OH ligand is protonated. is photochemically active, as expected for a complex which has the 2LF as the lowest lying excited state. Excitation leads to substitution of the weakest (e.g., Cl) ligand and therefore the stepwise aquation of the complex. We presume that the aquation in-
< previous page
page_284
next page >
< previous page
next page >
page_285
Page 285 volves an associative mechanism analogous to that responsible for the aquation of disappearance of
. The quantum efficiencies for
(pH 7, 23°C, 532-nm excitation) are 0.010 and 0.0026 in argon and air, respectively [29]. Overall the
aqueous chemistry of may be summarized as follows. The first aquation can occur thermally or photochemically. Following this step a pH-dependent equilibrium is established between cis-Cr(phen)2Cl(OH2)2+ and cis-Cr(phen)2Cl(OH)+. No net reaction would be observed from cis-Cr(phen)2Cl(OH2)2+ since the weakest (i.e., aqua) ligand would be labilized. Thermal or photochemical aquation of the cis-Cr(phen)2Cl(OH)+ would yield cis-Cr(phen)2(OH)(OH2)2+, which would now be part of an equilibrium also involving the diprotic and dibasic forms. A reversal of this sequence by substitution of one of the aqua-containing components by chloride could occur thermally or photochemically, at a sufficiently high chloride concentration. These reactions are summarized in Scheme 2.
Scheme 2. Thermal and photochemical reactions of
.
3.2 Dark Interactions with DNA As with the rhodium(III) complexes above, and have low association constants for interaction with DNA. These were too low to be measured with an ionic strength of I = 0.1, but at I = 0.01 (4°C) values of 1 × 103 and 0.4 × 103 and , respectively [29]. has been shown to be only slightly mutagenic were determined for by the Ames test and the mutagenicity has been ascribed to DNA damage by oxygen radicals produced by a redox mechanism [67]. Thermally activated covalent binding of the diaqua complex to DNA occurs at higher tem-
< previous page
page_285
next page >
< previous page
page_286
next page > Page 286
peratures. For example, a binding level of 111 nmol Cr/mg DNA was determined following a 2-hr thermolysis with calf thymus DNA at 1215°C [29]. A similar result (144 nmol Cr/mg DNA) was obtained following a 71-hr thermolysis at 37°C [29]. A much smaller level of covalent binding was observed for following a 71-hr thermolysis at 37°C (33 nmol/mg DNA), which could well be due to partial hydrolysis to the diaqua complex prior to reaction with DNA [29]. The DNA-chromium(III) adducts formed thermally are less stable than the photoinduced DNA-rhodium(III) products discussed above. This becomes evident when the DNA-chromium(III) product is subjected to the purification procedures which were successfully employed for the DNA-rhodium(III) product. There is a continuous loss of metal from the DNA during successive ethanol precipitations and significantly reduced levels of covalent binding after lyophilization of the nucleic acid or its chromatography on Sephadex G-25. Neither higher reaction temperatures nor extended reaction times seem to significantly impact the binding levels, which is an indication that all of the binding sites are rapidly and readily filled within the typical 2-hr reaction time. Furthermore, there is evidence that the thermal binding of involves two or more DNA sites and/or target heteroatoms. There is at least one mode of interaction which survives all of the purification procedures while other modes of attachment survive only exhaustive dialysis. 3.3 Photochemical Interactions with DNA Photolyses at 1215°C have shown that these chromium(III) complexes will covalently bind to DNA and that the binding level is comparable to those obtained using the rhodium(III) compounds. For example, a binding level of 111 nmol Cr/mg DNA was obtained from a 16.9-hr photolysate after exhaustive dialysis [29,68]. However, as was observed for the product of thermal reaction, additional purification procedures clearly indicate that the covalently bound chromium(III) DNA product is not as stable as that formed from rhodium(III). Thus, ethanol precipitation of DNA from a previously dialyzed solution leads to labilization of the chromium(III), possibly as a consequence of the NaCl which is added as part of this procedure and which might cause a shift in the aquation-chlorination equilibrium [62]. Size exclusion chromatography on Sephadex G-25 has also been found to labilize covalently bound chro-
< previous page
page_286
next page >
< previous page
next page >
page_287
Page 287 mium(III). Figure 3 is a histogram of such a chromatography and shows that almost equal amounts (1.0:0.92) of bound (fractions 810) and unbound (fractions 1316) metal are eluted, even though the DNA had been previously exhaustively dialyzed [29,69]. When the DNA-containing fractions from this separation were collected, lyophilized, and rechromatographed, additional chromium(III) was labilized with the ratio of bound to labilized metal now 1:0.42 [29]. In a separate experiment, a photolysate was dialyzed, the DNA solution lyophilized, and the DNA chromatographed and rechromatographed.
Fig. 3. Histogram (absorbance vs. fraction number) for the Sephadex chromatography of the previously dialyzed DNA solution from the 16.9-hr photolysis of
< previous page
page_287
with DNA.
next page >
< previous page
page_288
next page > Page 288
The histogram from this separation showed virtually completely bound chromium(III) (cf. Fig. 4) [29]. These results substantiate the proposal that there are multiple forms of bound chromium with varying degrees of stability and that there is a mode of binding which survives dialysis, lyophilization, and multiple chromatographies. The base selectivity of was determined from photolyses with polyribonucleotides. The binding levels following 3 days of dialysis are reported in Table 4. The results show the same trend as was seen for rhodium(III), albeit with somewhat less selectivity, i.e., preference for the purines, particularly guanosine. 3.4 Deoxyguanosine Effects on to DNA
Photochemistry and the Mechanism of Binding of Chromium(III)
. As was observed for the rhodium complex, the presence of dG affects the photoaquation of There is a twofold enhancement in the rate of disappearance of the complex in the presence of the nucleoside which is accompanied by an increased rate of formation of the diaqua product, . Interestingly, the rate of formation of the monoaqua species, cis-Cr(phen)2Cl(OH2)2+, is unaffected by the base [29]. The catalyzed conversion of the dichloride directly to the diaqua product has been attributed to a one-electron reductive quenching of the 2LF excited state of to
by dG, analogous to the conversion of by zinc amalgam (cf. Scheme 3) [70]. The
can be oxidized to the diaqua complex by molecular oxygen or a chromium(III) complex. The lack of any dG affect on the formation of cis-Cr(phen)2-
Scheme 3. Proposed mechanism for dG-catalyzed photoaquation of
< previous page
page_288
.
next page >
< previous page
page_289
next page > Page 289
Fig. 4. Histogram (absorbance vs. fraction number) for the Sephadex rechromatography of a first lyophilized and then a chromatographed dialyzed DNA solution from the 21-hr photolysis of
< previous page
page_289
with DNA.
next page >
< previous page
page_290
next page > Page 290
TABLE 4 Chromium Incorporation Values (nmol Cr/mg DNA) for Photolyses of with Polyribonucleotidesa Poly G
79 Poly C
Poly A
36 Poly Ub
16
aAll photolysis times were 115.3 hr. bPoly U was undetectable by absorption at 261 nm. Source: Reprinted from [29] by permission of the copyright owner. Cl(OH2)2+ was attributed to the instability of cis-CrII(phen)2Cl(OH2)+ formed in the redox pathway, hence any observed cis-CrIII(phen)2-Cl(OH2)2+ was formed via the LF substitution pathway and any dark aquation. , There are several possible mechanisms for the metallation of DNA upon its photolysis with i.e., a thermal reaction of diaqua photoproduct and either direct (or redox-catalyzed) substitution of the 2LF excited state. There are insufficient data to date to distinguish among these. A combination of these could be the source of the multiple binding sites indicated by the product stability studies, or there may be multiple sites for weak association of the complex with the DNA and a consequent mix of nucleophilic targets. 4 Future Directions Although both of the metal complexes discussed above meet some of the criteria outlined at the outset for a photocisplatin reagent, both systems are far from ideal. In particular, though the rhodium(III) complex has the desired thermal stability and photochemical lability, the wavelengths necessary to drive the photochemistry are shorter than one would like. Therefore, we have initiated sensitization studies with a visible absorbing dye, methylene blue (MB; λmax = 660 nm), and prelimi-
< previous page
page_290
next page >
< previous page
next page >
page_291
Page 291
Fig. 5. Structure of
complex.
nary results indicate that the dye sensitizes photoaquation of [62]. It is presumed that the sensitization occurs by energy transfer rather than electron transfer, since the rhodium(III)-MB redox couple is inappropriate for electron transfer between MB and rhodium(III) [71]. As regards the chromium(III) complex, this does have the desired longer wavelength absorption but suffers from poor thermal stability of the starting material as well as of the metallated nucleic acid product. The alternative trans complex shown in Fig. 5 has several potential advantages: light absorption at ~560 nm, a high quantum efficiency of photoaquation (φ = 0.30 for loss of Cl from with excitation at 565 nm [72]), an appropriate shape for intercalation, and the potential to form interstrand crosslinks (the O-Cr-O bond distance, 3.905 ± 0.02 Å, of trans-[Cr(NA)2(NH3)4](ClO4) is appropriate to span the base separation of 3.54.0 Å of B-DNA [77]). The preparation of such complexes is in progress. Acknowledgments Support by a National Institutes of Health (NIH) Research Service Award 5T32CA09634 from the Purdue Cancer Center to M. A. Billadeau and partial support by NIH grant RO1 HL53418-01 are gratefully acknowledged. Abbreviations φ
quantum efficiency
φdis
quantum efficiency for starting material disappearance
< previous page
page_291
next page >
< previous page
page_292
next page > Page 292
bpy
2,2'-bipyridyl
cisplatin
cis-diaminodichloroplatinum(II)
dA
deoxyadenosine
dapa
2,3-diaminopropionic acid
dG
2'-deoxyguanosine
dGMP
2'-deoxyguanosine-5'-monophosphate
dpphen
4,7-diphenyl-1,10-phenanthroline
en
ethylenediamine
ENDOR
electron nuclear double resonance
ESR
electron spin resonance
FAB-MS
fast atom bombardment-mass spectrometry
HPLC
high-performance liquid chromatography
IL
intraligand excited state
LF
ligand field, or d-d
MB
methylene blue
NA
carboxy-bound nicotinic acid
NHE
normal hydrogen electrode
NMR
nuclear magnetic resonance
8-OH-dG
8-hydroxy-2'-deoxyguanosine or 8-oxo-2'-deoxyguanosine
phen
1,10-penanthroline
UV
ultraviolet
References 1. B. Rosenberg, L. Vancamp, and T. Krigas, Nature, 205, 689 (1965). 2. B. Rosenberg, L. Vancamp, J. E. Trosko, and V. H. Mansour, Nature, 222, 385 (1969). 3. A. M. Pyle and J. K. Barton in Progress in Inorganic Chemistry: Bioinorganic Chemistry (S. J. Lippard, ed.), John Wiley and Sons, New York, 1990, p. 413 ff. 4. S. E. Sherman and S. J. Lippard, Chem. Rev., 87, 1153 (1987). 5. P. Umapathy, Coord. Chem. Rev., 95, 129 (1989). 6. P. Köpf-Maier and H. Köpf, Drugs of the Future, 11, 297 (1986). 7. L. Y. Kuo, M. G. Kanatzidis, M. Sabat, A. L. Tipton, and T. J. Marks, J. Am. Chem. Soc., 113, 9027 (1991).
< previous page
page_292
next page >
< previous page
page_293
next page > Page 293
8. E. Alessio, W. M. Attia, M. Calligaris, S. Cauci, L. Dolzani, G. Mestroni, C. Monti-Bragadin, G. Nardin, F. Quadrifoglio, G. Sava, M. Tamaro, and S. Zorzet in Platinum and Other Metal Coordination Compounds in Cancer Chemotherapy (M. Nicolini, ed.), Martinus Nijhoff, Boston, 1987, p. 617 ff. 9. G. Sava, S. Pacor, S. Zorzet, E. Alessio, and G. Mestroni, Pharmacol. Res., 21, 617 (1989). 10. S. Pacor, G. Sava, V. Ceschia, F. Bregant, G. Mestroni, and E. Alessio, Chem.-Biol. Interact., 78, 223 (1991). 11. S. Cauci, P. Viglino, G. Esposito, and F. Quadrifoglio, J. Inorg. Biochem., 43, 739 (1991). 12. S. Cauci, E. Alessio, G. Mestroni, and F. Quadrifoglio, Inorg. Chim. Acta, 137, 19 (1987). 13. G. Esposito, S. Cauci, F. Fogolari, E. Alessio, M. Scocchi, F. Quadrifoglio, and P. Viglino, Biochemistry, 31, 7094 (1992). 14. J. K. Barton and E. Lolis, J. Am. Chem. Soc., 107, 708 (1985). 15. N. Grover, N. Gupta, and H. H. Thorp, J. Am. Chem. Soc., 114, 3390 (1992). 16. A. R. Oseroff, G. Ara, D. Ohuoha, J. Aprille, J. C. Bommer, M. L. Yarmush, J. Foley, and L. Cincotta, Photochem. Photobiol., 46, 83 (1987). 17. H. J. H. den Hartog, C. Altona, J. H. van Boom, G. A. Vander Marel, C. A. G. Haasnoot, and J. Reedijk, J. Biomol. Struct. Dynam., 2, 1137 (1985). 18. J.-P. Girault, J.-C. Chottard, E. R. Guittet, J.-Y. Lallemand, T. Huynh-Dinh, and J. Igolen, Biochem. Biophys. Res. Commun., 109, 1157 (1982). 19. J. P. Cardonna, S. J. Lippard, M. J. Gait, and M. Singh, J. Am. Chem. Soc., 104, 5793 (1982). 20. J. Kozelka, G. A. Petsko, S. J. Lippard, and G. J. Quigley, J. Am. Chem. Soc., 107, 4079 (1985). 21. J. Kozelka, G. A. Petsko, G. J. Quigley, and S. J. Lippard, Inorg. Chem., 25, 1075 (1986). 22. J. Kozelka, S. Archer, G. A. Petsko, S. J. Lippard, and G. J. Quigley, Biopolymers, 26, 1245 (1987). 23. S. E. Sherman, D. Gibson, A. H. J. Wang, and S. J. Lippard, Science, 230, 412 (1985).
< previous page
page_293
next page >
< previous page
page_294
next page > Page 294
24. G. Admiraal, J. L. Vander Veer, R. A. G. de Graaff, J. H. J. den Hartog, and J. Reedijk, J. Am. Chem. Soc., 109, 592 (1987). 25. A. Eastman, Biochemistry, 22, 3927 (1983). 26. A. M. J. Fichtinger-Schepman, P. H. M. Lohman, and J. Reedijk, Nucl. Acids Res., 10, 5345 (1982). 27. R. E. Mahnken, M. A. Billadeau, E. P. Nikonowicz, and H. Morrison, J. Am. Chem. Soc., 114, 9253 (1992). 28. P.Horacek and J. Drobník, Biochim. Biophys. Acta, 254, 341 (1971). 29. M. A. Billadeau and H. Morrison, J. Inorg. Biochem., 57, 249 (1995). 30. H. J. Waterborg and H. R. Matthews, Methods in Molecular Biology, Vol. 2, Nucleic Acids, Humana Press, Clifton, New Jersey, 1984, pp. 13. 31. C. Labarca and K. Paigen, Anal. Biochem., 102, 344 (1980). 32. P. C. Ford, R. E. Hintze, and J. D. Petersen in Concepts of Inorganic Chemistry (A. W. Adamson and P. D. Fleischauer, eds.), John Wiley and Sons, New York, 1975, p. 203 ff. 33. P. C. Ford, Coord. Chem. Rev., 44, 61 (1982). 34. L. G. Van Quickenborne and A. Ceulemans, Coord. Chem. Rev., 48, 157 (1983). 35. A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, and A. von Zelewski, Coord. Chem. Rev., 84, 85 (1988). 36. Comprehensive Coordination Chemistry, Vols. 17 (Sir G. Wilkinson, R. D. Gillard, and J. A. McCleverty, eds.), Pergamon Press, New York, 1987. 37. K. Kalyanasundaram, Photochemistry of Polypyridine and Porphyrin Complexes, Academic Press, New York, 1992. 38. D. H. W. Carstens and G. A. Crosby, J. Mol. Spectrosc., 34, 113 (1970). 39. P. M. Gidney, R. D. Gillard, and B. T. Heaton, J. Chem. Soc. Dalton, 2621 (1972). 40. M. M Muri and W.-L. Huang, Inorg. Chem., 8, 1831 (1973). 41. S. Wieland, J. DiBenedetto, R. van Eldik, and P. C. Ford, Inorg. Chem., 25, 4893 (1986). 42. G. Krüger, S. Wieland, and R. van Eldik, Angew. Chem. Int. Ed. Engl., 26, 240 (1987).
< previous page
page_294
next page >
< previous page
page_295
next page > Page 295
43. T. Ohno, Coord. Chem. Rev., 64, 311 (1985). 44. G. Warren, E. Abbott, P. Schultz, K. Bennett, and S. Rogers, Mutat. Res., 88, 165 (1981). 45. J. M. Kelly, A. B. Tossi, D. J. McConnell, and C. OhUigin, Nucl. Acids. Res., 13, 6017 (1985). 46. M. Eriksson, M. Leijon, C. Hiort, B. Nordén, and A. Gräslund, J. Am. Chem. Soc., 114, 4933 (1992). 47. C. J. Murphy, M. R. Arkin, N. D. Ghatlia, S. H. Bossmann, N. J. Turro, and J. K. Barton, Science, 262, 1025 (1993). 48. E. C. Long and J. K. Barton, Acc. Chem. Res., 23, 273 (1990).
49. Based upon the pKa values of a similar complex, [50], the apparent charge was calculated to be +2.0 (11.4% 3+, 79.6% 2+, and 9.0% 1+) at pH 7 and an ionic strength of 0.02. 50. L. Lee, S. F. Clark, and J. D. Petersen, Inorg. Chem., 24, 3558 (1985). 51. The assignment of names to these adducts are based on HPLC retention times; adduct I elutes first while adduct III elutes last. Little is known about adduct III at this time. 52. R. J. Pugmire and D. M. Grant, J. Am. Chem. Soc., 93, 1880 (1971). 53. H. L. Harmon and H. Morrison, Inorg. Chem., 34, 4937 (1995). 54. M. A. Billadeau, K. V. Wood, and H. Morrison, Inorg. Chem., 33, 5780 (1994). 55. P. J. Elving, in Topics in Bioelectrochemistry and Bioenergetics, Vol. 1 (G. Milazzo, ed.), John Wiley and Sons, New York, 1976, p. 254 ff. 56. H. Kasai, Z. Yamaizumi, M. Berger, and J. Cadet, J. Am. Chem. Soc., 114, 9692 (1992). 57. P. Subramanian and G. Dryhurst, J. Electroanal. Chem., 224, 137 (1987). 58. S. Steenken, Chem. Rev., 89, 503 (1989). 59. E. O. Hole, W. H. Nelson, D. M. Close, and E. Sagstuen, J. Chem. Phys., 86, 5218 (1987). 60. B. Rakvin, J. N. Herak, K. Voit, and J. Hüttermann, Radiat. Environ. Biophys., 26, 1 (1987). 61. P. O'Neill, Radiat. Res., 96, 198 (1983).
< previous page
page_295
next page >
< previous page
page_296
next page > Page 296
62. R. E. Mahnken, doctoral dissertation, Purdue University, May 1991. 63. S. E. Torregrosa, masters thesis, Purdue University, May 1993. 64. M. P. Hancock, J. Josephsen, and C. E. Schäffer, Acta Chem. Scand., 30, 79 (1976). 65. R. G. Inskeep and J. Bjerrum, Acta Chem. Scand., 15, 62 (1961). 66. N. Serpone, M. A. Jamieson, M. S. Henry, M. Z. Hoffman, F. Bolletta, and M. Maestri, J. Am. Chem. Soc., 101, 2907 (1979). 67. K. D. Sugden, R. D. Geer, and S. J. Rogers, Biochemistry, 31, 11626 (1992). 68. By monitoring the chromium levels in the dialyzates, it was determined that 3 days of dialysis was necessary to remove the noncovalently bound chromium(III). 69. The absorbance of each fraction was measured at two wavelengths; 310 nm at which only chromium(III) absorbs and 260 nm at which both DNA and chromium(III) absorb. The amount of chromium(III) in each fraction was determined by the absorbance at 310 nm and the amount of DNA was determined by the absorbance at 260 nm corrected for absorbance due to chromium(III). 70. J. Josephsen and C. E. Schäffer, Chem. Commun., 61 (1970). 71. To our knowledge, neither the oxidation of MB nor the oxidation of
has been reported.
72. A. D. Kirk, K. C. Moss, and J. G. Valentin, Can. J. Chem., 49, 1524 (1971). 73. J. C. Chang, L. E. Gerdom, N. C. Baenzinger, and H. M. Goff, Inorg. Chem., 22, 1739 (1983).
< previous page
page_296
next page >
< previous page
page_297
next page > Page 297
10 Electrochemically Activated Nucleic Acid Oxidation Dean H. Johnston, Thomas W. Welch, and H. Holden Thorp Department of Chemistry, University of North Carolina, Chapel Hill, NC 27599-3290, USA
298
1. Introduction
2. Electrochemistry of One-Electron Couples Bound to DNA
299
2.1. General Considerations
301
2.2. Impact of DNA Diffusion
306
2.3. Impact of DNA Binding
309
3. Electrochemically Activated DNA Cleavage
309
3.1. Reductive Activation
310
3.2. Oxidative Activation
311
3.3. Metal-Mediated Oxidation
314
4. Oxidation Kinetics from Voltammetry
318
5. Conclusions
319
Acknowledgment
< previous page
299
page_297
next page >
< previous page
page_298
next page > Page 298 319
Abbreviations
320
References
1 Introduction Numerous tools for DNA diagnostics and other areas of biotechnology can be envisioned based on novel means for detecting DNA, RNA, and various substructures [14]. A recent advance in the area of DNA diagnostics is the ability to assemble spatially distinguished oligomers on the surface of a microchip using photoaddressable synthesis [57]. If these surfaces were built on a conducting or semiconducting electrode material, probing the structures or sequences of the immobilized oligomers would be particularly convenient and affordable using electrochemistry. The nucleobases are electroactive (which provides a basis for ionizing radiation damage to DNA [8]); however, the electrochemistry of DNA is difficult to interpret due to two important complications. First, the diffusion coefficients of polymeric DNAs are relatively low, which decreases the amount of mass transfer-limited diffusive current that can be obtained [9]. Second, the redox reactions of the nucleobases exhibit poor reversibility [10]. Described in this chapter are approaches to circumventing these two difficulties to obtain information on DNA structure and redox reactivity. Our approach has been to use metal complexes as mediators of DNA electrochemistry. This approach requires a complete understanding of the mass transport of DNA and DNA-bound metal complexes. Here appropriate choices of voltammetric technique and analysis become important. A means of quantitatively treating the DNA binding equilibrium of the metal complex is also important because the partitioning between bound and free metal complex will dictate the efficiencies with which the mediator is oxidized by the electrode and the DNA is oxidized by the mediator. Finally, understanding how coupled chemical reactions affect the voltammetric response will allow for quantitating the real-time kinetics of DNA-metal redox reactions in a convenient and affordable manner.
< previous page
page_298
next page >
< previous page
page_299
next page > Page 299
2 Electrochemistry of One-Electron Couples Bound to DNA 2.1 General Considerations Since the direct electrochemistry of DNA is hampered by slow diffusion and irreversibility, an initial approach has involved study of a redox-active probe (usually a metal complex) noncovalently associated with DNA [9,11,12]. We will first treat the case where the metal complex probe does nor react with DNA in either of its redox forms. Initial studies were performed on the tris chelate complexes and (M = Co or Fe, bpy = 2,2'-bipyridyl, phen = 1,10-phenanthroline) [9,11]. These complexes associate relatively weakly with DNA through a combination of surface binding and electrostatic interactions [1316]. Since each redox couple can exist in two redox states that can each be bound to DNA or free in solution, the electrochemistry of these couples must be analyzed in terms of a square scheme [9]:
Scheme 1 Analysis as a thermodynamic cycle shows that changes in the binding affinity as a result of changes in the charge on the metal complex will be reflected as a shift in the redox potential according to
where
and
are the redox potentials for the bound and free forms of the metal complex, respectively. For
, the E1/2 of the 3+/2+ couple shifts negatively, indicating that the 3+ ion is bound 1.7 times more strongly than the 2+ ion. This shift is therefore consistent with a greater electrostatic attraction between the more highly charged metal cation and the DNA polyanion. In contrast, the E1/2 for the
< previous page
page_299
next page >
< previous page
page_300
next page > Page 300
couple shifts positively, indicating that the 2+ ion is bound twice as strongly as the 3+ ion. This observation has been ascribed to a hydrophobic interaction between phen and the DNA that is more important than the electrostatic interaction [13,1517]. In addition to changes in the redox potential, binding of the metal complex to DNA also leads to a dramatic decrease in the measured current. This decrease occurs because the diffusion coefficient of the metal complex is at least an order of magnitude higher than that of the DNA. In a cyclic voltammogram, the current for two interconverting bound and free species can be described as [9]:
where Ct is the total concentration of metal complex, Df is the diffusion coefficient of the free metal complex, Db is the diffusion coefficient of the bound metal complex, Xf is the mole fraction of metal complex free in solution, and Xb is the mole fraction of bound metal complex. The electrochemical constants are collected in the term B, which at 25°C is equal to 2.69 × 105n3/2Aν1/2, where n is the number of electrons in the redox couple, A is the electrode area, and ν is the voltammetric sweep rate. In principle, it should be possible to use Eq. (2) to determine the distribution of bound and free metal complex, and hence the DNA binding constant (KB) of the metal complex. In practice, this endeavor is difficult because an accurate measure of Db must be available. With metal complexes that have low affinities, such as , very large ratios of DNA-to-metal complex (R) are required to achieve a situation where all of the metal complex is bound [18]. This condition is important because small concentrations of free metal complex dramatically distort the response, since Df >> Db. We have solved this problem using the probe Os(bpy)2(dppz)2+ (1, dppz = dipyridophenazine), which, like Ru dppz complexes [1922], intercalates into DNA and binds to DNA with high affinity. Binding of metal cations to DNA can occur either via site binding or territorial binding [23]. Territorial binding is noncooperative, i.e., binding of multiple cations does not influence the measured binding constant. This mode of binding is observed for most bpy and phen metal complexes at usual (50 mM) ionic strengths and is described by the equation [15]:
< previous page
page_300
next page >
< previous page
page_301
next page > Page 301
where Cb is the concentration of bound metal complex and Ct is the concentration of free metal complex. In absorbance titrations, the measured absorbance is a linear function of the extinction coefficients of the bound and free forms (εbεf), so that [18]:
where ε is the measured extinction coefficient at an arbitrary DNA concentration. When binding constants are larger, site binding occurs because two metal complexes cannot occupy the same binding site. This equilibrium is described by [9,15]:
where s is the site size excluded by the bound metal complex. Classical intercalators exhibit a site size of 2 due to neighbor exclusion [24], and dppz complexes of Ru exhibit binding constants of 106 M1 and site sizes of 2 [15,20]. Since Os and Ru have similar radii, the complexes should be nearly isostructural and should exhibit the same binding parameters. In fact, absorbance titration of 1 gives Kb of 4.0 × 106 M1 and s = 1.9 bp. At reasonable values of R, we can therefore assume that all of 1 is bound to DNA. 2.2 Impact of DNA Diffusion 2.2.1 Cyclic Voltammetry The cyclic voltammogram (CV) of 1 obtained at a tin-doped indium oxide (ITO) working electrode is shown in Fig. 1 [25]. Tindoped indium oxide was chosen as the working electrode because DNA does not adsorb to its negatively charged surface [26]. The complex undergoes a single one-electron oxidation to the corresponding Os(III) complex at an E1/2 of 0.72 V vs. Ag/AgCl. Addition of an excess (R = 35) of calf thymus DNA causes a slight shift in the E1/2 to less positive potential. The E1/2 for the
< previous page
page_301
next page >
< previous page
page_302
next page > Page 302
Fig. 1. Cyclic voltammograms of 1 with (B) and without (A) calfthymus DNA. Os(II/III) couple shifts negatively by 32 mV, which is similar to those of related 2+/3+ couples with 3+ bound more strongly [Eq. (1)] [9,26]. There is also a dramatic decrease in current, which from Eq. (2) must be ascribed to the intercalated complex diffusing with the DNA fragment. The drop in voltammetric current to a small fraction of its original magnitude reflects the slower diffusion of the biopolymer compared to the free complex. In cyclic voltammetry, plots of the peak current (ip) vs. the square root of the scan rate (ν1/2) should be linear according to the Randles-Sevcik Eq. (6), which describes peak voltammetric current for a diffusing species in classical cyclic voltammetry.
Equation (6) gives a value of 2.5 × 106 cm2 sec1 for Df for 1. The peak current for complexes bound to calf thymus DNA at R = 35 has a much smaller ν1/2 dependence and gives a Db of 1.5 × 108 cm2 sec1. The
< previous page
page_302
next page >
< previous page
page_303
next page > Page 303
diffusion coefficient for calf thymus DNA has been measured by dynamic light scattering and can be calculated from rigid rod polymer theory [2729]. This value (see Table 1) is significantly greater than that determined by CV. Why does CV underestimate the diffusion coefficient of DNA? Using CV and a digital potentiostat to determine diffusion coefficients is complicated by two important considerations. First, digital potentiostats do not apply a linear potential ramp to the solution. In fact, cyclic voltammetry using a digital potentiostat is actually cyclic staircase voltammetry (Scheme 2) because the potential is incremented in
Scheme 2
discrete steps [30,31]. Further, the current is sampled at the end of each pulse, after significant Faradaic electrochemistry has occurred, causing lower absolute currents in cyclic staircase voltammetry compared to analog CV. Second, in cyclic voltammetry experiments, peak currents
< previous page
page_303
next page >
< previous page
page_304
next page > Page 304
for simple one-electron couples are influenced by both rates of mass transfer and rates of heterogeneous electron transfer from the electrode to the analyte [32]. Decoupling these two rates is difficult and usually prohibits reliable determination of diffusion coefficients in cases where heterogeneous kinetics are slow. We have addressed the first consideration by varying the pulse width tp during the experiment. As the pulse width is decreased to zero, a cyclic staircase voltammogram reduces to a classical CV [30]. However, decreasing the pulse width over a wide range caused no quantitative change in the current response for DNA-bound 1. The second consideration was addressed by fitting the voltammograms using a software package called the COOL algorithm [31], which separates scalar contributions, such as analyte concentration and diffusion coefficient, from the characteristic shape of the voltammogram, which is determined by the rate of heterogeneous charge transfer. The scalar contributions are then analyzed in terms of the Cottrell Eq. (7):
where A is the electrode area. After these parameters are separated, the characteristic shape of the voltammogram can be fit to a kinetic model, which in our case would involve a single heterogeneous rate constant for oxidation of the bound or free metal complex. Fitting of the bound and free voltammograms shows that the rate of heterogeneous charge transfer is fast in the free case and slow (5 × 10-4 cm sec-1) in the bound case. This limitation on kinetics is undoubtedly due to charge transfer across the greater distance imposed by the electrostatic repulsion between the DNA and the partially negative surface of the ITO working electrode. This slower rate constant would lead to a lower peak current in the voltammograms that might give a lower apparent Db. Nonetheless, even after correcting for the slow heterogeneous kinetics, the diffusion coefficients determined from COOL fitting of CV data are in agreement with the results from the ip vs. ν1/2 plots in underestimating the diffusion coefficients. Thus, even when concerns over pulse width and heterogeneous kinetics are addressed, CV still underestimates Db, by an order of magnitude. 2.2.2 Normal Pulse Voltammetry The problems associated with determining diffusion coefficients from CV can be solved in two ways. In many cases, the current at a constant
< previous page
page_304
next page >
< previous page
page_305
next page > Page 305
potential that is beyond E1/2 can be monitored as a function of time in a chronoamperometry experiment and fit to Eq. (7) to obtain the diffusion coefficient [32]. Normal pulse voltammetry takes advantage of Eq. (7) in a similar way within the context of a voltammetric experiment. As shown in Scheme 2, normal pulse involves the application of a potential for some time period, but the system is allowed to relax back to equilibrium between the pulses. Thus each pulse is equivalent to a chronoamperometric experiment, i.e., Eq. (7), where the current is sampled at the end of the pulse. With long enough pulse widths, the diffusion-limited current is always sampled, and mass transfer can be separated from electron transfer kinetics [31]. The generalized appearance of the normal pulse voltammogram (NPV) is shown in Scheme 2. The current starts out at zero, passes through a sigmoidal transition region in the vicinity of E1/2, and maximizes at a level where the current is described solely by Eq. (6). The rate of the heterogeneous charge transfer dictates the slope in the transition region, but at sufficiently high potentials the diffusion coefficient can always be determined simply from Eq. (6). NPVs obtained for 1 with and without DNA are shown in Fig. 2. As the potential is stepped through the E1/2, the current exhibits a sigmoidal transition from essentially zero to a potential-independent mass transfer limited value. From plots of the diffusion-limited current vs. t-1/2 or from COOL fits, the determined diffusion coefficients are Df = 4.0 × 10-6 cm2 sec-1 and Db = 2.0 × 107 cm2 sec1. The slower charge transfer rate for bound 1 is apparent as a decrease in the slope in the transition region, but clearly, this change does not affect the current in the diffusionlimited plateau. Kinetics and mass transfer are therefore separated and the determined diffusion coefficients are in good agreement with known values. This consistency is in contrast to the dependence of peak current in CSV that, in addition to being sensitive to adsorption, underestimates the rate of diffusion of bound species by more than 50%. Because calculation of a diffusion coefficient involves squaring the measured current, a discrepancy of an order of magnitude appears between values measured by the two techniques. Normal pulse voltammetry was also used to observe the diffusion of 1 bound to an oligonucleotide. Figure 3 shows the normal pulse voltammograms of solutions containing 1 in the absence (A) and presence (B) of (dA)20·(dT)20 at R = 40. The impact on the mass transfer rate as measured by the COOL slope is not nearly as great as with polymeric calf thymus DNA (240 bp) and the calculated diffusion coefficient is Db =
< previous page
page_305
next page >
< previous page
page_306
next page > Page 306
Fig. 2. Normal pulse voltammograms of 1 with (B) and without (A) calf thymus DNA. 1.2 × 106 cm2 sec1 for (dA)20·(dT)20. The collected results are summarized in Table 1. The agreement of the NPV results with values from theory and nonelectrochemical techniques shows that the effect of slow DNA diffusion on the electrochemical response is as expected. Furthermore, the approximation that the diffusion coefficient of the bound metal complex is the same as the diffusion coefficient of DNA itself has been confirmed. This information then provides a basis for using electrochemistry to determine DNA binding constants of redox-active probes and for studying coupled chemical reactions of DNA-bound species. 2.3 Impact of DNA Binding With an understanding of the diffusion properties available, Eq. (2) can now be used to quantitate DNA binding of reversible couples. Equation
< previous page
page_306
next page >
< previous page
page_307
next page > Page 307
Fig. 3. Normal pulse voltammograms of 1 with (B) and without (A) (dA)20·(dT)20. TABLE 1 DNA Diffusion Coefficients from Electrochemistry, Light Scattering, and Theory Technique
Calf thymus DNA (cm2 sec1)
(dA)20·(dT)20 (cm2 sec1)
Cyclic voltammetry
1.5 × 108
Normal pulse voltammetry
2.0 × 107
1.2 × 106
Light scatteringa
2.1 × 107
1.1 × 106
Theoryb
2.1 × 107
1.1 × 106
aFrom [28,29]. bFrom [27].
< previous page
page_307
next page >
< previous page
page_308
next page > Page 308
Fig. 4. Plots of
vs. DNA for
at (A) 15 mM [Na+] and (B) 75 mM [Na+]. i0 is the current in the absence of DNA (= if).
(2) shows that the distribution of bound and free species is related to the current squared. Thus the mole fraction bound can be written as
where i is the current measured at any DNA concentration, if is the current in the absence of DNA, and ib is the current of the fully bound species. This expression can now be used in conjunction with the site or territorial binding equations to quantitate KB [15]. A difficulty with quantitating small KB values (as for phen or bpy complexes) by absorption titration is that saturation is usually not achieved at reasonable DNA concentrations. In this case, εb [in Eq. (4)] must be a floating parameter in the fit. However, for normal pulse titrations we can assume that ib will be governed by the diffusion coefficient of DNA, which we have measured. Titrations can then be fit according to
< previous page
page_308
next page >
< previous page
page_309
next page > Page 309
where is known and KB is the only floating parameter. Such a fit for is shown in Fig. 4 at two different ionic strengths. The determined binding constants are KB = 95 M1 at [Na+] = 75 mM and KB = 950 M1 at [Na+] = 15 mM. These values are in good agreement with values from fluorescence titration, and the dependence on ionic strength is as predicted from polyelectrolyte theory [15]. An advantage of electrochemical titration is that (as is evident in Fig. 1) the changes in current as a percentage of the total signal are much larger than analogous changes in absorbance bands upon binding. For example, a KB of 102 M1, as seen in Fig. 4, is difficult to determine by absorption titration [18]. 3 Electrochemically Activated DNA Cleavage 3.1 Reductive Activation A variety of DNA cleaving molecules may be activated via electrochemical reduction in the presence of dioxygen. Hecht and coworkers studied the electrochemical activation of oxygenated Fe-bleomycin [33,34], which mediates the oxidative cleavage of DNA. The cyclic voltammogram of Fe-bleomycin in the absence of dioxygen is quasireversible (∆E = 0.10 V) with an E1/2 of 0.08 vs. Ag/AgCl. In the presence of dioxygen, however, significant enhancement of the reductive current is observed, consistent with formation of an ''activated" Fe-bleomycin/O2 complex that is more prone to reduction. Anaerobic cyclic voltammograms of Fe-bleomycin in the presence of DNA show a significant decrease in current, indicative of electrostatic binding of the Fe-bleomycin to DNA. Electrolysis of Fe-bleomycin solutions in the presence of dioxygen and oligonucleotides followed by product analysis showed the resulting product ratios to be identical to those obtained by chemical activation of Fe-bleomycin. The proposed catalytic cycle for Fe-bleomycin activation and DNA cleavage is shown in Scheme 3. Bard and coworkers studied DNA cleavage by electrochemically activating MnIII and FeIII porphyrin complexes [35]. These complexes behave much the same as the Fe-bleomycin system and are capable of DNA cleavage upon reductive activation in the presence of dioxygen.
< previous page
page_309
next page >
< previous page
page_310
next page > Page 310
Scheme 3 Bard and coworkers showed that these complexes interact strongly with DNA due primarily to their high positive charge. Reductive electrolysis of solutions in the presence of dioxygen caused significant cleavage of plasmid DNA. 3.2 Oxidative Activation 3.2.1 Direct Electrochemistry The interaction of nucleic acids with charged surfaces is extremely important in both biological and analytical systems. For this reason, a large body of work has focused on the direct electrochemistry of DNA, RNA, and other biomolecules adsorbed to an electrode surface [10,3652]. The primary interest has been in determining the properties and structure of the adsorbed biomolecule. The reductive electrochemistry of DNA adsorbed to mercury or graphite electrodes has been shown to be highly sensitive in the determination of submicrogram quantities of DNA [52]. However, as this chapter is primarily concerned with the oxidative electrochemistry of DNA, the extensive literature on reduction of nucleic acids will not be covered. Whereas the reductive chemistry of DNA involves reduction of cytosine and thymine residues [4952], the guanine and adenine residues are found to be the most easily oxidized bases in DNA. Using differential pulse voltammetry at pyrolytic graphite electrodes, two
< previous page
page_310
next page >
< previous page
page_311
next page > Page 311
well-resolved peaks are observed for calf thymus DNA, corresponding to the oxidation of guanine (Ep = 0.89 V vs. SCE) and adenine (Ep = 1.17 V) residues [10,36,37]. The anodic peaks observed for oxidation of adsorbed DNA are found to be highly sensitive to the nucleic acid structure. Comparison of native and denatured forms of various nucleic acids shows significantly larger amounts of current for the denatured form [10,3638]. This difference has been ascribed to the reduced flexibility of the native (double-stranded) form, which precludes close contact of the base residues to the electrode surface. In addition, small lengths of DNA tend to give larger currents, also due to the greater contact with the electrode surface [42]. In thermal denaturation studies, Brabec [38] showed that the ''A" peak in the differential pulse voltammogram increased at a lower temperature than the "G" peak, indicating that the A-T rich regions of the DNA melt at a lower temperature than the G-C rich regions. More recent work has probed the extent of DNA denaturation as a function of adsorption time and adsorption potential [47]. A related set of studies has used this technique to probe the binding of netropsin to natural and synthetic nucleic acids [39,45]. Netropsin is known to bind in A-T-rich regions of DNA, and it was shown that the A peak in the differential pulse voltammogram is significantly reduced whereas the G peak changes are much less pronounced. Analysis of the relative peak heights (A and G) for several different natural and synthetic forms of DNA showed that the peaks heights are directly related to the relative percent G-C content [41], which has been proposed as a rapid and convenient method for the determination of percent G-C content in various forms of DNA. In addition, Tolbert et al. [53] developed a type of chemically modified electrode that selectively oxidizes ribonucleosides such as cytidine, uridine, adenosine, and guanosine but does not oxidize deoxyribonucleosides such as thymidine. 3.3 Metal-Mediated Oxidation 3.3.1 Oxo and Hydrogen Transfer
Complexes based on DNA based on the redox couples [26,5456]:
< previous page
(tpy = 2,2',2"-terpyridyl) mediate the electrochemical cleavage of
page_311
next page >
< previous page
page_312
next page > Page 312
where E1/2(III/II) = 0.49 V and E1/2(IV/III) = 0.62 V at pH 7. The Ru(tpy)(bpy)O2+ complex is capable of cleaving DNA by two redox pathways: sugar oxidation at all nucleotides at the 1' position and inner-sphere oxidation of guanosine nucleotides at the base [22,57,58]. Electrocatalytic oxidation can be observed through application of a and DNA. These reactions lead to detectable cleavage on potential of 0.8 V to a solution of sequencing and plasmid gels. An important implication of Eq. (2) is that when bpy is replaced with ligands that impart higher affinities (such as phen and dppz), longer electrolysis times are required to achieve comparable cleavage [26,55,56]. Stronger binding of the complex leads to reduced current, so that less of the active oxidant Ru(tpy)(bpy)O2+ is generated. This observation emphasizes the need for a complete understanding of DNA diffusion and weak binding equilibria because weak binding complexes will be the most effective cleavage electrocatalysts. As stated at the outset, an important goal in metal-mediated electrochemical DNA cleavage has been to use voltammetric data to obtain rate information on the oxidation of DNA by the mediator. The most convenient way to obtain this kinetic information would be through the observation of catalytic current in cyclic voltammograms. Catalytic currents are obtained when an exogenous agent rereduces the metal complex multiple times during a voltammetric scan, so that the reduced form is continually reoxidized by the electrode. The degree of catalytic enhancement in the current is then directly related to the rate of the chemical reaction. Unfortunately, catalytic currents are not observed with Ru(tpy)(bpy)O2+ because the rates of the hydrogen atom and oxo transfer reactions are too slow to allow multiple turnovers during sweeps at common scan rates. 3.3.2 Outer-Sphere Electron Transfer We have recently undertaken the study of trans-[Rev(O)2L4]+ (L = substituted pyridine) systems and their interaction with DNA [59]. These complexes are oxidized to the corresponding Re(VI) complexes at potentials between 0.5 and 1.5 V, depending on L [60,61]. The trans[Re(O)2(4-OMe-py)4]+ complex (4-OMe-py = 4methoxypyridine), which
< previous page
page_312
next page >
< previous page
page_313
next page > Page 313
undergoes reversible one-electron electrochemistry (E1/2(VI/V) = 1.00 V) in buffered aqueous solution [60], displays dramatic catalytic current enhancement in the presence of DNA. As shown in Figure 5 (upper part), a reversible cyclic voltammogram is obtained for [Re(O)2(4-OMe-py)4]+ alone. Upon addition of calf thymus DNA, a large catalytic en-
Fig. 5. Cyclic voltammogram of trans-[Re(O)2(4-OMe-py)4]+ with (dashed) and without (solid) calf thymus DNA. The scan rate was 25 mV/sec and the buffer contained 100 mM NaCl and 5 mM phosphate buffer (pH 7).
Cyclic voltammogram of trans-[Re(O)2(4-OMe-py)4]+ with poly(dA)·poly(dT) (solid) and poly(dG)·poly(dC) (dashed). (Reprinted with permission from [59].)
< previous page
page_313
next page >
< previous page
page_314
next page > Page 314
hancement is seen in the forward (anodic) wave and the return wave nearly retraces the forward wave. This typical catalytic wave is indicative of efficient oxidation of the DNA by [ReVI(O)2(4-OMe-py)4]2+. Further studies show that this large catalytic enhancement is not observed in the presence of poly(dA)·poly(dT), but is observed for poly(dG)·poly(dC) (Figure 5; lower part). Guanine is the most easily oxidized of the nucleic acid bases, both by oxo transfer and by outer-sphere, one-electron oxidation [6264]. That the mechanism of DNA oxidation is outer-sphere electron transfer is confirmed by the observation that (E1/2(III/II) = 1.02 V) and other tris chelate complexes with high enough potentials, which cannot participate in inner-sphere reactions, give results identical to those shown in Figure 5 (5-Cl-phen = 5-chloro-1,10-phenanthroline). Controlled-potential electrolysis at 1.2 V of φX174 plasmid DNA in the presence of [Re(O)2(4-OMe-py)4]+ induces relaxation of the plasmid DNA from form I (supercoiled) to form II (nicked circular). More important, similar electrolysis using the 5'-end 32P-labeled oligonucleotide d(5'-TACGCAAGGGCAT-3') produces piperidine-labile lesions specifically at guanine; no cleavage was observed without piperidine treatment. Recently, it has been shown that one-electron oxidation of guanine leads to the formation of 8-oxo-guanine via hydrolysis of the guanine radical cation [65]. The presence of 8-oxo-guanine is known to produce a base-labile lesion in DNA [62,66]. As discussed above, the observation of a significant catalytic current also depends on a relatively low binding affinity for the metal complex, and we have determined the binding affinity of [Re(O)2(4-OMe-py)4]+ to be only 10 M1 by emission titration using published procedures [18]. 4 Oxidation Kinetics from Voltammetry One of the attractive features of studying electrochemical activation of nucleic acid oxidation is the ability to obtain kinetic information from the experimental current response. The classic work of Nicholson and Shain [67] relates the cyclic voltammetry current response to kinetic parameters for a variety of mechanisms. In addition, with the advent of powerful desktop computers, software such as the COOL package [31] is
< previous page
page_314
next page >
< previous page
page_315
next page > Page 315
able to analyze pulse voltammetric data (cyclic voltammetry, square-wave voltammetry, normal pulse voltammetry) to obtain kinetic parameters. Unfortunately these two techniques are only applicable to those relatively simple mechanisms for which an analytical solution exists, prohibiting analysis of the complicated mechanisms which often occur with reactions involving DNA. The simplest mechanism for such electrocatalytic DNA oxidation is shown in Scheme 4. The mechanism consists of an electrochemical
Scheme 4 oxidation (E) step, and a homogeneous chemical oxidation step (C') which regenerates the initial reduced species. The mechanism shown in Scheme 4 has been analyzed extensively [67,68] and the theoretical current response is known. However, this scheme requires that the substrate be present in excess so as to produce pseudo-first-order conditions. This condition is difficult to obtain with DNA due to the relatively low solubility of DNA. A mechanism which takes the second-order oxidation reaction into account (Scheme 5) is therefore more suitable.
Scheme 5 The mechanism shown in Scheme 5 does not have a known analytical solution. Adding to this complexity are the following: (1) this scheme does not take into account the possible effects of electrostatic binding and (2) this system contains components with widely differing diffusion coefficients. Using the CV simulation package DigiSim [69], virtually any mechanism including those with second-order reactions and widely different diffusion coefficients can be simulated. We successfully analyzed our data using the mechanism shown in Scheme 5 as well as variations of this mechanism. Of course, Scheme 5 does not account for binding of the metal
< previous page
page_315
next page >
< previous page
page_316
next page > Page 316
complex to DNA, as would be required according to Scheme 1. Extension of Scheme 1 to include the one-electron oxidation of the DNA by the bound metal complex gives the mechanism shown in Scheme 6. This
Scheme 6 scheme is fundamentally similar to Scheme 5 but considers electrostatic binding. Simulations of this scheme show very close agreement with data collected for oxidation of DNA by a variety of
complexes [70].
In spite of close qualitative agreement between data and simulations found using Scheme 6, analysis of large sets of data is rather unwieldly. Application of the steady-state approximation to Scheme 6 in the limit of low binding constants allows analysis using the two-step mechanism shown in Scheme 5. Experimentally this simplification is accomplished by increasing the ionic strength of the buffered aqueous solution [15]. Simulations using both schemes show identical current responses for equivalent oxidation rates under the correct experimental conditions ([Na+] = 780 mM). Use of this simplified scheme allows straightforward fitting of the experimental data and determination of rate information. Shown in Fig. 6 are the cyclic voltammograms of with and without calf thymus DNA, showing the catalytic enhancement. In this case of high ionic strength, the current can be analyzed in terms of the modified EC' mechanism (Scheme 5). Cyclic voltammograms were analyzed by fitting the complete current-potential curves using the DigiSim analysis package [69]. The input parameters were the E1/2
< previous page
page_316
next page >
< previous page
page_317
next page > Page 317
Fig. 6. Cyclic voltammograms of
and calf thymus DNA. (A) Scan of
at 25 mV/sec in 700 mM NaCl/50 mM Na-phosphate buffer. and 3.0 mM (nucleotide) calf thymus (B) Voltammogram of DNA (solid line). (B) Simulated cyclic voltammogram (broken line) using DigiSim with kox = 9.0 × 103 M1 sec1. for the metal complex and the diffusion coefficients for the metal complex and the DNA, all of which are well determined in separate experiments. Therefore the sole parameter obtained from the fit was the second-order rate constant for DNA oxidation, k = 9.0 × 103 M1 sec1. The same rate constant was determined over a wide range of scan rates, and the high quality of the fit is apparent in Fig. 6. The same rate constant was confirmed by stopped-flow spectrophotometry on reactions of authentic with calf thymus DNA and by square wave voltammetry under pseudo-first-order conditions and fitting with COOL [70]. Understanding the dependence of electron transfer rates on distance and driving force is an area of active experimentation and high theoretical sophistication [7177]. If the driving force for electron transfer is significantly less than the reorganization energy (λ), a plot of
< previous page
page_317
next page >
< previous page
page_318
next page > Page 318
Fig. 7. Driving force dependence of the rate constant for electron transfer from guanine in calf thymus DNA to the metal complex. The best fit gave a slope of 0.49. Rate constants were determined by cyclic voltammetry (circles), square wave voltammetry (squares), and stopped flow spectrophotometry (triangles). RT ln k vs. driving force (when corrected for work terms associated with approach of the reactants) should yield a straight line with a slope of 1/2 [73]. The rate constants for oxidation of DNA by a number of derivatives were determined, and a Marcus plot of these rate constants and driving forces is shown in Fig. 7 where the slope and linearity are in excellent agreement with the theoretical prediction. This result shows that DNA(guanine)-metal electron transfer follows Marcus theory, which provides a firm theoretical basis for future analyses of rate data. 5 Conclusions The studies of electrochemistry of DNA-bound metal complexes described here show that using voltammetry to study DNA-metal redox
< previous page
page_318
next page >
< previous page
page_319
next page > Page 319
reactions relies on a complete understanding of four phenomena. First, diffusion coefficients for DNA must be available, since the large difference in diffusion coefficient for DNA and the metal complex complicates the analysis of currents obtained in voltammograms. Second, the binding equilibrium (either site or territorial) must be understood including a quantitative determination of the binding constant (and site size if appropriate). Third, the heterogeneous rate constant for electron transfer to the mediator must be known. Finally, the mechanism of coupled chemical reactions must be available to allow for extraction of rate data from the voltammetry. We have identified systems that are understood at these levels where continued experimentation may lead to the ability to gather in real time voltammetric rate data that reflects key structural or reactivity properties of nucleic acids. Acknowledgment We thank our talented colleagues who began this work, especially Neena Grover, Bill Kalsbeck, Greg Neyhart, Kathy Glasgow, and Nishi Gupta. Our work in this area has been supported by the David and Lucile Packard Foundation and the National Science Foundation. Abbreviations
1
Os(bpy)2(dppz)2+
A
electrode area
BLM
bleomycin
bpy
2,2'-bipyridyl
Cb
concentration of bound metal complex
Ct
concentration of free metal complex
CSV
cyclic staircase voltammetry
CV
cyclic voltammetry
Db
diffusion coefficient of bound metal complex
Df
diffusion coefficient of free metal complex
dppz
dipyridophenazine
E1/2
redox potential
< previous page
page_319
next page >
< previous page
page_320
next page > Page 320
F
Faraday's constant
ib
current due to bound metal complex
if
current due to free metal complex
ip
peak current
ITO
tin-doped indium oxide
KB
DNA binding constant
v
scan rate
NPV
normal pulse voltammogram
4-OMe-py
4-methoxypyridine
phen
1,10-phenanthroline
R
ratio of DNA (nucleotide phosphate) to metal complex
s
site size
SCE
standard calomel electrode
tp
pulse width
tpy
2,2',2''-terpyridyl
X
mole fraction
References 1. T. Ried, A. Baldini, T. C. Rand, and D. C. Ward, Proc. Natl. Acad. Sci. USA, 89, 13881392 (1992). 2. Z. Du, L. Hood and R. K. Wilson, Meth. Enzymol., 218, 104121 (1993). 3. R. Kaiser, T. Hunkapiller, C. Heiner, and L. Hood, Meth. Enzymol., 218 (1993). 4. Y. Jenkins and J. K. Barton, J. Am. Chem. Soc., 114, 87368738 (1992). 5. S. P. A. Fodor, R. P. Rava, X. C. Huang, A. C. Pease, C. P. Holmes, and C. L. Adams, Nature, 364, 555556 (1993).
6. D. Noble, Anal. Chem., 67, 201A204A (1995). 7. W. Bains, Chem. Br., 122125 (1995). 8. S. Steenken, Chem. Rev., 89, 503520 (1989). 9. M. T. Carter, M. Rodriguez, and A. J. Bard, J. Am. Chem. Soc., 111, 8901 (1989). 10. V. Brabec and G. Dryhurst, J. Electroanal. Chem. Interfac. Electrochem., 89, 161173 (1978).
< previous page
page_320
next page >
< previous page
page_321
next page > Page 321
11. M. J. Carter and A. J. Bard, J. Am. Chem. Soc., 109, 75287530 (1987). 12. M. T. Carter and A. J. Bard, Bioconj. Chem., 1, 257 (1990). 13. A. M. Pyle, J. P. Rehmann, R. Meshoyrer, C. V. Kumar, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 111, 3051 (1989). 14. A. M. Pyle and J. K. Barton, Prog. Inorg. Chem., 38, 413 (1990). 15. W. A. Kalsbeck and H. H. Thorp, J. Am. Chem. Soc., 115, 71467151 (1993). 16. W. A. Kalsbeck and H. H. Thorp, Inorg. Chem., 33, 34273429 (1994). 17. S. Satyanarayana, J. C. Dabrowiak, and J. B. Chaires, Biochemistry, 31, 9319 (1992). 18. S. R. Smith, G. A. Neyhart, W. A. Kalsbeck, and H. H. Thorp, New J. Chem., 18, 397406 (1994). 19. R. E. Holmlin and J. K. Barton, Inorg. Chem., 34, 78 (1995). 20. C. Hiort, P. Lincoln, and B. Nordén, J. Am. Chem. Soc., 115, 3448 (1993). 21. A. E. Friedman, J. C. Chambron, J. P. Sauvage, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 112, 4960 (1990). 22. G. A. Neyhart, N. Grover. S. R. Smith, W. A. Kalsbeck, T. A. Fairley, M. Cory, and H. H. Thorp, J. Am. Chem. Soc., 115, 4423 (1993). 23. G. S. Manning, Acc. Chem. Res., 12, 443 (1979). 24. K. W. Jeanette, S. J. Lippard, G. A. Vassiliades, and W. R. Bauer, Proc. Natl. Acad. Sci. USA, 71, 3839 (1974). 25. T. W. Welch, A. H. Corbett, and H. H. Thorp, J. Phys. Chem., 99, 1175711763 (1995). 26. N. Grover, N. Gupta, P. Singh, and H. H. Thorp, Inorg. Chem., 31, 2014 (1992). 27. M. A. Tracy and R. Pecora, Annu. Rev. Phys. Chem., 43, 525 (1992). 28. W. Eimer and R. Pecora, J. Phys. Chem., 94, 2324 (1991). 29. H. T. Goinga and R. Pecora, Macromolecules, 24, 6128 (1991). 30. P. He, Anal. Chem., 67, 986992 (1995). 31. J. Osteryoung, Acc. Chem. Res., 26, 7783 (1993). 32. A. J. Bard and L. R. Faulkner, Electrochemical Methods, John Wiley and Sons, New York, 1980.
< previous page
page_321
next page >
< previous page
page_322
next page > Page 322
33. R. B. Van Atta, E. C. Long, S. M. Hecht, G. A. van der Marel, and J. H. van Boom, J. Am. Chem. Soc., 111, 2722 (1989). 34. S. M. Hecht, E. C. Long, R. B. Van Atta, E. De Vroom, and B. J. Carter, On the Mechanism of Bleomycin Activation and Polynucleotide Strand Scission (C. Bleasdale and B. T. Goldings, eds.), Mol. Mech. Bioorg. Processes (Royal Soc. Chem., Cambridge, UK, 1989), pp. 100114. 35. M. Rodriguez, T. Kodadek, M. Torres, and A. J. Bard, Bioconj. Chem., 1, 123 (1990). 36. V. Brabec and G. Dryhurst, Stud. Biophys., 67, 2324 (1978). 37. V. Brabec, Biophys. Chem., 9, 289297 (1979). 38. V. Brabec, Biopolymers, 18, 23972404 (1979). 39. V. Brabec, Stud. Biophys., 81, 99100 (1980). 40. V. Brabec, Radiat. Environ. Biophys., 17, 129141 (1980). 41. V. Brabec, Bioelectrochem. Bioenerg., 7, 6982 (1980). 42. V. Brabec and J. Koudelka, Bioelectrochem. Bioenerg., 7, 793805 (1980). 43. V. Brabec, Stud. Biophys., 82, 1726 (1981). 44. V. Brabec, Biophys. Chem., 13, 187191 (1981). 45. V. Brabec, Bioelectrochem. Bioenerg., 9, 245252 (1982). 46. E. Palecek, in Topics in Bioelectrochemistry and Bioenergetics, Vol. 5 (G. Milazzo, ed.), John Wiley and Sons, Chichester, 1983, p. 65. 47. V. Brabec, Bioelectrochem. Bioenerg., 11, 245255 (1983). 48. V. Brabec, V. Glezners, and V. Kadysh, Collect. Czech. Chem. Commun., 48, 12571271 (1983). 49. E. Palecek, Bioelectrochem. Bioenerg., 15, 275295 (1986). 50. E. Palecek, Bioelectrochem. Bioenerg., 20, 179194 (1988). 51. E. Palecek, V. Kolár, and F. Jelen, Bioelectrochem. Bioenerg., 23, 285299 (1990). 52. E. Palecek, F. Jelen, C. Teijeiro, V. Fucík, and T. M. Jovin, Anal. Chim. Acta, 273, 175186 (1993). 53. A. M. Tolbert, R. P. Baldwin, and L. M. Santon, Anal. Lett., 22, 683702 (1989). 54. N. Grover and H. H. Thorp, J. Am. Chem. Soc., 113, 7030 (1991). 55. N. Gupta, N. Grover, G. A. Neyhart, W. Liang, P. Singh, and H. H. Thorp, Angew. Chem. Int. Ed. Engl., 31, 1048 (1992).
< previous page
page_322
next page >
< previous page
page_323
next page > Page 323
56. N. Gupta, N. Grover, G. A. Neyhart, P. Singh, and H. H. Thorp, Inorg. Chem., 32, 310 (1993). 57. C.-C. Cheng, J. G. Goll, G. A. Neyhart, T. W. Welch, P. Singh, and H. H. Thorp, J. Am. Chem. Soc., 117, 29702980 (1995). 58. G. A. Neyhart, C.-C. Cheng, and H. H. Thorp, J. Am. Chem. Soc., 117, 14631471 (1995). 59. D. H. Johnston, C.-C. Cheng, K. J. Campbell, and H. H. Thorp, Inorg. Chem., 33, 63886390 (1994). 60. J. C. Brewer and H. B. Gray, Inorg. Chem., 28, 33343336 (1989). 61. J. C. Brewer, H. H. Thorp, K. M. Slagle, G. W. Brudvig, and H. B. Gray, J. Am. Chem. Soc., 113, 3171 (1991). 62. X. Chen, C. J. Burrows, and S. E. Rokita, J. Am. Chem. Soc., 113, 5884 (1991). 63. S. V. Jovanovic and M. G. Simic, J. Phys. Chem., 90, 974978 (1986). 64. L. P. Candeias and S. Steenken, J. Am. Chem. Soc., 114, 699704 (1992). 65. H. Kasai, Z. Yamgizumi, M. Berger, and J. Cadet, J. Am. Chem. Soc., 114, 9692 (1992). 66. X. Chen, C. J. Burrows, and S. E. Rokita, J. Am. Chem. Soc., 114, 322 (1992). 67. R. S. Nicholson and I. Shain, Anal. Chem., 36, 706723 (1964). 68. J. J. O'Dea, J. Osteryoung, and R. A. Osteryoung, Anal. Chem., 53, 695701 (1981). 69. M. Rudolph, D. P. Reddy, and S. W. Feldberg, Anal. Chem., 66, 589A600A (1994). 70. D. H. Johnston, K. C. Glasgow, and H. H. Thorp, J. Am. Chem. Soc., 117, 89338938 (1995). 71. D. N. Beratan, J. N. Onuchic, J. R. Winkler, and H. B. Gray, Science, 258, 17401741 (1992). 72. B. E. Bowler, A. L. Raphael, and H. B. Gray, Prog. Inorg. Chem., 38, 259322 (1990). 73. T. M. McCleskey, J. R. Winkler, and H. B. Gray, J. Am. Chem. Soc., 114, 69356937 (1992). 74. D. S. Wuttke and H. B. Gray, Curr. Opin. Struct. Biol., 3, 555563 (1993).
< previous page
page_323
next page >
< previous page
page_324
next page > Page 324
75. C. C. Moser, J. M. Keske, K. Warncke, R. S. Farid, and P. L. Dutton, Nature, 355, 796802 (1992). 76. D. S. Wuttke, M. J. Bjerrum, J. R. Winkler, and H. B. Gray, Science, 256, 10071009 (1992). 77. S. M. Risser, D. N. Beratan, and T. J. Meade, J. Am. Chem. Soc., 115, 25082510 (1993).
< previous page
page_324
next page >
< previous page
page_325
next page > Page 325
11 Electron Transfer between Metal Complexes Bound to DNA: Is DNA a Wire? Eric D. A. Stemp and Jacqueline K. Barton Division of Chemistry and Chemical Engineering and the Beckman Institute, 127-72, California Institute of Technology, Pasadena, CA 91125, USA
326
1. Introduction
329
2. Early Studies with Organic and Transition Metal Donors and Acceptors
332
3. Metallointercalators As Donors and Acceptors
333
3.1. Dipyridophenazine Complexes of Ruthenium
3.2. Bis(phenanthroline)(dipyridophenazine)osmium(II): A Fast Red Emitting DNA Light Switch
3.3. Phenanthrenequinone Diimine Complexes of Rhodium(III) As Acceptors
4. Electron Transfer Quenching of DNA-Bound Metallointercalators
4.1. Intercalators vs. Groove-Bound Complexes
336
338
338
341
4.2. Electron Transfer at a Fixed Distance
4.3. Direct Spectral Evidence for Electron Transfer
< previous page
335
page_325
347
next page >
< previous page
page_326
next page > Page 326
5. Fast Spectroscopy of DNA-Mediated Electron Transfer
350
350
5.1. Loading Independent Rates 5.2. Sensitivity in Electron Transfer to π Stacking
353
354
5.3. Variations in DNA Sequence
6. Comparisons of Metallointercalating Systems with Others
356
358
7. Theory
360
8. Future Studies
361
Acknowledgments
361
Abbreviations
362
References
1 Introduction Long-range transport of electrons between metal centers in biological matrices is fundamental to essential life processes such as photosynthesis and oxidative phosphorylation. While the effects of distance and driving force on electron transfer through proteins are well documented [15], the influence of the intervening medium between redox centers is not well understood. Substantial progress has been made recently in predicting the relative efficiencies of electron transfer pathways through proteins using the σ bond as a benchmark [6]. However, while organic model systems have suggested that π pathways have an enhanced ability to conduct electrons [7,8], the role of aromatic moieties in facilitating biological electron transfer remains unclear. For example, two studies with cytochrome c have shown that the presence of an aromatic residue in predicted electron transfer pathways is not essential to either intraprotein [9] or interprotein [10] electron transfer. However, these studies addressed the importance of only one isolated aromatic residue on the rates of electron transfer. Considerable evidence exists in solid state materials showing that extended stacked π networks such as those found in one-dimensional conductors comprising stacked phthalocyanines clearly facilitate charge transport [11,12].
< previous page
page_326
next page >
< previous page
page_327
next page > Page 327
Double-helical DNA provides an ideal molecular medium for charge mediation because of its highly ordered stack of electronically coupled bases (Fig. 1). This biological π stack furthermore offers an amenable system for study in that double-helical DNA is structurally well described and may be characterized using common biochemical
Fig. 1. A B-DNA double helix, by looking down (top), and perpendicular (bottom) to the helical axis.
< previous page
page_327
next page >
< previous page
page_328
next page > Page 328
methodologies [13,14]. Moreover, fast charge transport through DNA may have physiological relevance either in the mediation of or protection from free radical damage often associated with nucleic acid-based disease. Scientists with a range of perspectives have regarded the DNA helix as a possible medium for long-range electron transport almost since the first proposal of its double-helical π-stacked structure [15]. DNA conductivity, measured on oriented fibers [16], was suggested to be an element of carcinogenesis, and radiation biologists have found that neutron bombardment of DNA at 77 K leads to radicals capable of migrating up to 200 base pairs along the DNA polymer [17]. More recently, the disproportionation of the antitumor drug daunomycin semiquinone into its quinone and hydroquinone forms has been attributed to electron transfer through the helix [18], and electron transport between nitroacridines has also been found to occur at significant rates [19]. Thus, it is not surprising that the examination of the DNA helical polymer as an efficient medium for long-range electron transfer has become an area of burgeoning interest for chemists. Researchers have recently focused on systematic studies of electron transfer between small donors and acceptors bound to DNA. Figure 2 illustrates a general scheme for a photoinduced electron transfer cycle between donors and acceptors bound to the DNA duplex. Studies have been carried out using both organic and transition metal donors and acceptors bound to DNA in a variety of modes: electrostatically, by intercalation, in a groove-bound mode, as well as covalently attached. Here, rather than
Fig. 2. A generalized DNA-mediated photoinduced electron transfer cycle between donors (D) and acceptors (A).
< previous page
page_328
next page >
< previous page
page_329
next page > Page 329
making an exhaustive review of work carried out [20], we focus on studies from our own laboratories to illustrate some lessons that have been learned and some of the general chemical features that are emerging concerning this DNA wire. 2 Early Studies with Organic and Transition Metal Donors and Acceptors Some of the earliest chemical studies utilized the ubiquitous organic intercalator and nucleic acid stain ethidium bromide as a photophysical probe of DNA-mediated electron transfer. Investigating the binding properties of antitumor drugs, Baguley et al. found that amsacrine quenched the fluorescence of DNA-bound ethidium, presumably by a reductive mechanism [21,22]. DNA was also found to enhance by 103 the rates for oxidative quenching of ethidium by methylviologen [23]. More recently, the quenching of ethidium and acridine orange by N,N'-dimethyl-2,7-diazapyrenium (DAP2+) in DNA has been shown also to proceed via electron transfer [24]. Realization of the shape complementarity between DNA and chiral octahedral transition metal complexes [25] led us to studies of electron transfer between metal centers using DNA as the medium. Tris(phenanthroline)ruthenium(II), an analog of the exhaustively studied
, was among the first metal complexes to be used as a nucleic acid
probe [2630]. In the presence of DNA, shows an enhancement in luminescence, hypochromism in the metal-to-ligand charge transfer (MLCT) absorption band, and the retention of polarization in emission on the microsecond time scaleall consistent with an intimate association of the complex with the biopolymer. With a d6 configuration, is inert to substitution and racemization, and can be resolved into its enantiomers. The stereoselectivity exhibited in these photophysical experiments and in DNA helical unwinding assays strongly suggested that associates with B-form DNA through at least two noncovalent binding modes, i.e., intercalation and surface binding, as illustrated in Fig. 3. Specifically, in the right-handed DNA helix, the intercalative binding mode was found to favor the ∆ isomer and a hydrophobic surface binding interaction against the minor
< previous page
page_329
next page >
< previous page
page_330
next page > Page 330
Fig. 3. Illustration of noncovalent binding modes for complexes bound to duplex DNA. Shown are the Λ isomer (above) bound against the minor groove in a surface-bound mode, and the ∆ isomer (below) bound by intercalation with one ligand partially inserted into the base stack in the major groove. groove was found to favor the complementary Λ isomer. This trend was confirmed by subsequent nuclear magnetic resonance (NMR) studies on a series of isostructural tris(phenanthroline) metal complexes bound to an oligonucleotide duplex [31,32]. The enantioselectivity was found to vary with sequence and ionic strength, depending on the relative contributions of intercalation vs. surface binding, and, under typical condi-
< previous page
page_330
next page >
< previous page
page_331
next page > Page 331
tions for study, both isomers were found to bind B-DNA somewhat weakly (Kb≤ 104 M1 cm1), with the ∆ isomer being favored. Tris(phenanthroline)metal complexes therefore became the logical first choices for investigations of DNA-dependent electron transfer between metal complexes, given the wealth of spectroscopic information available on their electron transfer reactions in solution [33] and on their interactions with DNA. The quenching of acceptors was found to be strongly accelerated in the presence of DNA [34]. Using ∆- and
by Co(III) as the
photoexcited electron donors, it was shown that quenches the ∆ isomer more efficiently than the Λ isomer, providing the first example of stereoselective electron transfer in the presence of DNA. It was also found that the quenching efficiency was inversely proportional to the DNA strand length. This result suggested that diffusion might play a major role in the accelerating effect observed, since a noncovalent association of the metal complexes with the helix could lead to a reduction in a diffusional search from three dimensions in bulk solution to approximately one dimension along the DNA polymer. The enhancements in electron transfer efficiency brought about by DNA could therefore be understood as a combination of several factors: (1) the increase in local concentration of metal complexes associated with the helix; (2) the increased efficiency of noncovalently bound metal complexes in diffusing along the helix; and (3) the possibility of electron transfer at long range through the π stack. These possibilities were explored in a more detailed analysis of quenching by a series of metal quenchers, [M = Co(III), Rh(III), or Cr(III)], in solution and in a frozen glass [35]. These studies demonstrated that the quenching reaction could be partitioned between the intercalative and surface-bound modes, with the latter mode being quenched more efficiently. At higher metal:DNA ratios, where considerations of relative binding affinity do not dominate, quenching of enantiomers occurs preferentially for the Λ isomer, the isomer favored for surface binding. This effect may be distinguished from that observed in the pilot study, in which the lower overall concentrations led to stereoselectivities that reflected the relative binding affinities of the enantiomers for the helix. The effect of concentration was also examined in a subsequent study, which suggested that once local concentration effects were taken into account, the rate of electron transfer in the vicinity of DNA might actually be even lower than in bulk solution [36].
< previous page
page_331
next page >
< previous page
page_332
next page > Page 332
To minimize the effects of diffusion along the helix and instead focus on the issue of long-range electron transfer through the π stack, the quenching was also studied under conditions of restricted mobility [35]. In aqueous glycerol at 253 K, both surface-bound and intercalated
molecules remained substantially quenched by
. In addition, the quenching had to occur over long distances, as the low loadings of donor and acceptor on the helix corresponded to an average separation of >15 Å. These observations suggested that electron transfer through DNA may indeed display a weak dependence on distance. The most efficient quencher among isostructural the acceptor of intermediate driving force, perhaps because its electronic states are acceptors was particularly well coupled to the excited donor via the DNA π way. These studies clearly demonstrated, then, the utility of combining chiral octahedral metal complexes with DNA to form efficient electron transfer assemblies and suggested that the π-stacked bases of DNA could act as a lowresistance medium through which electrons could flow. Nonetheless in these studies the competition between intercalative and surface binding modes, and hence between π-mediated and diffusional quenching, complicated the system. The effects of different binding modes on long-range charge transport also needed greater consideration in our understanding of studies with organic donors and acceptors. The range and variety of complexes with different binding interactions available using coordination chemistry actually underscores the utility of transition metal chemistry in characterizing this system. To reveal the true role of the DNA π-stacking interactions in electron transfer, then, it became clear that the next generation of studies might best be accomplished using transition metal donors and acceptors which bind DNA exclusively and tightly by intercalation, and ideally using compounds whose intercalative interactions with DNA were structurally well characterized. 3 Metallointercalators As Donors and Acceptors In order to explore how the DNA π stack may mediate long-range electron transfer, our recent work has focused on donors and acceptors
< previous page
page_332
next page >
< previous page
page_333
next page > Page 333
which bind double-helical DNA avidly through a π-stacked intercalative interaction. We have utilized dipyridophenazine (dppz) complexes of ruthenium(II) and osmium(II), [Ru(phen)2dppz]2+ and [Os(phen)2dppz]2+, as the photoexcited donors and 9,10-phenanthrenequinonediimine (phi) complexes of rhodium(III), e.g., Rh(phi)2phen3+, as the ground state acceptor. Octahedral complexes containing these ligands have been shown to bind duplex DNA through intercalation into the π stack with binding affinities of ≥106 M1, and extensive biophysical and structural studies have been carried out to characterize their interactions with the DNA duplex [37]. 3.1 Dipyridophenazine Complexes of Ruthenium Excitation of dppz complexes of ruthenium(II) in organic solvents leads to a directional charge transfer from the metal to the dppz moiety, nominally a triplet state, with a lifetime of 270 nsec in ethanol [38]. In aqueous solution, however, the complex does not luminesce because the metal-to-ligand charge transfer state is strongly quenched by proton transfer from solvent to the phenazine nitrogens. In contrast, the luminescence is maintained upon intercalation into double-stranded DNA [39], as the stacked bases protect the dppz ligand from water; the emission intensity is seen to be increased by >103 bound to DNA compared to its absence [40,41]. This DNA ''light switch" effect is illustrated in Fig. 4. Since the excited state involves charge transfer onto the intercalated ligand, irradiation provides a facile path to direct the electron into the DNA π way of stacked bases. Clearly then, dppz complexes of ruthenium represent ideal candidates for studies of DNA-mediated electron transfer, since the excited state is well coupled with the base stacks of DNA and any nonintercalated molecules in water are nonemissive. Given the "light switch" characteristic of these dppz complexes of ruthenium, their luminescence bound to nucleic acids must sensitively depend on and probe how well the nucleic acid serves to protect the phenazine ligand from solvent [40]. The highest luminescence is observed for the complex bound to triple-helical DNA, where likely a stacking arrangment of three bases together serves to shield the ligand. In contrast, little luminescence is observed for the complex in the presence of double-helical RNA; poor intercalation from the major groove
< previous page
page_333
next page >
< previous page
page_334
next page > Page 334
Fig. 4. Ru(phen)2dppz2+ as a molecular ''light switch" for DNA. Shown are the steady-state luminescence spectra of Ru(phen)2dppz2+ in aqueous solution in the presence and absence of DNA. occurs for metallointercalators into double-stranded RNA owing to the deep and narrow groove in the A conformation. In binding to each nucleic acid, a biexponential decay in emission is observed. In the presence of BDNA (nucleotide/Ru = 100), for example, the luminescence of rac-Ru(phen)2dppz2+ decays with τ1 = 130 nsec (80%) and τ2 = 730 nsec (20%). The lifetimes are quite different for the two enantiomers, as the Λ isomer decays with τ = 30 nsec (80%) and τ2 = 160 nsec (20%), while the ∆ isomer decays with τ1 = 150 nsec (80%) and τ2 = 800 nsec (20%) [42]. Since the ∆ isomer accounts for ~ 85% of the luminescence of the racemic compound, it is not surprising that the decay kinetics are nearly identical for racemic and ∆-Ru(phen)2dppz2+. The biexponential decay of excited Ru(phen)2dppz2+ (*Ru) in the presence of DNA has been attributed to two distinct intercalative orientations for the complex [41]. In the "side-on" mode, the complex is considered to be canted toward one strand, maximizing stacking of the phenazine ligand along the long axis of the base pair; in this arrangment one of the phenazine nitrogens is more exposed to solvent. In the "head-on" mode, the dppz ligand is instead perpendicular to the long axis of the base pair and both phenazine nitrogen atoms are well pro-
< previous page
page_334
next page >
< previous page
page_335
next page > Page 335
tected from solvent. Differential quenching studies and the photophysical characteristics of dppz derivatives have supported this assignment [41,43]. Recently one- and 2-dimensional NMR studies of selectively deuterated ∆-Ru(phen)2dppz2+ bound to d(GTCGAC)2 have been carried out [44]. These studies first confirm the preferential intercalation of the dppz ligand into the helix and that the intercalation arises from the major groove of the duplex. The NMR studies, moreover, have been useful in identifying at least two intercalative major groove orientations based on upfield chemical shifts for the dppz ligand and intermolecular nuclear Overhauser effects (NOEs) to the A5H8 resonance in the major groove. It is noteworthy in these studies that the chemical shifts are consistent with at least one symmetrical and one asymmetrical intercalative orientation, a result which fully supports the ''head-on" and "side-on" binding models, assigned using photophysical techniques. 3.2 Bis(phenanthroline)(dipyridophenazine)osmium(II): A Fast Red Emitting DNA Light Switch Os(phen)2dppz2+, like its ruthenium analog, intercalates into DNA, as evidenced by the hypochromism in the dppz ππ* absorption band and in its characteristic light switch effect [45]. This compound has become key to our photophysical studies of DNA-mediated electron transfer, for the osmium complex provides an isostructural donor analog to Ru(phen)2dppz2+ with identical DNA binding properties but distinct electronic properties. As expected, the intercalated racemic osmium complex has much shorter lifetimes than its ruthenium counterpart: 0.8 nsec (49%), 2 nsec (33%), and 11 nsec (18%). The excited state oxidation potential is 0.78 V vs. NHE, ~0.15 V more reducing than *Ru. Importantly in terms of its spectroscopic characteristics, the tail of the 3MLCT absorption band for Os(phen)2dppz2+ extends beyond 700 nm and allows for selective excitation of the donor in the presence of phi complexes of rhodium(III), for which the lowest energy absorption band approaches zero by 600 nm. Moreover, the red-shifted emission of Os(phen)2dppz2+, with its maximum at 740 nm, results in an even smaller spectral overlap between donor emission and acceptor absorp-
< previous page
page_335
next page >
< previous page
page_336
next page > Page 336
tion than for Ru(phen)2dppz2+. These nonoverlapping spectra ensure that any quenching of Os(phen)2dppz2+ by Rh(phi)2X3+ (X = bpy, phen) proceeds via electron transfer and not by a Förster energy transfer mechanism [46]. 3.3 Phenanthrenequinone Diimine Complexes of Rhodium(III) As Acceptors Phi complexes of rhodium(III) have proven to be valuable probes of nucleic acid structure and recognition [37]. These metal complexes generally bind to duplex DNA with K ≥ 106 M1 by intercalation of the phi, and the specific sites which are recognized may be tuned by varying the ancillary, nonintercalating ligands. NMR studies of sitespecifically bound complexes show that they also intercalate into the major groove [4749]. Importantly, the sites recognized by the different phi complexes have been determined through DNA photocleavage experiments [50]. Upon irradiation with ultraviolet light (λ < 350 nm), these phi complexes of rhodium all appear to promote DNA (or RNA) strand scission directly at the site of binding. This photoreaction proceeds via abstraction of the C3' hydrogen atom from the sugar by the activated, intercalated phi ligand, not by base oxidation or by involving an intermediate diffusing species. By varying the ancillary ligands, then, an array of complexes, some sequence-neutral [51], some sequence-selective [52,53], and others with sequence specificity mimicking that of DNA-binding proteins [5456] have been prepared. Rh(phi)2bpy3+, for example, binds in a sequence-neutral fashion and thus may be utilized as a high-resolution photofootprinting agent [51]. Rh(phen)2phi3+ preferentially cleaves 5'-pyr-pyr-pur-3' sequences of DNA with an open major groove by shape selection, since only at somewhat opened sites are clashes of the phenanthroline hydrogens with the flanking base pairs avoided [52]. ∆-α-[Rh(R,R-dimethyltrien)phi]3+ was designed specifically to recognize the sequence 5'-TGCA-3' [55], and indeed two-dimensional NMR studies have now demonstrated the site-specific binding of the metallointercalator to a decamer containing the central 5'-TGCA-3' site [49], with stereospecific hydrogen bonding and methylmethyl contacts precisely as predicted. A remarkably high level of DNA site specificity has been observed with ∆-Rh(diphenylbpy)2phi3+,
< previous page
page_336
next page >
< previous page
page_337
next page > Page 337
which enantiospecifically binds to the duplex sequence 5'-CTCTAGAG-3' as a noncovalent dimer, closely mimicking the recognition properties of a DNA-binding protein [54]. The sequence-neutral Rh(phi)2X3+ (X = bpy, phen) complexes are particularly well suited to act as electron acceptors in studies of electron transfer for several reasons. First, the sequence-neutral binding characteristic of the intercalation ensures that noncovalent binding of the acceptor on the helix is randomly distributed. Moreover, the photocleavage properties of the rhodium complex allow one to assay the distribution of complexes on the strand in the absence and presence of bound donors. Electronically these complexes are also well suited for study in that typically the lowest energy absorption band in a Rh(phi)3+ complex is characterized by a charge transfer (LMCT), permitting the phi ligand to extract electrons from the DNA π way [57]. Thus the rhodium intercalator is poised either to accept electrons from the π way
Fig. 5. Schematic illustration of Os(phen)2dppz2+, the photoinduced donor, and Rh(phi)2phen3+, the intercalating acceptor, poised for DNA-mediated electron transfer through a π stack.
< previous page
page_337
next page >
< previous page
next page >
page_338
Page 338 or to report its position of binding via photocleavage. A metallointercalating donor and acceptor bound for longrange electron transfer through the DNA π stack are illustrated in Fig. 5. 4 Electron Transfer Quenching of DNA-Bound Metallointercalators 4.1 Intercalators vs. Groove-Bound Complexes Long-range electron transfer through DNA occurs only when both donor and acceptor are intercalated into the DNA π stack. This requirement of intercalative stacking for fast quenching has been observed repeatedly in a range of systems (Table 1), but perhaps it is best illustrated by comparing the quenching of DNA-bound *Ru(phen)2dppz2+ by the two acceptors, Rh(phi)2phen3+, an intercalator, and , which binds in the DNA groove electrostatically and through hydrogen bonding (Fig. 6) [43]. The 3+/2+ reduction potentials for Rh(phi)2phen3+ and . NHE, respectively, so that the driving forces for electron transfer from *Ru(II) are nearly identical. is found to quench DNA-bound *Ru(phen)2dppz2+ dynamically on a nanosecond time scale. In the absence of quencher, the MLCT excited state of Ru(phen)2dppz2+ bound to calf thymus DNA decays as a biexponential, with τ1 = 130 nsec (80%) and τ2 = 730 nsec (20%), as measured by nanosecond luminescence spectroscopy. The quenching behavior of
approximately follows Stern-Volmer kinetics. When DNA-
bound Ru(phen)2dppz2+ is titrated with , there is a concomitant and comparable linear increase in both I0/I and τ0/τ, consistent with dynamic quenching on the nanosecond time scale. The dynamic nature of the quenching of Ru(phen)2dppz2+ by
suggests that the reaction involves a diffusing species, and this
mobile species must be because the luminescence polarization [30] and slow dissociation (koff < 70 sec1) [44] for Ru(phen)2dppz2+ bound to DNA indicate that the location of the donor is fixed during its excited state lifetime. It is interesting that the quenching results also support the assignments of two intercalative binding orientations [41], at least qualitatively, as the amount of
< previous page
page_338
quenching correlates with
next page >
< previous page
next page >
page_339
Page 339 TABLE 1 Luminescence Quenching of Photoexcited Donors in the Presence of DNAa Donor
Acceptor
[Ru(phen)2dppz]2+ [Rh(phi)2phen]3+
[Ru(phen)2dppz]2+ [Rh(phi)2phen]3+
[Ru(phen)2dppz]2+ [Ru(NH3)6]3+ [Ru(bpy)3]2+
[Rh(phi)2phen]3+
[Ru(phen)2dppz]2+ methyl viologen Pd porphyrinsc
methyl viologen
Ethidiumd
methyl viologen
Ethidiume
DAP2+
Medium
Nucleotide/ donor
Calf thymus DNA in buffer
28-mer Oligonucleotide in buffer Calf thymus DNA in buffer Calf thymus DNA in buffer Calf thymus DNA in buffer Calf thymus DNA in buffer Calf thymus DNA in buffer Calf thymus DNA in buffer
100
100
100 100 100 40 100 40
Io/I at 5 eq quencherb
Intercalative binding
3.5
Both donor and acceptor
11
1.5 1.3 1.1 1.03 1.9 3.0
Both donor and acceptor Donor only Acceptor only Donor only Donor only Donor only Both donor and acceptor
aQuenching measured at ambient temperature, in buffers where pH is close to neutral. The first four entries of the table are from Ref. 43, the fifth represents an unpublished result; the other entries are from the sources given below. bValues with ethidium reported for 2 equivalents of quencher. cRef. 58. dRef. 23. eRef. 24.
< previous page
page_339
next page >
< previous page
page_340
next page > Page 340
Fig. 6. (top) bound in the DNA minor groove and Rh(phi)2phen3+ (below) intercalated in the DNA major groove. the amount of solvent exposure, with the shorter lifetime being more efficiently quenched than the longer lifetime. Note that this preference is the opposite of what would be expected for a diffusive quencher, since excited Ru(phen)2dppz2+ bound in the head-on mode lives longer than when bound in the side-on mode, leading to more encounters of acceptor with photoexcited donor that is bound in the head-on mode.
< previous page
page_340
next page >
< previous page
page_341
next page > Page 341
A distinctly different quenching behavior is evident with the intercalated acceptor. Upon titration of Ru(phen)2dppz2+ in calf thymus DNA with Rh(phi)2phen3+, the initial luminescence intensity drops markedly and is accompanied by only minimal changes in the nanosecond lifetimes. This behavior manifests itself in a sharply upward curving Stern-Volmer plot of I0/I vs. quencher, while the increase in τ0/τ with quencher concentration is roughly linear and small. These observations are consistent with the majority of the quenching being static, i.e., occurring on a time scale fast relative to the resolution (10 nsec) of the instrument. Repeating the experiment on a 28mer oligonucleotide yields the same results, but the quenching is even more efficient. The fact that the quenching is more efficient on a short piece of DNA strongly indicates that the reaction occurs between fixed species, since the opposite trend was found earlier for the more mobile tris(phenanthroline) complexes [34]. Later, we see that this electron transfer does indeed occur with a subnanosecond quenching constant. Thus, efficient quenching is observed only for cases in which both donor and acceptor are intercalated, as seen for *Ru(phen)2dppz2+ quenched by Rh(phi)2phen3+ and for ethidium quenched by DAP2+ [24]. The same poor quenching seen with
also applies to other electron transfer pairs in which only one of the reactants is
intercalated, e.g., the quenching of by Rh(phi)2phen3+ [43], and the excited state quenching of ethidium [23] or palladium porphyrins [58] by methylviologen. Indeed, when local concentration effects are taken into account, one finds that electron transfer is actually inhibited between an intercalated reagent and a nonintercalator. Clearly then, it is required that both reactants be intercalators in order to exploit the π-stacked array of DNA bases. 4.2 Electron Transfer at a Fixed Distance While the quenching of *Ru(phen)2dppz2+ by Rh(phi)2phen3+ suggested that electron transfer through DNA exhibits a weak distance dependence, a distribution of donor-acceptor distances exists in these experiments with noncovalently bound donors and acceptors. Clearly, the best way to examine the distance dependence of the electron transfer rate is to engineer both donor and acceptor into a covalent assembly [59]. As
< previous page
page_341
next page >
< previous page
page_342
next page > Page 342
Fig. 7. Schematic illustration of a covalently modified DNA duplex assembly to probe long-range electron transfer and the companion experiments using metallated oligonucleotides hybridized to unmodified complements. Irradiation of the ruthenium-modified duplex permits the characterization of intercalation through luminescence, while photolysis of a rhodium-modified duplex yields the position of intercalation through DNA photocleavage. shown in Fig. 7, the metal complexes may be tethered to the 5' termini of complementary strands of a 15-mer DNA duplex. This mixedmetal hybridized duplex yields a self-contained electron transfer complex, boasting a fixed donor-acceptor distance and freedom from bimolecular effects. The individual components of the assembly furthermore allow characterization of both the mode and sites of binding on the DNA duplex. Upon hybridization of the ruthenated oligonucleotide to its unmetallated complement, the luminescence of the ruthenium complex offers information on the binding of the complex to the duplex. Moreover, the DNA photocleavage characteristic of rhodium complexes may be exploited with the rhodium-modified oligonucleotide annealed to radioactively end-labeled but unmetallated complement, so as to report the position of binding of the rhodium intercalator. We first examined the behavior of the components of our covalent electron transfer assembly. As expected, the single-stranded ruthenium-
< previous page
page_342
next page >
< previous page
page_343
next page > Page 343
modified oligonucleotide luminesces only weakly, either in solution or in the presence of noncomplementary singlestranded DNA, consistent with the high accessibility of water to the dppz ligand. When the ruthenium-modified oligonucleotide is annealed to its unmodified complement, however, one observes the strong luminescence that is the signature of the intercalated complex [60]. Importantly, dilution experiments show that this intercalation is intraduplex at the micromolar concentrations typically used for photophysical measurements. Luminescence titrations of the ruthenated duplex with free donor indicate that the 15-mer contains one intercalator, as designed, and that the duplex is accessible to other intercalators [59]. When the ruthenated 15-mer duplex is titrated with free Ru(phen)2dppz2+, one obtains a binding isotherm which saturates at 2 eq of free donor per duplex. The analogous experiment with unmetallated duplex shows saturation at about 3 eq of Ru(phen)2dppz2+, consistent with competitive binding of the complex to the 15-mer duplex and an average binding site size of about 4 base pairs. The close agreement between the two experiments demonstrates that the intercalation of the tethered ruthenium complex is not hindered by the presence of the other intercalators on the duplex. In addition, the luminescence characteristics of the tethered Ru(phen)2dppz2+ are quite similar to those for the free donor bound to the oligonucleotide, indicating that the binding interaction is essentially the same in both cases. Irradiation of phi complexes of rhodium at 313 nm results in DNA strand scission at the site of intercalation, and this photoinduced DNA cleavage may be utilized to report the position of intercalation of the tethered rhodium when in a duplex. The complementary strand is radioactively labeled at its 5' end, annealed to the rhodium-modified strand, and the photocleavage reaction followed by gel electrophoresis. The covalently bound rhodium complex cleaves with high specificity at sites 1 and 2 from the 3' terminus of the 32P-labeled strand, whereas free Rh(phi)2phen3+ cleaves at all positions of an unmetallated duplex. This experiment indicates that the tethered intercalator binds with similar probability one or two base pairs in from the 5' end of the modified strand. The cleavage specificity exhibited by the covalently attached rhodium(III) complex further confirms that intercalation of these covalently attached complexes is largely intramolecular. Since the linker arm is the same for both metal complexes, it is apparent that
< previous page
page_343
next page >
< previous page
page_344
next page > Page 344
both the rhodium(III) and the ruthenium(II) complexes are able to intercalate one and two bases from the 5' sites of attachment. Using 3.4 Å as the distance between stacked heterocycles, hybridization of ruthenated 15-mer hybridized to its rhodium-modified complement therefore creates an assembly, illustrated in Fig. 8, in which the intercalated ligands of the two metal complexes are separated by ≥41 Å through the DNA π stack. Figure 9 shows the steady-state emission spectra of the ruthenated oligonucleotide annealed to its complementary strand and also annealed to its rhodium-modified complement. The difference is striking. Although one might certainly expect some quenching at this separation, complete quenching of the donor luminescence is observed. Timecorrelated single-photon counting (SPC) measurements established a lower limit of 109 sec1 for the quenching constant in the assembly, and picosecond transient absorption measurements on the analogous system with noncovalently bound intercalators indeed show the quenching rate to be 1010 sec1. Electron transfer through the DNA duplex is found to be very efficient over a long molecular distance. The unimolecular nature of this remarkable quenching has been confirmed in several companion experiments. First, the luminescence of the ruthenium-modified duplex is unaffected by the addition of the doubly modified duplex, indicating that there are no adventitious quenchers present in the doubly modified sample. Moreover, the luminescence of ruthenium-modified duplex is only minimally quenched by the presence of 1 eq of rhodiummodified duplex, which shows that the quenching is unimolecular in the doubly modified assembly, consistent with the photocleavage experiments done with rhodium-modified duplex. Both metal complexes must be bound to the same duplex in order for quenching to occur. As noted earlier, titration of the ruthenium-modified duplex with Ru(phen)2dppz2+ demonstrates that tethered and free Ru(phen)2dppz2+ intercalate independently and that the 15 mer can accommodate about three intercalators. In an analogous quenching experiment, addition of one equivalent of Rh(phi)2phen3+ to the ruthenium-modified duplex results in considerable but not complete quenching of the luminescence; some duplexes will bind two rhodium intercalators, thus leaving some ruthenium-modified duplexes without a bound quencher. Therefore, total quenching requires covalent attachment of both donor and acceptor to the same helix.
< previous page
page_344
next page >
< previous page
page_345
next page > Page 345
Fig. 8. A model of the doubly modified 15-mer oligonucleotide duplex. In this model, the metallointercalating donor and acceptor are covalently attached to the 5' ends of complementary strands and stacked within the resultant DNA duplex. The donor-acceptor distance through the DNA π stack is 41 Å. (This figure is adapted with permission from Ref. 59.)
< previous page
page_345
next page >
< previous page
page_346
next page > Page 346
Fig. 9. Emission spectra of the ruthenated 15-mer oligonucleotide hybridized to its complement and of the mixed-metal electron transfer DNA duplex assembly. Although intense emission is observed with the Ru-modified duplex hybridized to the unmodified complement, complete quenching of the emission is apparent with complement modified to contain the Rh intercalator. (This figure is reprinted with permission from Ref. 59.) A model system using the tris(phenanthroline) metal complexes
and
further illustrates the requirement of
does intercalate into DNA, its binding constant is weak (
< previous page
page_347
next page > Page 347
luminescence and its quenching by are identical to that of free ruthenium, indicating that the tethered complex does not intercalate. No significant intraduplex electron transfer quenching occurs in an 8-mer with and covalently bound to the 5'-termini [59]. Hence, it appears that charge transport occurs much more readily through noncovalent π interactions than via a covalent σ bond network, consistent with the requirement of intercalation for fast quenching. These results may be compared to the work of Meade and Kayyem, who see moderate (~106 sec1) quenching rates for a nonintercalating donor and an acceptor covalently bound to the 5' ends of an 8-mer duplex [61]. 4.3 Direct Spectral Evidence for Electron Transfer Assigning a quenching process to electron transfer requires the detection of the products of electron transfer, in this case Ru(III) and Rh(II). Several studies of quenching in covalently linked Ru(II)-Rh(III) polypyridyl complexes have been carried out [6264]. The mechanism for quenching in these systems was attributed to electron transfer because of a lack of spectral overlap between donor emission and acceptor absorption and because of the strong thermodynamic driving force for electron transfer. More recently, in a Ru(II)-Rh(III) dyad [64], evidence for formation of a Ru(III) intermediate with a 100-psec lifetime was provided by transient absorption measurements in which the rhodium chromophore was excited. It is noteworthy that no evidence for reduced acceptor was obtained since Rh(II) is not well characterized due to its instability. In none of the cases described was an electron transfer intermediate observed when the photoexcited Ru(II) was quenched by ground state Rh(III), probably because the rapidity of the back electron transfer relative to the photoinduced forward rate prevents accumulation of the intermediate in appreciable concentrations, as noted by Indelli et al. [64]. Clearly then, while the quenching of Ru(II) by Rh(III) is well documented, the observation of the electron transfer products formed by that quenching has proven to be nontrivial. The first direct evidence for electron transfer through DNA between metallointercalators (Fig. 10) was obtained using donor and acceptor analogs, Ru(DMP)2dppz2+ (DMP = 4,7-dimethylphenanthroline)
< previous page
page_347
next page >
< previous page
page_348
next page > Page 348
Fig. 10. The photoinduced electron transfer cycle for a ruthenium(II) donor and rhodium(III) acceptor. and Rh(phi)2bpy3+ [65]. In the presence of DNA, Ru(DMP)2dppz2+ exhibits both a marked hypochromism in the dppz π-π* band as well as a >100-fold enhancement in luminescence intensity, clearly indicating the compound intercalates into the polymer. Bound to DNA in the absence of Rh(III) (nucleotide/Ru = 100), the decay of the Ru(DMP)2dppz2+ MLCT excited state is biexponential with lifetimes of 32 nsec (27%) and 107 nsec (73%), as measured by SPC. Upon addition of ∆-Rh(phi)2bpy3+, the luminescence intensity decreases markedly, with only a minimal change in the luminescence lifetimes, consistent with quenching on a time scale fast relative to the measurement ( τ(MLCT)] transient grows in upon addition of Rh(III). As expected for a transient intermediate formed by redox quenching, the signal size is found to increase in parallel with the amount of quenching during the titration. The spectral characteristics of the intermediate are also consis-
< previous page
page_348
next page >
< previous page
page_349
next page > Page 349
tent with its assignment as a product of electron transfer. The wavelength dependence for the transient formed by Rh(III) quenching of *Ru(DMP)2dppz2+ is seen to be typical of oxidation of ruthenium polypyridyl complexes. Namely, the transient is positive from 300 to 360 nm and negative from 400 to 460 nm, in close agreement with the Ru(III)-Ru(II) difference spectrum for the DNA-bound donor, generated by electron transfer to the nonintercalating and spectrally silent quencher,
.
It is noteworthy that, in contrast to the fast quenching seen by ∆-Rh(phi)2bpy3+ bound to DNA, the decay of the intermediate is found to be relatively slow, with observable lifetimes ranging from microseconds to milliseconds. We can account for the difference between the photoinduced and recombination rates, given that the donor for the forward reaction is in an excited state and therefore may be better coupled to proposed [66,67] high-energy bridging states in DNA. The recombination reaction, on the other hand, occurs from a ground state donor [Rh(II)] which is low in energy compared to the molecular bridge. Here the differences between the intercalating Rh(III) acceptor and the nonintercalating acceptor are also particularly illustrative. The Rh(III) acceptor, able to couple into the DNA π way, is a much more efficient quencher. However, for the back electron transfer, the high-energy bridge is no longer readily accessible for electron transfer from the ground state Rh(II), and hence the rate of the . Interestingly, in this ground state reaction, the avid recombination reaction is slower even than that of binding of the rhodium intercalator (koff < 103 sec1) may in fact serve to inhibit the back electron transfer by to diffuse along holding the Rh(II) donor at a fixed distance from the Ru(III) acceptor. The ability of the helix would then serve to make its recombination reaction faster than that of the rhodium intercalator. It is important to note that all of these observations are found also to apply for excited state Os(phen)2dppz2+ quenched by ∆-Rh(phi)2bpy3+, demonstrating that *Os quenching also proceeds via electron transfer [68]. In summary, then, these results establish that electron transfer is the primary mechanism for the fast quenching between ruthenium(II) and rhodium(III) metallointercalators bound to DNA, as well as for osmium(II) and rhodium(III) DNA-bound metallointercalators. What makes the long range electron transfer reaction so efficient, then, must be the DNA medium through which the electrons travel.
< previous page
page_349
next page >
< previous page
page_350
next page > Page 350
5 Fast Spectroscopy of DNA-Mediated Electron Transfer How fast does this DNA-mediated electron transfer proceed? The rapidity of these long range electron transfer reactions has required ultrafast spectroscopic techniques to determine rates of reaction. The quenching of a series of dppz complexes by ∆-Rh(phi)2bpy3+ in DNA has therefore been examined using picosecond transient absorption spectroscopy and nanosecond luminescence spectroscopy. These studies allow one to explore and quantitate the different parameters affecting fast electron transfer through the DNA τ stack. 5.1 Loading Independent Rates After excitation, return of *Ru(phen)2dppz2+ to the ground state may be followed by the recovery of the bleach in the MLCT absorption band near 440 nm. When ∆-Ru(phen)2dppz2+ is excited in the absence of quencher, it decays as a biexponential, with lifetimes of 160 nsec (80%) and 850 nsec (20%), as expected. As the quencher is added to the ruthenium complex bound to DNA, however, a fast component appears in the 420-nm ground state recovery. This rate constant is measured to be 8 × 109 sec1 [69]. Importantly, this rate constant is found to be independent of the rhodium concentration. Instead of this component becoming faster during the titration, only its contribution to the overall decay increases. Time-correlated single photon counting measurements on the picosecond timescale reveal that the majority of the quenching occurs on a timescale fast relative to the instrument response, setting a lower limit of 3 × 1010 s1 on the quenching constant. In addition, as no 1010 s1 component is present in the emission dynamics, the kinetics measured by transient absorption spectroscopy must represent the back electron transfer, and this direct observation of the recombination reaction further confirms that electron transfer is the quenching mechanism. The fraction of the signal decaying at 1010 sec1 correlates with the static emission quenching measured by luminescence experiments, which indicates that the picosecond quenching is the predominant de-
< previous page
page_350
next page >
< previous page
page_351
next page > Page 351
cay pathway. By comparison, the rate constant for deactivation of *Ru(phen)2dppz2+ by H2O is only 4 × 109 sec1 and thus this quenching cannot be attributed to displacement of the donor from the DNA by the rhodium intercalator. In addition, neither the rate constant for back electron transfer nor the quenching profile change when D2O is substituted for H2O, further demonstrating that the electron transfer is DNA dependent and is not directly associated with solvent. The insensitivity of the rate constant for back electron transfer to rhodium concentration suggests that the electron transfer must occur over a considerable distance and must display a shallow dependence on distance. Taking the binding of donor and acceptor on the helix to be random, an assumption supported by photocleavage experiments, one may predict the contribution of various donor-acceptor separations as a function of metallointercalator concentration. For example, at a nucleotide/donor ratio of 100, 0.5 eq of acceptor results in 2% of the donor-acceptor separations corresponding to nearest neighbors, while 3 eq of acceptor puts 12% of the Ru-Rh pairs in closest contact. In contrast, the titration shows that 20% and 53% of *Ru(phen)2dppz2+ molecules are quenched at 0.5 and 3 eq of acceptor, respectively. These numbers eliminate nearest-neighbor interactions as the sole source of quenching in a random association of complexes on the helix and require that some of the electron transfer events occur over long distance. Fitting of the steady-state emission quenching profile to a one-dimensional sphere-of-action model [70], one obtains an interaction distance of ~11 base pairs (35 Å) for ∆-Ru(phen)2dppz2+/∆-Rh(phi)2bpy3+ (Fig. 11). The loading-independence in the rate constant for back electron transfer persists upon changes to the intercalated ligand, the ancillary ligands, and the central metal (Table 2). The generality of this phenomenon makes it difficult to attribute the fast reactions to a clustering of donor and acceptor together on the helix because the large cooperativity calculated is inconsistent with both structural modeling and photocleavage studies of donor-acceptor interactions. Moreover, the quenching is associated with the DNA, not with the solvent, as the rate of back electron transfer shows no solvent-isotope effect and differs considerably from the decay of free donor in buffer. Thus these results from donors and acceptors noncovalently associated with DNA suggest a shallow distance dependence for electron transfer through DNA. Indeed, the same conclusion must be reached from the fast quenching
< previous page
page_351
next page >
< previous page
page_352
next page > Page 352
Fig. 11. Stern-Volmer plot of the quenching of luminescence intensity (I) of ∆-Ru(phen)2dppz2+ by ∆-Rh(phi)2phen3+ bound to DNA. These data have been fit to a sphere of action model for quenching along the DNA helix within a critical distance of 35 Å. observed in the covalent assembly [59], where the Ru(II) donor is separated from the Rh(III) acceptor by >40 Å. ∆-Os(phen)2dppz2+ qualitatively follows the same quenching profile as its ruthenium(II) analog, with a somewhat faster recovery constant of 1.1 × 1010 sec1, that is also independent of the loading of acceptor on the helix. As the excited state lifetimes of ∆-Os(phen)2dppz2+ bound to DNA are much shorter (1.5 and 9.2 nsec) than for ∆*Ru(phen)2dppz2+, the unchanging rate constant cannot be an anomaly introduced by the large difference between the intrinsic decay rates and the rates of back electron transfer. Perhaps the most important advantage of using *Os(phen)2dppz2+ as donor is that the >100-nm red shift in the emission maximum also makes it very unlikely that quenching could proceed via an energy transfer mechanism. In addition, the intrinsically shorter lifetimes of *Os(phen)2dppz2+ aided in the direct detection of the long-lived electron transfer intermediate in the same system, and this observation strengthens our assertion that the fast quenching occurs by electron transfer.
< previous page
page_352
next page >
< previous page
page_353
next page > Page 353
TABLE 2 Rate Constants for Back Electron Transfera ∆G (eV)b
ket (sec1)c
∆-Ru(phen)2dppz2+
1.66
8.9 × 109
Ru(DMP)2dppz2+d
1.59
1.1 × 1010
∆-Os(phen)2dppz2+
1.21
1.1 × 1010
Λ-Ru(phen)2dppz2+
1.66
4.5 × 109
Donor
aConditions: nucleotide/donor = 100, calf thymus DNA, quencher = ∆-Rh(phi)2bpy3+, in 5 mM Tris, 50 mM NaCl, pH 8.5. The data are from Ref. 69. bExcited state reduction potentials were determined from the excited state energy and the ground state reduction potential (vs. NHE). cRate constants reflect the kinetics of ground state recovery measured in picosecond transient absorption measurements. dDMP = 4,7-dimethyl-1,10-phenanthroline. 5.2 Sensitivity in Electron Transfer to π Stacking Just as we have observed a requirement of intercalation for fast DNA-mediated electron transfer, we see also that the quenching appears to be most sensitive to the degree of stacking of the intercalated complex into the DNA helix. Perhaps the best example of the relationship between electron transfer and stacking of the metallointercalators comes from comparison of Ru(phen)2dppz2+ enantiomers. The excited state life-times of intercalated ∆-Ru(phen)2dppz2+ are about five times longer than those for Λ-Ru(phen)2dppz2+ [42], reflecting the complementary binding and thus better stacking of the ∆ isomer with the right-handed helix [25]. Moreover, we find that the ∆ isomer is quenched more efficiently by rhodium than is the Λ isomer, and the rate constant for back electron transfer is twice as fast [69]. This difference is remarkable for two reactions with the same driving force (to a first approximation), and
< previous page
page_353
next page >
< previous page
page_354
next page > Page 354
the better quenching of the better stacked ∆ enantiomer reflects the sensitivity of the electron transfer reaction to stacking. Variations in the driving force for photoinduced electron transfer through DNA appear primarily to affect the rate constant for back electron transfer and not the fraction of quenched donor, supporting the assertion that the stacking interaction is the most important factor in controlling electron transfer through DNA. For example, ∆Ru(phen)2dppz2+ and ∆-Os(phen)2dppz2+ should bind identically to DNA and have equivalent titration profiles with Rh(phi)2bpy3+ as quencher. However, the rate constant for back electron transfer is somewhat faster (1.1 × 1010 sec1) for ∆-Os(phen)2dppz2+ than for its isostructural ruthenium(II) analog (8 × 109 sec1). A particularly interesting story arises from the comparison of the quenching efficiencies of Rh(phi)2phen3+ and Rh(phen)2phi3+ as acceptors [71]. Both of these compounds bind to DNA with high affinity (Kb > 106 M1), yet while the former is an efficient quencher of *Ru(phen)2dppz2+ luminescence, Rh(phen)2phi3+ does not show measurable quenching in the presence of DNA. However, both rhodium complexes quench equally well in sodium dodecyl sulfate micelles, suggesting that the lack of quenching by Rh(phen)2phi3+ in DNA cannot be explained on thermodynamic grounds alone. Rather, it appears that Rh(phen)2phi3+ does not quench because it is poorly coupled in the π stack. Earlier photo-cleavage studies support this assertion, as Rh(phen)2phi3+ has been shown to preferentially bind to sites with an open major groove, because of steric clashes between the ancillary phen ligands and the DNA backbone [52,72]. The faster dissociation constant for Rh(phen)2phi3+ from DNA, obtained by NMR, compared to Rh(phi)2phen3+ also reflects the poorer stacking of Rh(phen)2phi3+. Thus, it appears that the quenching of Ru(phen)2dppz2+ by Rh(phen)2phi3+ is inhibited by a poor stacking interaction of the acceptor with the DNA. Clearly then, while driving force has a small effect on the rate constant for electron transfer, the reaction is governed primarily by the interaction of the π-stacked DNA bases with the metallointercalators. 5.3 Variations in DNA Sequence Among the fundamental questions in electron transfer studies is how the composition of the intervening medium between donor and acceptor
< previous page
page_354
next page >
< previous page
next page >
page_355
Page 355 influences the reaction. Using the alternating polymers poly d(GC) and poly d(AT), the quenching of ∆*Ru(phen)2dppz2+ by ∆-Rh(phi)2bpy3+ has also been examined as a function of DNA sequence [69]. The quenching in poly d(AT) is found to be similar to that in mixed-sequence calf thymus DNA, but the back electron transfer is slightly slower (7 × 109 sec1). In contrast, there is much less quenching in poly d(GC) and the rate constant for back electron transfer is 2 × 108 sec1, about 40 times slower than in poly d(AT) (Table 3). This strong sequence dependence may indicate the importance of the intervening medium in facilitating long-range electron transfer through DNA. Since both polymers have B-form conformations [13], the sequence-dependent quenching cannot be attributed to gross structural changes in the DNA. Indeed, both donor and acceptor bind with high affinity to both poly d(AT) and poly d(GC), as shown by photophysical studies of Ru(phen)2dppz2+ [40,42] and DNA photocleavage studies with Rh(phi)2bpy3+ [51]. However, ∆-*Ru(phen)2dppz2+ displays shorter excited state lifetimes in the dGC polymer than in the dAT polymer, indicating that the stacking of the intercalated ligand is different in the two sequences. In addition, the sequence-dependent electron transfer provides another argument against cooperative binding of donor and TABLE 3 Sequence-Dependent Electron Transfera Bound to DNA DNA poly d(AT) poly d(GC)
Quenched by Rh
τ (ns)b
%
τ (ps)c
%
120
75
135
70
720
25
37
10
4700
47
280
90
aNucleotide/donor = 100; quenching at 3 eq of ∆-Rh(phi)2bpy3+ in 5 mM Tris, 50 mM NaCl, pH 8.5. All the data are from Ref. 69. bMeasured by nanosecond emission experiments. cThese lifetimes reflect the kinetics of back electron transfer determined in transient absorption measurements.
< previous page
page_355
next page >
< previous page
page_356
next page > Page 356
acceptor, since such an association of the cationic metal complexes should be equally likely on either polyanion. 6 Comparisons of Metallointercalating Systems with Others It is instructive at this point to compare our results with other studies of electron transfer through DNA. The most relevant studies are those using intercalators as both donor and acceptor, and we focus on these first. Baguley et al. found that ethidium fluorescence is quenched by amsacrine and several other 9-anilinoacridine derivatives [21,22]. A qualitative correlation of the quenching efficiency with electron-donating groups on the quencher and a lack of spectral overlap led them to propose that the ethidium fluorescence is deactivated via a reductive quenching mechanism. The quenching occurred on the nanosecond time scale, as measured by SPC, but the lifetime of the quenched ethidium was not resolvable. The intensity quenching did not become significant until higher loadings of the photoexcited acceptor and ground state donor on the DNA, a result which suggests that the electron transfer is not as efficient in this system as with the metallointercalators *M(phen)2dppz2+/Rh(phi)2X3+ (M = Ru, Os; X = bpy, phen). Interestingly, these workers also observed sequence-dependent quenching, which was more efficient in poly d(AT) than in poly d(GC). A more detailed study of electron transfer through DNA was performed by Brun and Harriman [24]. In this work, ethidium and acridine orange were utilized as the photoexcited electron donor. DAP2+, a methyl viologen pyrene derivative which intercalates into DNA under low-salt conditions, was the electron acceptor. This study demonstrated the formation of reduced acceptor and thus provided the first evidence for an intermediate formed by electron transfer through DNA. As seen with the reductive quenching of ethidium by amsacrine, strong quenching was seen only at high loadings. The quenching of ethidium by DAP2+ led to a triexponential fluorescence decay, in which the lifetimes were 0.7, 8, and 20 nsec, respectively. Applying the nearest-neighbor exclusion rule and assuming a nonrandom distribution of donors and acceptors on the DNA, these lifetimes were assigned to donoracceptor
< previous page
page_356
next page >
< previous page
page_357
next page > Page 357
separations of 3, 4, and 5 base pairs. These results indicated a much stronger distance dependence than what was observed with our system of metallointercalators, whether covalently or noncovalently bound to DNA. This value might result from poorer coupling of the ethidium singlet into the DNA π stack or, alternatively, it might be a result of their assumption of a nonrandom distribution. While the authors assign separations of 35 base pairs between reactants based on a nonrandom distribution, there is no direct evidence that these are in fact the actual distances in contrast to the metallointercalator system, where a control is provided by a covalent assembly. Certainly, this work would benefit from studies where the reactants are clearly separated by a fixed distance, although this experiment could be difficult to carry out if indeed the distance dependence is as strong as reported, because the short lengths of the oligonucleotides necessary to achieve quenching may preclude duplex formation. Meade and Kayyem designed a system with nonintercalating metal complexes which promises to yield valuable information on the distance dependence of electron transfer through DNA [61]. In this system, both donor and accept or are tethered to the 2' position of a deoxyribose at the 5' termini. The high exposure of the nonintercalated reactants to solvent makes this system well suited for flash-quench experiments, which are useful in studying ground-state electron transfer. This system with covalent modifications is complementary to ours in that it should reveal how σ pathways mediate coupling into the DNA π stack, whereas electron transfer between intercalators probes the π stack directly. In an 8-mer duplex containing Ru(bpy)2(Im)3+ (Im = imidazole) as acceptor and Ru(NH3)4(py)2+ (py = pyridine) as donor, Meade and Kayyem reported an electron transfer rate constant of 1.6 × 106 sec1. The authors conservatively compare their electron transfer rate for a 21-Å through-space separation to a value of 2.5 × 106 sec1 for a 20-Å donor-acceptor distance in ruthenated cytochrome c [9], one of the most efficient protein electron transfer pathways reported to date. However, such a fast electron transfer almost certainly could not occur through-space, and thus the electron transfer is in fact much more efficient in the 8-mer duplex than in the cytochrome. Although this rate is considerably slower than that between metallointercalators, the difference is perhaps expected because the electron must pass through several σ bonds to access the π-stacked bases en route to the acceptor. In
< previous page
page_357
next page >
< previous page
page_358
next page > Page 358
addition, the high exposure of these nonintercalating metal complexes to solvent would likely result in a higher value of outer-sphere reorganization energy, thus lowering the electron transfer rate relative to the intercalated species. 7 Theory Given the intriguing results obtained thus far, it becomes important to establish a theoretical framework through which to consider experiments on DNA-mediated electron transfer. A number of theoretical studies have been carried out to assess the ability of DNA to facilitate charge transport. In a model involving discrete reduced bases in the transport, Dee and Baur predicted that electrons could hop from base to base on a very fast time scale, approaching femtoseconds [73]. Using the pathway model applied successfully to proteins, which assumes an exponential distance dependence, Risser et al. predicted that electronic coupling could actually increase with distance at short separations for a system with covalently bonded but nonintercalating reactants [74]. Interestingly, the coupling was predicted to depend both on the DNA sequence and whether or not the reactants are on the same strand. Recently, Beratan and coworkers extended this analysis to assess the ability of the DNA bridge to facilitate electron transfer between pairs of intercalators or covalently attached nonintercalators [67]. In this two-state nonadiabatic model for weakly coupled systems, they find that all efficient coupling of donor and acceptor proceeds through the πstacked bases rather than through the DNA backbone. This pathway approach is consistent with the strong distance dependence of the quenching of ethidium but can produce a weaker distance dependence in the limit where the tunneling energies are within thermal access of the bridge; in this limit, the decay in electron transfer rate with distance is calculated to be smaller for poly d(A)·poly d(T) than for poly d(G)·poly d(C), which is in agreement with experimental quenching results. Using multilevel Redfield treatment of bridge-mediated electron transfer, Felts et al. formulated a model which explains how distance-independent electron transfer may occur [66]. The essential charac-
< previous page
page_358
next page >
< previous page
page_359
next page > Page 359
teristics of the treatment involve a tight-binding system comprising a donor, intervening bridge elements, an acceptor, contact with a thermal bath, and a high-frequency oscillator coupling vibrations to the electron transfer event. The model agrees with standard formulations such as the Golden Rule or superexchange. However, in the limit where the donor orbital energies approach the energy of the bridge orbitals, the electron can thermally access the bridge, becoming rapidly dephased by bath fluctuations into the spatially delocalized states of the bridge en route to the acceptor. These dephasing processes, caused by energy fluctuations in individual sites, introduce an adiabatic channel for which electron transfer can be distance-independent, in contrast to the strong distance dependence expected for the nonadiabatic channel. There are three requirements for such a distance independence: (1) a ''conduction" band of wave-like states in the bridge, feasible so long as a minimum level of site couplings exists in the bridge, (2) system-bath fluctuations that disrupt the coherence of the bridge; and (3) a small enough energy gap between donor and bridge that an electron can remain in the bridge throughout the transfer process. This treatment may explain the sequence dependence of the quenching between metallointercalators. More efficient quenching is seen for poly d(AT) than for poly d(GC). In light of the Redfield treatment, this result might be expected because poly d(AT) exhibits an uncommonly large conformational flexibility for B-DNA [75], perhaps providing the strong energy fluctuations necessary for dephasing and efficient quenching. In contrast, the rigidity of poly d(GC) should hinder the adiabatic channel for electron transfer. This theory may also account for the onset of weak distance dependence in electron transfer rates measured for metal complexes attached to polyproline peptides [5]. Indeed, for separations of more than three residues (>18 Å), the distance dependence becomes quite shallow, allowing electron transfer over 40 Å on a submillisecond time scale. This mechanism might also be functioning in the microsecond time scale electron transfer from Ru(NH3)4(py)2+ to Ru(bpy)2(Im)3+ in an 8-mer duplex, for which the through-bond separation is greater than 30 Å [61]. Since energy fluctuations help to open the adiabatic channel for electron transfer in the Redfield treatment, one predicts that the distance independence of the electron transfer rate will be diminished at temperatures low enough to eliminate thermal access of the bridge from
< previous page
page_359
next page >
< previous page
page_360
next page > Page 360
the donor orbitals. Thus, quenching experiments performed as a function of temperature would aid in determining whether or not this model is in fact valid for electron transfer through DNA. In contrast, for nonadiabatic electron transfer, one expects tunneling and thus a temperature-independent rate to be observed at low temperatures. Experiments are underway to test these models. 8 Future Studies Experimental studies carried out thus far indicate that the DNA double helix represents a remarkable medium for long-range charge transport. The DNA helix does indeed represent a molecular π stack, facilitating charge transport over long molecular distances and in a manner that depends most sensitively on the π stacking of the aromatic heterocyclic base pairs. The analog, then, for these studies is perhaps not long-range electron transfer through proteins, utilizing isolated aromatic residues, but electron mobility in solid state π-stacked materials, such as the stacked phthalocyanines. We need next, then, to describe the parameters which govern and distinguish electron transport in a molecular π stack. With solid state materials, determining the distance dependence of electron transfer is difficult to establish, but in a molecular π stack the careful variations in donor-acceptor distance becomes feasible, albeit synthetically challenging. Using the DNA helix as a medium for electron transport furthermore offers opportunities to develop a range of novel, site-selective biosensors. We need to determine variations in electron transfer efficiency as a function of sequence and perturbations in π stacking. Figure 12 illustrates the remarkable variations in stacking available in comparing the different duplex conformations of A-, B-, and Z-DNA. Can we begin to harness electron transfer rates as a probe of DNA structure and dynamics? Finally, the observation that electron transport through the DNA helix is facile begs the question of whether such fast electron transport serves a role in the biological expression of DNA. Indeed, this high electron mobility could facilitate long range damage to DNA. Here, too,
< previous page
page_360
next page >
< previous page
page_361
next page > Page 361
Fig. 12. A view down the DNA π stacks for B- (left), A- (center), and Z- (right) DNA duplexes. it is time for chemists to elucidate the molecular characteristics of the DNA double helix and to begin to explore whether the DNA π stack may also function physiologically as a wire. Acknowledgments We are especially grateful to our coworkers and collaborators, as named in the individual references, for their efforts. In addition, we thank the NIH (GM49216) for their financial support of this research and the American Cancer Society for a postdoctoral fellowship to E.D.A.S. Abbreviations
bpy
2,2'-bipyridyl
DAP2+
N,N'-dimethyl-2,7-diazapyrenium
DMP
4,7-dimethyl-1,10-phenanthroline
dppz
dipyridophenazine
Im
imidazole
LMCT
ligand-to-metal charge transfer
MLCT
metal-to-ligand charge transfer
NHE
normal hydrogen electrode
< previous page
page_361
next page >
< previous page
page_362
next page > Page 362
NMR
nuclear magnetic resonance
NOE
nuclear Overhauser effect
phen
1,10-phenanthroline
phi
9,10-phenanthrenequinonediimine
pur
purine
py
pyridine
pyr
pyrimidine
SPC
single-photon counting
References 1. R. A. Marcus and N. Sutin, Biochim. Biophys. Acta, 811, 265 (1985). 2. C. C. Moser, J. M. Keske, K. Warncke, R. S. Farid, and P. L. Dutton, Nature, 355, 796 (1992). 3. J. R. Winkler and H. B. Gray, Chem. Rev., 92, 369 (1992). 4. D. N. Beratan, J. N. Onuchic, J. N. Betts, B. E. Bowler, and H. B. Gray, J. Am. Chem. Soc, 112, 7915 (1990). 5. S. S. Isied, M. Y. Ogawa, and J. F. Wishart, Chem. Rev., 92, 381 (1992). 6. G. McLendon, Acc. Chem, Res,, 21, 160 (1988). 7. H. Heitele, M. E. Michel-Beyerle, and P. Finckh, Chem. Phys. Lett., 134, 273 (1987). 8. J. Seth, V. Palaniappan, T. E. Johnson, S. Prathapan, J. S. Lindsey, and D. F. Bocian, J. Am. Chem. Soc., 116, 10578 (1994). 9. D. R. Casimiro, J. H. Richards, J. R. Winkler, and H. B. Gray, J. Phys. Chem., 97, 13073 (1993). 10. A. M. Everest, S. A. Wallin, E. D. A. Stemp, J. M. Nocek, A. G. Mauk, and B. M. Hoffman, J. Am. Chem. Soc., 113, 4337 (1991). 11. P. G. Schouten, J. M. Warman, M. P. de-Haas, M. A. Fox, and H.-L. Pan, Nature, 353, 736 (1991). 12. T. J. Marks, Science, 227, 881 (1985). 13. W. Saenger in Principles of Nucleic Acid Structure (C. R. Cantor, ed.), Springer-Verlag, New York, 1984. 14. J. Sambrook, E. F. Fritsch, and T. Maniatis in Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, New York, 1989.
< previous page
page_362
next page >
< previous page
page_363
next page > Page 363
15. A. Szent-Györgyi, Proc. Natl. Acad. Sci., U.S.A., 46, 1444 (1960). 16. R. S. Snart, Biopolymers, 6, 293 (1968). 17. P. M. Cullis, J. D. McClymont, and M. C. R. Symons, J. Chem. Soc. Faraday Trans., 86, 591 (1990). 18. C. Houee-Levin, M. Gardes-Albert, A. Rouscilles, C. Ferradini, and B. Hickel, Biochemistry, 30, 8216 (1991). 19. R. F. Anderson, K. B. Patel, and W. R. Wilson, J. Chem. Soc. Faraday Trans., 87, 3739 (1991). 20. A detailed general review may be found in Chapter 13 by T. J. Meade in Volume 32 of this series. 21. B. C. Baguley and M. LeBret, Biochemistry, 23, 937 (1984). 22. L. M. Davis, J. D. Harvey, and B. C. Baguley, Chem.-Biol. Interact., 62, 45 (1987). 23. P. Fromherz and B. Rieger, J. Am. Chem. Soc., 108, 5361 (1986). 24. A. M. Brun and A. J. Harriman, J. Am. Chem. Soc., 114, 3656 (1992). 25. J. K. Barton, Science, 233, 727 (1986). 26. J. K. Barton, A. T. Danishefsky, and J. M Goldberg, J. Am. Chem. Soc., 106, 2172 (1984). 27. C. V. Kumar, J. K. Barton, and N. J. Turro, J. Am. Chem. Soc., 107, 5518 (1985). 28. J. K. Barton, J. M. Goldberg, C. V. Kumar, and N. J. Turro, J. Am. Chem. Soc., 108, 2081 (1986). 29. A. M. Pyle, J. P. Rehmann, R. Meshoyrer, C. V. Kumar, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 111, 3051 (1989). 30. A. E. Friedman, C. V. Kumar, N. J. Turro, and J. K. Barton, Nucl. Acids Res., 19, 2595 (1991). 31. J. P. Rehman and J. K. Barton, Biochemistry, 29, 1701 (1990). 32. J. P. Rehman and J. K. Barton, Biochemistry, 29, 1710 (1990). 33. C. Creutz, A. D. Keller, N. Sutin, and A. P. Zipp, J. Am. Chem. Soc., 104, 3618 (1982). 34. J. K. Barton, C. V. Kumar, and N. J. Turro, J. Am. Chem. Soc., 108, 6391 (1986). 35. M. D. Purugganan, C. V. Kumar, N. J. Turro, and J. K. Barton, Science, 241, 1645 (1988).
< previous page
page_363
next page >
< previous page
page_364
next page > Page 364
36. G. Orellana, A. Kirsch-De Mesmaeker, J. K. Barton, and N. J. Turro, Photochem. Photobiol., 54, 499 (1991). 37. C. M. Dupureur and J. K. Barton in, Comprehensive Supramolecular Chemistry, Vol. 5 (J.-M. Lehn, ed), Pergamon Press, Oxford, 1995. 38. E. Amouyal, A. Homsi, J.-C. Chambron, and J.-P. Sauvage, J. Chem. Soc. Dalton Trans., 1841 (1990). 39. A. E. Friedman, J.-C. Chambron, J.-P. Sauvage, N. J. Turro, and J. K. Barton, J. Am. Chem. Soc., 112, 4960 (1990). 40. Y. Jenkins, A. E. Friedman, N. J. Turro, and J. K. Barton, Biochemistry, 31, 10809 (1992). 41. R. M. Hartshorn, and J. K. Barton, J. Am. Chem. Soc., 114, 5919 (1992). 42. C. Hiort, P. Lincoln, and B. Nordén, J. Am. Chem. Soc., 115, 3448 (1993). 43. C. J. Murphy, M. R. Arkin, N. D. Ghatlia, S. Bossman, N. J. Turro, and J. K. Barton, Proc. Natl. Acad. Sci. USA, 91, 5315 (1994). 44. C. M. Dupureur and J. K. Barton, J. Am. Chem. Soc., 116, 10286 (1994). 45. R. H. Holmlin and J. K. Barton, Inorg. Chem., 34, 7 (1995). 46. G. J. Kavarnos and N. J. Turro, Chem. Rev., 86, 401 (1986). 47. S. S. David and J. K. Barton, J. Am. Chem. Soc., 115, 2984 (1993). 48. J. G. Collins, T. P. Shields, and J. K. Barton, J. Am. Chem. Soc., 116, 9840 (1994). 49. B. P. Hudson, C. M. Dupureur, and J. K. Barton, J. Am. Chem. Soc., 117, 9379 (1995). 50. A. Sitlani, E. C. Long, A. M. Pyle, and J. K. Barton, J. Am. Chem. Soc., 114, 2303 (1992). 51. K. Uchida, A. M. Pyle, T. Morii, and J. K. Barton, Nucl. Acids. Res., 17, 10259 (1989). 52. A. M. Pyle, E. C. Long, and J. K. Barton, J. Am. Chem. Soc., 111, 4520 (1989). 53. A. H. Krotz, L. Y. Kuo, T. P. Shields, and J. K. Barton, J. Am. Chem. Soc., 115, 3877 (1993). 54. A. Sitlani, C. M. Dupureur, and J. K. Barton, J. Am. Chem. Soc., 115, 12589 (1993).
< previous page
page_364
next page >
< previous page
page_365
next page > Page 365
55. A. H. Krotz, B. P. Hudson, and J. K. Barton, J. Am. Chem. Soc., 115, 12577 (1993). 56. R. H. Terbrueggen and J. K. Barton, Biochemistry, 34, 8227 (1995). 57. A. H. Krotz, L. Y. Kuo, and J. K. Barton, Inorg. Chem., 32, 5963 (1993). 58. A. M. Brun and A. Harriman, J. Am. Chem. Soc., 116, 10383 (1994). 59. C. J. Murphy, M. R. Arkin, Y. J. Jenkins, N. D. Ghatlia, S. H. Bossmann, N. J. Turro, and J. K. Barton, Science, 262, 1025 (1993). 60. Y. Jenkins and J. K. Barton, J. Am. Chem. Soc., 114, 8736 (1992). 61. T. J. Meade and J. F. Kayyem, Angew. Chem. Int. Ed., 34(3), 352 (1995). 62. K. Kalyanasundaram, M. Graetzel, and M. K. Nazeeruddin, J. Phys. Chem., 96, 5865 (1992). 63. K. Nozaki, T. Ohno, and M. Haga, J. Phys. Chem., 96, 10880 (1992). 64. M. T. Indelli, C. A. Bignozzi, A. Harriman, J. R. Schoonover, and F. Scandola, J. Am. Chem. Soc., 116, 3768 (1994). 65. E. D. A. Stemp, M. R. Arkin, and J. K. Barton, J. Am. Chem. Soc., 117, 2375 (1995). 66. A. K. Felts, W. T. Pollard, and R. A. Friesner, J. Phys. Chem., 99, 2929 (1995). 67. S. Priyadarshy, S. M. Risser, and D. N. Beratan, unpublished results. 68. R. E. Holmlin, E. D. A. Stemp, and J. K. Barton, unpublished results. 69. M. R. Arkin, E. D. A. Stemp, R. E. Holmlin, J. K. Barton, A. Hoermann, E. Olson, and P. F. Barbara, unpublished results. 70. F. Perrin, Compt. Rend., 178, 1978 (1924). 71. M. R. Arkin, E. D. A. Stemp, and J. K. Barton, submitted for publication. 72. D. Campisi, T. Morii, and J. K. Barton, Biochemistry, 33, 4130 (1994). 73. D. Dee and M. E. Baur, J. Chem. Phys., 60(2), 541 (1974). 74. S. M. Risser, D. N. Beratan, and T. J. Meade, J. Am. Chem. Soc., 115, 2508 (1993). 75. M. Vorlickova and J. Kypr, J. Biomol. Struct. Dynam., 3, 67 (1985).
< previous page
page_365
next page >
< previous page
page_367
next page > Page 367
12 Porphyrin and Metalloporphyrin Interactions with Nucleic Acids Robert F. Pasternack1 and Esther J. Gibbs2 1Department of Chemistry, Swarthmore College, Swarthmore, PA 19081, USA 2Department of Chemistry, Goucher College, Towson, MD 21204, USA
368
1. Background
368
1.1. Introduction
1.2. Structures and Solution Properties of Porphyrins and Metalloporphyrins
371
2. Porphyrin Interactions with Duplex DNA
371
2.1. Early Efforts to Establish Binding Modes
378
2.2. Specifics of Porphyrin-DNA Interactions
379
2.3. Does Axial Ligation Play a Role?
380
2.4. Kinetics
3. Binding of Porphyrins and Metalloporphyrins to Other Nucleic Acid Structures
3.1. Interactions with Duplex and Transfer RNAs
382
382
383
3.2. Interactions with Branched DNA
385
3.3. Binding Studies with DNA Ψ Aggregates
< previous page
370
page_367
next page >
< previous page
page_368
next page > Page 368 386
4. Porphyrin Assemblies on Nucleic Acids
386
4.1. Assemblies on Duplex DNA
4.2. A Brief Introduction to Resonance Light Scattering
387
388
4.3. Porphyrin Assemblies Revisited
390
5. New Directions and Closing Remarks
390
Acknowledgment
391
Abbreviations
391
References
1 Background 1.1 Introduction Studies of the interactions of small molecules (referred to generically in this context as ''drugs") with nucleic acids was a well-established research area prior to the discovery in 1979 that porphyrins are capable of intercalating into DNA [1]. Yet this first report led to a surge of activity in the field which continues to this day. What then are the unique, appealing features of porphyrinic species, especially with respect to their interactions with nucleic acids? It is almost impossible when addressing the question of relevance to resist alluding to the clinical potential of this class of compounds [25]. In addition to their chemical and photochemical versatility, porphyrins have been known for many years to accumulate spontaneously in malignancies. Photodynamic therapy, for example, used for the treatment of several types of cancer takes advantage of both porphyrin accumulation and photosensitization properties. In addition, fluorescence by many porphyrin derivatives makes them useful for diagnosis of even incipient cancer cells and porphyrins have recently been reported as having some potential for the treatment of AIDS [6,7]. However, what seemed especially intriguing to us is that meso-substituted porphyrins of the type shown in Fig. 1 must, by virtue of their structure, challenge DNA and therefore, provide information about DNA in ways which other smaller, more compact drugs do not. When Henry Sobel first showed a space-filling model of an intercalated porphyrin at a meeting at the University of Rochester in the early
< previous page
page_368
next page >
< previous page
page_369
next page > Page 369
Fig. 1. Structures of meso-substituted porphyrins used for these studies. The commonly used symbol for each porphyrin is shown next to the appropriate R-group(s). 1980s, it became clear that the question to be addressed was not whether a porphyrin could intercalate into DNA but rather how, i.e., by what mechanism the process occurs. The flat, rigid porphine core is of the proper dimensions to be held between adjacent base pairs in an intercalated complex; the positively charged peripheral substituents are restricted to lie outside the double helix, several in proximity to negatively charged phosphate groups. However, it should be noted that these peripheral substituents are nearly perpendicular to the porphine plane and that the barrier to rotation to coplanarity is considerable [8,9]. Is some remarkable distortion of either the porphyrin or the DNA required to effect intercalation? It was the question of interaction kinetics which first drew us to porphyrin/nucleic acid research and, as so often happens, this initial foray quickly led to other, more engaging issues. The remarkable repertoire of reactivity of porphyrins and metalloporphyrins, the ease with
< previous page
page_369
next page >
< previous page
page_370
next page > Page 370
which their properties may be ''tuned" via peripheral substitution or choice of metal for insertion, and the richness and sensitivity of the spectroscopic features of these species make them particularly suitable for such investigations. These studies have been aided immeasurably by the extensive research activity on water-soluble porphyrin and metalloporphyrin derivatives. A number of excellent reviews of structure and chemical behavior of such derivatives have appeared [1012]. However, it may be useful for those readers who have not worked with these substances to begin with a brief consideration of some of the more relevant solution properties of the cationic porphyrins shown in Fig. 1 and of their metal complexes. 1.2 Structures and Solution Properties of Porphyrins and Metalloporphyrins The form of the porphyrins shown in Fig. 1 is referred to as the "free base." These species can be induced to add more protons at the pyrrole nitrogen positions to form mono-and diacid species. Generally, for meso-substituted porphyrins, the monoacid form is stable over a very limited pH range. In fact, it had been widely believed, until studies by Hambright and coworkers [13], that the two protons were added in a concerted fashion. Interactions of porphyrins with nucleic acids have for the most part been limited to free base or metallo forms, although several recent reports propose the binding to DNA of a diacid porphyrin derivative [1416]. Many chemical and physical properties of porphyrins can be correlated with their basicity, as, for example, their tendency to self-aggregate and the ease with which certain metal derivatives add axial ligands [17,18]. Therefore, this is a parameter of some interest. Of the tetracationic porphyrins shown in Fig. 1, the order of their basicities is H2T2 < H2T4 < H2TAP. They are all characterized by a very intense absorption in the blue-violet region of the spectrum (the Soret band) having a molar absorptivity of about 15 × 105 M-1 cm-1 and, consequently, micromolar concentration levels are most often used for spectral measurements. None of these tetracationic free base porphyrins dimerize or otherwise aggregate at these concentrations. Some controversy had existed as to the state of aggregation of H2T4 but that issue seems to
< previous page
page_370
next page >
< previous page
page_371
next page > Page 371
have been resolved in the conclusion that this porphyrin is monomeric in aqueous solution [17,1922]. In marked contrast is the solution behavior of the dicationic porphyrins c- and t-H2Pagg (see Fig. 1), which tend to form extended self-aggregates even at the concentration levels considered here and especially in the presence of added electrolyte [23,24]. The aggregation produces a substantially red-shifted Soret band (of Lorentzian line shape) and a significant lowering of the quantum yield for fluorescence [25]. A variety of metal ions can be inserted into these porphyrin derivatives with the replacement of the central protons. Of particular interest is the preferred coordination number of the metal center as illustrated in Fig. 2 for the ''T4" series of porphyrins [18,2629]. Transition metal ions having a "d8" electron configuration commonly form square coplanar complexes and, indeed, such is the case here. Although palladium(II), platinum(II), and gold(III) porphyrin derivatives can be induced to add axial ligands if a high enough concentration of an appropriate Lewis base is present, under most conditions they remain four-coordinate. Similarly, Cu(II)T4 shows little tendency to add axial ligands and Ni(II)T4 exists in water as a roughly 1:1 mixture of (diamagnetic) four-coordinate and (paramagnetic) six-coordinate forms, which are readily and rapidly interconverted [28,30]. ZnT4 is a five-coordinate species in aqueous solution with the zinc displaced from the porphyrin plane. The remaining metal derivatives of interest hereVO(II)T4, Co(III)T4, Mn(III)T4, and Fe(III)T4also tend to add axial ligands to form five- or six-coordinate species. Aggregation tendencies of metal derivatives can be at some variance with the free base porphyrins from which they are derived. CuTAP, NiTAP, and AgTAP, for example, show greater tendency to aggregate than does H2TAP [31] and whereas t-CuPagg behaves like the metal-free derivative and aggregates extensively, t-AuPagg remains monomeric under identical conditions of concentration and ionic strength [32]. 2 Porphyrin Interactions with Duplex DNA 2.1 Early Efforts to Establish Binding Modes Several lines of experimental evidence were offered by Fiel and coworkers for the intercalation of H2T4 into DNA, of which perhaps the
< previous page
page_371
next page >
< previous page
page_372
next page > Page 372
Fig. 2. Structures of metal derivatives of H2T4. Typical examples of each are (a) M = Cu2+ (b) M = Zn2+ (c) M = Mn3+ (Reproduced with permission from McGraw-Hill Yearbook of Science and Technology, 1989, p. 262.)
< previous page
page_372
next page >
< previous page
page_373
next page > Page 373
most compelling are the increase in viscosity with increasing drug load, the unwinding of circular DNA, and the increase in DNA melting point temperature [1,33]. The hypochroism of the porphyrin Soret absorption band and the avidity of the binding were also noted as being consistent with an intercalation bonding model. Two induced circular dichroism (CD) features were observed in the Soret region for the H2T4-DNA complex under the conditions employed, a negative band at about 440 nm and a positive one at shorter wavelength. Bisignate CD spectra of this type can arise from chromophore aggregation but, as pointed out by these authors, such is not the case here. The profile of the CD spectrum is roughly independent of drug load and represents two distinct binding modes, reported as intercalation and external, electrostatic binding. (However, the binding mode assignment made for the induced CD bands is the reverse of the one later demonstrated to be correct [34,35].) At about this time studies from our laboratories, including kinetic and spectroscopic investigations, helped to clarify the above binding model [34,36]. What is especially relevant for a contribution to this series is that the connection among porphyrin binding mode, base pair composition, and spectroscopic signatures grew out of systematic studies with metalloporphyrins. Metal complexes of the ''T4 porphyrin" which exist as (or could easily be converted to) fourcoordinate species are capable of intercalating, as is the metal-free, free base form. Other metal derivatives containing axial ligands are blocked from intercalation and instead form external complexes in which the metalloporphyrin nestles into a DNA groove [34,3739]. This model is consistent with the earlier observation (based on supercoiled-DNA unwinding studies) that, unlike H2T4, FeT4 does not intercalate into DNA [33]. It points out that the key factor in preventing intercalation of this latter derivative is the presence of axial ligands at the metal center. Based on work with synthetic DNAs, it was demonstrated that intercalation is favored at GC-rich regions of DNA while external binding is more favorable in AT-rich regions [34]. Theoretical studies later showed that intercalation is indeed favored over groove binding by about 100 kJ at GC base pairs while outside binding is more favorable at AT base pairs by some 40 kJ [40]. In addition, it was proposed that there is a very convenient spectroscopic signature for these interactions (Fig. 3); a negative induced CD band in the Soret region is diagnostic for intercalation while a positive feature indicates external, groove binding [34]an assignment then confirmed by Carvlin and Fiel [41].
< previous page
page_373
next page >
< previous page
page_374
next page > Page 374
Fig. 3. Induced CD spectra in the Soret region. Poly(dG-dC)2 with (·····) H2T4; () NiT4; (···) CuT4; (·) AuT4. Poly(dA-dT)2 with the same porphyrin derivatives. t-H2Pagg when added to B-form DNA in the presence of 100 mM NaCl. (---) 20% glycerol-water mixture in which the porphyrin is not appreciably aggregated and (·····) in an aggregated state. [Reproduced with permission from R. F. Pasternack and E. J. Gibbs, J. Inorg. Organomet. Polym., 3, 77 (1993).]
< previous page
page_374
next page >
< previous page
page_375
next page > Page 375
These generalizations about porphyrin-DNA interactions provided a basis for interpreting a broad range of experimental results including the observation of two induced CD bands for H2T4 interactions with natural DNAs, whereas only one (negative) band is obtained with poly(dG-dC)2 and only one (positive) band is obtained with poly(dA-dT)2 [34]. It should be pointed out, however, that no general correlation for all drug-DNA interactions exists between binding mode (intercalation vs. outside, groove binding) and sign of the induced CD feature. For some drug molecules, e.g., proflavin, the sign of the induced CD signal upon intercalation into DNA is positive [42], unlike the negative induced CD signal observed for porphyrins. The binding mode of H2T4 to natural DNAs depends sensitively on solution conditions [43]. At fixed concentration of porphyrin and DNA, H2T4 tends to intercalate at low salt concentration (providing a negative induced CD feature near the Soret maximum at 446 nm), but as salt is added, external binding becomes relatively more favorable with the appearance of its attendant positive induced CD feature (Fig. 4). To our knowledge only one other case has been documented in which the sign of the induced CD reverses with salt concentration as a result of a change in binding mode. Tuite and Nordén [44] recently proposed a redistribution of methylene blue from GC intercalation to AT external sites with increasing salt concentration to explain the change in sign of the induced CD from negative to positive. A number of studies employing other techniques helped to confirm and extend the model presented above. Nuclear magnetic resonance (NMR) investigations, particularly those by Marzilli, Wilson, and coworkers, have contributed to the understanding of these porphyrin-DNA interactions. Their initial report [45], combining NMR and viscometric titration techniques, confirmed that NiT4 and H2T4 intercalate while axially liganded ZnT4 does not. Additional studies in their labs probed various aspects of the binding model in some considerable detail, a subject to which we will return later. Kelly et al. [46] demonstrated that fluorescence and topoisomerase studies of DNA complexes are useful for distinguishing intercalators (like H2T4) from nonintercalators (like ZnT4). Geacintov and coworkers [47] using linear flow dichroism techniques also concluded that five-coordinate ZnT4 is an outside binder, with the plane of the porphine core making an angle of 6267° with respect to the twofold DNA helical axis, whereas H2T4 is
< previous page
page_375
next page >
< previous page
page_376
next page > Page 376
Fig. 4. Schematic diagram of intercalated and externally bound H2T4-DNA complexes. (Adapted with permission from Ref. 48.)
< previous page
page_376
next page >
< previous page
page_377
next page > Page 377
perpendicular to this axis as expected for an intercalator. They confirmed that the binding mode of the metal-free derivative changes with salt concentration in the manner previously proposed [43]. Further evidence for H2T4 being distributed between intercalated and outside bound forms comes from fluorescence lifetime studies [48,49] while electron spin resonance (ESR), luminescence, and electron nuclear double resonance (ENDOR) measurements are consistent with CuT4 intercalation [5053] and resonance Raman (RR) results [5458] support the model relating porphyrin structure and base pair composition to binding mode [34,35,59]. An important factor in determining the binding mode(s) of a given porphyrin to duplex DNA is the nature of the peripheral substituents. H2T3 (see Fig. 1) and its metal derivatives behave much the same as the ''T4" porphyrins whereas neither H2T2 nor H2TAP (nor their metal derivatives) gives any evidence of intercalation [41,54,60]. Analogs of H2T4 in which the quaternization of the para-pyridyl nitrogens is accomplished with moieties larger than CH3 have been investigated and, in general, the size of the substituent at this position plays little role in affecting the binding patterns of the porphyrin [61,62]. The intercalation of cis- but not trans-ortho-H2Ph2(NMePy)2 has been demonstrated by Sari and coworkers [63]. The intercalation complex formed by the cis isomer presumably has the porphyrin in a somewhat displaced position so as to relieve the steric constraints imposed by the ortho-N-methyl substituents. This nonsymmetrical intercalation model is consistent with theoretical studies of Ford et al. [64] and is favored by Nonaka et al. based upon infrared (IR) and RR studies [56]. The mixed periphery porphyrins shown in Fig. 1 (cis- and trans-para-H2Ph2-(NMePy)2; c- and t-H2Pagg) were also investigated for their binding avidity. Early reports based upon quenching of ethidium fluorescence (a method due to LePecq and Paoletti which assumes that the quenching is due to displacement of ethidium from the DNA double helix) suggested that these modifications had only a modest influence on the binding [65,66]. However, this last conclusion has to be reassessed [67]. First, the LePecq-Paoletti method must be applied with extreme caution to drugs whose absorption bands overlap with the fluorescence emission from ethidium, especially given the finding that the DNA matrix facilitates quenching via energy transfer [68,69]. Second, several of the porphyrins which were studied by Sari et al., tend to form
< previous page
page_377
next page >
< previous page
page_378
next page > Page 378
extended assemblies on the DNA surface, a subject to which we will return in some detail later in this chapter [23,32]. 2.2 Specifics of Porphyrin-DNA Interactions Early findings had led to the conclusion that intercalation is favored in regions of DNA rich in GC base pairs while AT base pairs are preferred for external, groove binding [34,36]. Dabrowiak and coworkers were able to confirm (in part) and expand on this model through footprinting studies [39,70]. They found that MnT4, an external binder, accumulates in an end-on fashion in the minor groove of DNA and especially in regions containing strings of at least three AT base pairs [39,71]. More recent work confirms the preference by FeT4, MnT4, and CoT4 for triplets containing only A and T bases [72]. It is suggested that external binding causes a kink in the DNA structure resulting from partial insertion of the porphyrin ligand between adjacent base pairs [47,7274]. The organization of DNA is thus disrupted at the site of binding by the breaking of hydrogen bonds between A and T, to allow for the partial ''inclusion" of the porphyrin. AT-rich regions are known to be easier to "melt" than GC regions and the binding energy is sufficient in the former case to overcome the loss of stabilization due to the decreased Watson-Crick base pairing. Footprinting results with intercalators were less conclusive. Ford et al. [74] attempted to account for the failure to obtain definitive results for intercalators by recalling published reports on Z-DNA which showed that one H2T4 molecule is capable of converting from 10 to 20 base pairs to a B configuration [75,76]. Such extensive deformations, they proposed, might also occur in other DNA forms, even the B form itself, and need to be considered when interpreting footprinting results. NMR, in contrast to footprinting, clearly shows that GC base pairs are preferred for intercalation and that the sequence 5'CpG3' is particularly favorable [77,78]. At first it was proposed that H2T4 intercalation occurred exclusively in this sequence (although metalloporphyrins such as NiT4, PdT4, and CuT4 show no such requirement), but that position has been relaxed [3,52,55,79,80]. Based on absorption and CD results it had been concluded that H2T4 can intercalate into the poly(dG-dC)·poly(dA-dT) polymer, which contains no CpG sites [81], and results of a detailed
< previous page
page_378
next page >
< previous page
page_379
next page > Page 379
NMR study are consistent with intercalation between AT base pairs, as long as they are flanked by GCs [79] (although this claim has recently been challenged [3]). It does appear certain, however, that although not required, the CpG sequence is particularly attractive for porphyrin intercalation [40]. However, it is well to bear in mind that from their modeling and footprinting studies, Ford et al. [74] conclude that binding to a particular dinucleotide pair can be strongly affected by the sequence in neighboring regions and caution against predictions based purely on local environment. In summary, certain generalizations concerning the relationship of DNA composition to porphyrin binding mode seem secure. Intercalation is favored in regions of DNA rich in GC base pairs [34] with the CpG site being particularly attractive for this binding mode [3]. External binding by monomeric, ground state porphyrins occurs primarily in the minor groove and especially in regions rich in AT base pairs (although some major groove binding has been detected for ZnT4 [82] and for MnT4 [83] at a high drug load or if no AT base pairs are present). The external, groove binding is end-on for the most part and causes some disruption of the AT hydrogen bonding pattern. If the steric features of the porphyrin are such as to prevent close contact with the minor groove of DNA, the AT specificity is relaxed [39] and the binding becomes almost entirely based on electrostatic interactions (''territorial" binding). 2.3 Does Axial Ligation Play a Role? Nucleoside and nucleotide binding to H2T4 and a number of its metal derivatives was studied using absorption spectroscopy, kinetic, and NMR techniques [84]. Values of the stability constants are comparable for a given nucleotide with the metal-free, copper(II), nickel(II), palladium(II), and zinc(II) derivatives. Virtually no interactions are observed for the six-coordinate iron(III), cobalt(III), or manganese(III) porphyrins. It was concluded from these findings that coordination of mononucleotides to axial sites does not play a major role in the formation of these complexes. ZnT4 behaves much the same as square coplanar porphyrin complexes of copper(II), nickel(II), and palladium(II) although it involves a five-coordinate metal site with the zinc out of the porphyrin plane and attached to a water molecule. This geometry
< previous page
page_379
next page >
< previous page
page_380
next page > Page 380
leaves the distal side of the ZnT4 available for reaction. What appears to be essential for the formation of complexes with mononucleotides is the availability of at least one side of the porphyrin plane without interference from an axial ligand. Furthermore, the nickel(II) derivative, which exists in solution in rapid equilibrium between a four- and sixcoordinate form [28,30], is converted on interacting with the nucleotides to the four-coordinate form; binding occurs preferentially to the nonaxially coordinated species for which ππ overlap can be maximized. Even when free in solution without the structural constraints imposed by being part of a helical polymer, nucleotides interact with porphyrins primarily through stacking interactions involving van der Waals forces. In a similar manner, axial ligation appears not to play a major role in the binding of ground state porphyrins to nucleic acids. Only for MnT4 binding to DNA has evidence been offered for some degree of ligation involving an ''A" or "T" base [85]. On the other hand, a strong and detailed case has been made for axial binding of excited state porphyrins and, in particular, CuT4 in the presence of "T" (or "U") residues [86,87]. At high-power laser excitation, an exciplex forms in the major groove at a site containing a minimum of four alternating AT base pairs. The metal is displaced but not removed from the porphyrin plane [86,88]. McMillin and coworkers demonstrated that consistent with this model, excited, externally bound CuT4 is five-coordinate [52]. The translocation of electronically excited CuT4 from a GC site where it is intercalated to AT sites where it forms the exciplex is a very rapid process, k > 3 × 107 sec-1 [89]. A mechanism proposed for intercalation of porphyrins into natural DNAs requires such a rapid translocation process [36]. 2.4 Kinetics The design of experiments to study the kinetics of porphyrin interactions with DNA was facilitated by findings relating binding mode to porphyrin/metalloporphyrin structure and DNA composition as described above. From these considerations, poly(dG-dC)2 proved to be the DNA polymer of choice for studies of intercalation by H2T4, CuT4, and NiT4 [36]. The amplitude profiles of the temperature jump effects obtained for these systems were shown to be consistent with titration
< previous page
page_380
next page >
< previous page
page_381
next page > Page 381
results with respect to both size of the equilibrium constants (~105106 M-1) and near-neighbor exclusion (n ~ 2). The rate constants for the intercalation process obtained in this study are somewhat smaller than those for drug molecules such as proflavin, not having steric constraints, but only modestly so; a value of ∆(∆G≠) was estimated at less than 13 kJ/mol. It was concluded from these results that coplanarity of the pyridyl substituent with the porphine ring system is not required for intercalation and that open regions and/or structural fluctuations must be present in the duplex to account for the relatively rapid incorporation of the porphyrin molecule. Modeling studies performed by Monaco and Zhao [90] and Ford et al. [64] are consistent with this conclusion of noncoplanarity of porphyrin substituents in the transition state complex. Chou and Mao [91] proposed a ''quake" model for DNA local disturbances, in part to account for the intercalation of porphyrins. They suggested that in the normal course of (dynamic) events for DNA, energy becomes transiently concentrated in small regions, leading to a simultaneous breaking of several adjacent hydrogen bonds (the "quake"). A number of processes associated with DNA reactivity, e.g., intercalation of drugs and premelting, occur in these regions. Other more recent models of DNA dynamics are also consistent with collective excitations involving many (1020) base pairs [92]. Why then do some porphyrins intercalate while others do not? For some systems, intercalation may be thermodynamically unfavorable compared to other bonding modes. For example, whereas t-H2Pagg and t-CuPagg intercalate into DNA of the appropriate composition at low Na+ concentration, with increasing ionic strength, the free energy of self-stacking at the surface of the DNA becomes more favorable than intercalation and the binding mode changes [23,32]. For some derivatives external groove binding can be energetically more favorable than intercalation even in the absence of aggregation. Marzilli, Dixon, and coworkers [93] suggested that the electron density at the porphine core is an important factor in determining binding preference; basic, electron-rich porphyrins are proposed as being less likely to intercalate. For some porphyrins, a kinetic barrier to intercalation exists. It may be that the deciding factor is not whether the porphyrin can enter the duplex but rather whether steric features of the porphyrin (size or placement of peripheral substituents, presence of one or more axial ligands) interfere with the subsequent closing of the duplex required to
< previous page
page_381
next page >
< previous page
page_382
next page > Page 382
keep the guest molecule in place. This may be the basis for the inability of the very acidic H2T2 to intercalate. However, this is not to imply that there cannot be cases in which the substituents at the porphyrin periphery are simply too large to permit passage through the transiently open regions. Reactions of porphyrins and metalloporphyrins with poly(dA-dT)2 were shown to be much faster than intercalation into poly(dG-dC)2 [36], a result confirmed in studies by Marzilli and coworkers [94,95]. External binding in the DNA groove is, as expected, much more rapid than insertion into the duplex. Competition kinetic studies, in which H2T4 was distributed between the two synthetic polymers, poly(dA-dT)2 and poly(dG-dC)2, were completely consistent with the results obtained for the thermodynamics and kinetics of the binary systems involving the porphyrin and one of the polymers only [36]. These studies led to the proposal that, in the mechanism of binding to natural DNAs, even intercalating porphyrins and metalloporphyrins first bind externally, presumably in AT-rich regions, and then redistribute to their intercalation binding sites via a direct internal transfer; in other words, a translocation along the duplex not involving dissociation. Evidence for such a kinetic process has been obtained by Strahan et al. [89] and Sugimoto et al. [96]. 3 Binding of Porphyrins and Metalloporphyrins to Other Nucleic Acid Structures 3.1 Interactions with Duplex and Transfer RNAs The association of H2T4 with duplex RNAs and, in particular with poly(A)·poly(U) and poly(I)·poly(C), was studied via spectrophotometric titrations [97]. Large bathochromic shifts with pronounced hypochromicity of the Soret band were observed for both polymers, reminiscent of the spectral features obtained for porphyrins intercalated into DNA. Unpublished work in our labs is consistent with an intercalation bonding mode for H2T4 with these synthetic duplex RNAs; negative induced CD signals are obtained in the Soret region of the spectrum and slow interaction kinetics are observed. In fact, the rate of porphyrin reaction with duplex RNAs is considerably slower than for poly(dG-dC)2. Under
< previous page
page_382
next page >
< previous page
page_383
next page > Page 383
similar conditions of temperature and ionic strength, the equilibrium constants for binding to the two RNAs are similar to that obtained for poly(dG-dC)2 [34,43,97]. However, much larger values for n (number of consecutive lattice sites made inaccessible by the binding of a molecule of the drug) were obtained for the RNAs (n ~ 1215) than for DNA (n ~ 2). Studies with transfer RNA have also produced some intriguing results. Foster and coworkers investigated the interactions of H2T4 and some of its metal derivatives with tRNAphe using a number of spectroscopic techniques including CD and NMR [98,99]. A persuasive argument was made for a unique mode of interaction for H2T4, CuT4, and ZnT4 with tRNA. The authors point out that the binding is neither intercalation in the usual sense of the term (i.e., insertion between two adjacent base pairs) nor a simple electrostatic, external interaction. Rather, the porphyrin is purported to locate at a specific site (leading to 1:1 stoichiometry) in a fold of the tertiary structure. This site, found at the outside of the elbow bend in a tRNA loop, is different from the typical binding location for intercalators like ethidium and proflavin. Six-coordinate MnT4 does not bind at this unique position on tRNA, leading the authors to propose that the interaction is restricted by the size of the site. However, it might be noticed that in many respects this interaction pattern parallels the one observed for mononucleotide binding to porphyrins described above [84]. A similar van der Waals interaction may be involved, for which an open side of the porphyrin must be available. 3.2 Interactions with Branched DNA The binding of porphyrins to synthetic DNAs of the type shown in Fig. 5, containing either a three-way or four-way branched junction, has been reported [100,101]. Chemical footprinting techniques provide evidence that the region in the four-way junction near the branch point is a high-affinity binding site for H2T4, CuT4, and NiT4. On the other hand, the nature of the interaction of ZnT4, CoT4, and MnT4, all of which contain axial ligands, is less easily defined. These derivatives also bind to four-way junction DNA, but in a manner different from porphyrins not having axial ligands although the reactivities of the cutting reagents employed for this study are consistent with some degree of insertion at
< previous page
page_383
next page >
< previous page
page_384
next page > Page 384
Fig. 5. Schematic drawings of a four-way (left) and three-way (right) junction used for porphyrin binding studies. (Adapted with permission from Refs. 100 and 101.) this special site. Similarly, binding of H2T4, H2T3, and H2TAP occurs at a high-avidity site located in the three-way junction region. The data are consistent with a model in which, at low drug load, the porphyrins are inserted almost exclusively at this location between the base pairs flanking the junction, a binding mode which the authors prefer to call ''inclusion" to distinguish it from "intercalation," which they reserve for insertional binding into duplex regions. As the drug load is increased, binding to secondary sites occurs but in patterns different for H2T4, which can intercalate, than for H2TAP, which cannot. In contrast, H2T2 binds weakly at best under all conditions investigated. Again, parallels exist for porphyrin inclusion at junctions and their interaction with tRNA or mononucleotides. For these interactions, porphyrins group themselves according to those that can form van der Waals-based donoracceptor adducts and those that cannot due to structural constraints. The structural restrictions which apply for porphyrin intercalation are not generally applicable for these cases; H2TAP, which can form a tight junction complex, does not intercalate. The absorption spectra obtained for H2TAP with such junctions are sufficiently unique
< previous page
page_384
next page >
< previous page
page_385
next page > Page 385
that this porphyrin may be useful for identifying such regions in natural DNAs. 3.3 Binding Studies with DNA Ψ Aggregates Psi (for ''polymer salt-induced") aggregates, Ψ, of DNA have been extensively studied because of their unusual spectral properties and their similarity to some biologically compacted DNA. Whereas DNA can be caused to condense without any dramatic effect on its CD spectrum by, for example, the addition of , other reagents, such as synthetic cationic polypeptides, can produce remarkable changes. The CD signal in the ultraviolet region not only loses its bisignate profile, becoming either positive or negative depending on the peptide chosen, but also becomes enormous in magnitude. Tinoco and coworkers accounted for these features in terms of ordered assemblies of DNA involving periodic repeats [102]. The very large signals are purported to arise from the electronic coupling of the DNA bases in the assembly, i.e., the entire assembly can be considered to behave as an antenna in which electronic communication occurs over long distances. As an extension of their work on such Ψ species, the Tinoco and Maestre group considered the binding of drugs to these assemblies [103] for which they utilized several porphyrins which had been shown to intercalate (H2T4, CuT4, NiT4) and one nonintercalator (MnT4). Whereas the nonintercalator produced no measurable CD signal in the Soret region, the intercalators provided very large signals whose sign correlated with whether they were bound to Ψ(+) or Ψ(). The previously discussed pattern relating the sign of the (relatively small) induced CD signal to the binding mode is now overridden by a correlation with the superhelical sense of the DNA assembly. In more recent work [104], we concluded that the distinction between intercalators and nonintercalators in producing enhanced CD signals when bound to Ψ-type DNA could be overstated; such large induced signals were obtained with the five-coordinate, nonintercalating ZnT4. However, the key conclusions of Tinoco, Maestre, and coworkers remain intact. The electron density of the porphyrin couples to that of the polynucleotide "antenna" and to other bound molecules. This self-recognition and communication among the porphyrin moieties is accomplished without direct contact. We have shown
< previous page
page_385
next page >
< previous page
page_386
next page > Page 386
through the use of resonance light scattering (to be discussed later in Sec. 4.2) that the porphyrins do not form extended self-aggregates under the conditions at which they produce enhanced CD signals. 4 Porphyrin Assemblies on Nucleic Acids 4.1 Assemblies on Duplex DNA The question of aggregation state of DNA-bound porphyrins has interested workers in the field almost from the start. An early report indicated that H2TAP, when added to DNA under high drug load conditions, aggregates on the DNA surface [60]. The induced CD signal under these conditions is bisignate and conservative. Although the magnitude of the features is not very different from those for monomeric porphyrins, it was thought that this complex involved an extended array of porphyrins. However, other evidence (CD and RLS; vide infra) suggests that these aggregates probably are more modest in size. Similar results were obtained for NiT4 with poly(dA-dT)2 [34] but, once again, it is likely that the aggregation is limited, in contrast to the extended assemblies described below. In studies of DNA-bound t-H2Pagg (see Fig. 1), induced CD signals of an unusual shape and size were observed [23,24]. The signals are bisignatebut markedly nonsymmetricaland from one to two orders of magnitude larger than for bound H2T4 or H2TAP; values of approach 103 M1 cm1 (see Fig. 3). The shape and especially the size of the CD signals could be best accounted for in terms of an extended, electronically coupled, organized array of porphyrin molecules. Such arrays of chromophores on DNA are not without precedent. The aggregation of acridine orange on polynucleotide surfaces, for example, has been investigated from both an experimental [105] and theoretical point of view [106]. In addition to the metal-free derivative, t-CuPagg was found to produce large induced CD signals whereas t-AuPagg gives induced CD patterns similar to those for CuT4 or other monomeric fourcoordinate metalloporphyrin intercalators on binding to DNA [32]. Although it was apparent that the DNA-bound t-H2Pagg or t-CuPagg porphyrins are assembled in some extended manner involving periodic
< previous page
page_386
next page >
< previous page
page_387
next page > Page 387
repeats (to account for the size and bisignate nature of the induced CD signals), we could be less certain as to the state of aggregation of the DNA. The ultraviolet CD spectrum due to DNA is virtually unaffected by the porphyrin aggregation; it resembles quite closely the CD spectrum of native DNA in the absence of drugs. However, CD spectra are not necessarily sensitive to aggregation state. Indirect evidence was offered for the dispersed nature of the DNA but it was not until the development and application of resonance light scattering to this system that the matter was finally resolved [107]. 4.2 A Brief Introduction to Resonance Light Scattering Light scattering experiments typically involve measurements away from absorption bands. What has apparently not been generally appreciated is that light scattering at wavelengths within the absorption band enveloperesonance light scattering (RLS)can be extremely informative if the absorption is not too great and the aggregate size is sufficiently large. The RLS effect is observed as increased scattering intensity at or very near the wavelength of absorption of an aggregated molecular species. The effect can be enhanced by several orders of magnitude when strong electronic coupling exists among the chromophores [107109]. DNA, aggregated by the addition of hexaamminecobalt(III) at low salt conditions, scatters light extensively in the UV region (Fig. 6). When t-H2Pagg is added to such condensed DNA, an RLS profile is obtained in the visible region. The UV scattering profile due primarily to DNA is almost unaffected, still showing the impact of its being in a condensed state. When these two profiles are compared to one obtained in the UV region for a t-H2Pagg/DNA solution in which the DNA has not been previously condensed with hexaamminecobalt(III), it is apparent that the DNA is not extensively aggregated. Yet the porphyrin provides a scattering profile and CD spectrum in the visible range, characteristic of a highly aggregated species. Thus, the combination of RLS and CD measurements provides a very powerful experimental strategy for the study of chromophore assemblies and, in this case, confirms the model proposed for DNA-bound porphyrin arrays.
< previous page
page_387
next page >
< previous page
page_388
next page > Page 388
Fig. 6. Light scattering measurements of DNA and DNA/t-H2Pagg complexes. (---) DNA condensed with hexaamminecobalt(III); () t-H2Pagg added to DNA previously condensed with hexaammine cobalt(III); (........) t-H2Pagg-DNA complex to which no hexaammine cobalt(III) has been added. (Reproduced with permission from Ref. 107.) 4.3 Porphyrin Assemblies Revisited The combined spectroscopic evidence points to a model in which t-H2Pagg and t-CuPagg are extensively aggregated on a DNA surface. Once the concentrations and ionic strengths have been reached which constitute the threshhold for this highly cooperative process, the porphyrins relocate from their intercalated and/or minor groove-bound positions to form an extended, organized assembly on the nucleic acid backbone. Other conditions remaining equal, the assemblies form more readily in the order [23,24]: poly(dG-dC)2 > calf thymus DNA > poly(dA-dT)2, an order reflecting the stiffness of the polynucleotide backbone on which the aggregates form. The aggregation of the porphyrin can occur even in the absence of a polynucleotide but the information needed for the formation of a helical assembly (as detected through the remarkably large CD signals) is provided by the polymerthe polymeric scaffolds serve as ''templates" for the organized assembly.
< previous page
page_388
next page >
< previous page
page_389
next page > Page 389
These studies relating properties of porphyrins, polymer templates, and medium to assembly formation illustrate how exquisitely sensitive drug-nucleic acid interactions can be to solvent conditions. This dependence was previously indicated in studies of the change in binding mode of H2T4 to DNA as a function of salt concentration. The influence of bulk ionic strength is seen even more clearly in the highly cooperative processes described here. Figure 7, for example, shows the manner in which the magnitude of CD signals reporting assembly formation depends on salt concentration. The curves shown are for a cooperativity factor of 57 for Hill-type dependencies. Recently, a porphyrin derivative was reported as having such a strong tendency to aggregate on DNA that, under all conditions of drug load and ionic strength considered, extremely large, bisignate-induced CD features were observed [93]. Although RLS experiments have not as yet been reported for this system, there is every reason to believe that extended assemblies are formed in this case as well.
Fig. 7. Induced CD spectra for two aggregating porphyrin-DNA complexes as a function of NaCl concentration. ( ) t-H2Pagg; ( ) t-CuPagg. (Reproduced with permission from Ref. 107.)
< previous page
page_389
next page >
< previous page
page_390
next page > Page 390
5 New Directions and Closing Remarks Porphyrin/nucleic acid research is evolving in a number of diverse directions for which the fundamental solution chemistry described here serves as a basis. Among these studies are the synthesis of new, highly specific and efficient metalloporphyrin nucleases (as discussed in Chap. 13 of this volume) and the production of supramolecular assemblies. Such assemblies have been characterized for double-stranded [107] and single-stranded polynucleotide templates [110,111] and offer promise for the construction of nanodevices to serve as reaction centers, molecular wires, or light-harvesting arrays involving long-range energy and/or electron transfer. Model studies have shown that several of these processes are facilitated by the nucleic acid polymer. The new generation of synthetic nucleases can serve not only as useful laboratory probes but have considerable clinical potential. Recent reports describe efforts to model and synthesize porphyrin derivatives useful for the treatment of nucleic acid-based diseases including AIDS and various forms of cancer. Pitié et al. [112], using an MnT4 derivative tethered to a 19-mer oligonucleotide, reported the selective cleavage of DNA containing the codon of the TAT gene of HIV-1. The cleavage process is initiated by chemical activation while an alternative approach, reported by Mastruzzo et al. [113], involves selective photodamage. By covalently linking free base and zinc porphyrins to complementary antisense RNA molecules, a catalyzed photocrosslinking reaction can be targeted to occur at specific sites through hybrid formation. Modeling studies have defined a bisarginyl derivative of a tricationic intercalating porphyrin which is purported to be capable of targeting the palindromic sequence d(GGCGCC)2, encountered in the genomes of retrovirus HIV-1 and oncogenes [114]. These various substances have been designed so as to heighten their selectivity and effectiveness at recognizing and destroying target polynucleotides. Acknowledgment The present understanding of the interactions of porphyrins and metalloporphyrins with nucleic acids represents the contributions from many laboratories. Central to this effort at Swarthmore and Goucher Colleges
< previous page
page_390
next page >
< previous page
page_391
next page > Page 391
has been the role played by undergraduate collaborators. They have contributed to our research programs over the years through their interest, enthusiasm, commitment, and creativity. It is to these students that we dedicate this chapter. Abbreviations
CD
circular dichroism
ENDOR
electron nuclear double resonance
ESR
electron spin resonance
c-H2Pagg
cis-bis(methylpyridinium-4-yl)diphenylporphine (see Figure 1)
t-H2Pagg
trans-bis(methylpyridinium-4-yl)diphenylporphine (see Figure 1)
ortho-H2Ph(NMePy)2
bis(methylpyridinium-2-yl)diphenylporphine
para-H2Ph(NMePy)2
bis(methylpyridinium-4-yl)diphenylporphine
H2T2
tetrakis(methylpyridinium-2-yl)porphine (see Figure 1)
H2T3
tetrakis(methylpyridinium-3-yl)porphine (see Figure 1)
H2T4
tetrakis(methylpyridinium-4-yl)porphine (see Figure 1)
H2TAP
tetrakis(methylanilinium-4-yl)porphine (see Figure 1)
HIV
human immunodeficiency virus
IR
infrared
NMR
nuclear magnetic resonance
RLS
resonance light scattering
RR
resonance Raman
UV
ultraviolet
Ψ aggregates
polymer salt-induced (= Psi) aggregates
References
1. R. J. Fiel, J. C. Howard, E. H. Mark, and N. Datta Gupta, Nucl. Acids Res., 6, 3093 (1979).
< previous page
page_391
next page >
< previous page
page_392
next page > Page 392
2. M. R. Hamblin and E. L. Newman, J. Photochem. Photobiol. B, 23, 3 (1994). 3. L. G. Marzilli, New J. Chem., 14, 409 (1990). 4. T. J. Dougherty in Advances in Photochemistry (D. H. Volman, G. S. Hammond, and D. C. Neckers, eds.), John Wiley and Sons, New York, 1992, p. 275. 5. R. D. Levere, Y.-F. Gong, A. Kappas, D. J. Bucher, G. P. Wormser, and N. G. Abraham, Proc. Natl. Acad. Sci. USA, 88, 1756 (1991). 6. D. W. Dixon, L. G. Marzilli, and R. F. Schinazi, Ann. N.Y. Acad. Sci., 616, 511 (1990). 7. M. Asanaka, T. Kurimura, H. Toya, K. Ogaki, and Y. Kato, AIDS, 3, 403 (1989). 8. S. S. Eaton and G. R. Eaton, J. Am. Chem. Soc., 97, 3660 (1975). 9. S. S. Eaton, D. M. Fiswild, and G. R. Eaton, Inorg. Chem., 17, 1542 (1978). 10. D. K. Lavallee, Coord. Chem. Rev., 61, 55 (1985). 11. P. Hambright in Porphyrins and Metalloporphyrins (K. M. Smith, ed.), Elsevier, Amsterdam, 1975, p. 233. 12. W. Schneider, Struct. Bond., 23, 123 (1975). 13. H. Baker, P. Hambright, and L. Wagner, J. Am. Chem. Soc., 95, 5942 (1973). 14. G. Pethö, N. B. Elliott, M. S. Kim, M. Lin, D. W. Dixon, and L. G. Marzilli, J. Chem. Soc., Chem. Commun., 20, 1547 (1993). 15. H. J. Schneider and M. Wang, J. Org. Chem., 59, 7473 (1994). 16. N. E. Mukundan, G. Pethö, D. W. Dixon, M. S. Kim, and L. G. Marzilli, Inorg. Chem., 33, 4676 (1994). 17. R. F. Pasternack, R. Huber, P. Boyd, G. Engasser, L. Francesconi, E. Gibbs, P. Fasella, G. C. Venturo, and L. de C. Hinds, J. Am. Chem. Soc., 94, 4511 (1972). 18. R. F. Pasternack, L. Francesconi, D. Raff, and E. Spiro, Inorg. Chem., 12, 2606 (1973). 19. R. F. Pasternack, Ann. N.Y. Acad. Sci., 206, 614 (1973). 20. F. J. Vergeldt, R. B. M. Koehorst, A. van Hoek, and T. J. Schaafsma, J. Phys. Chem., 99, 4397 (1995). 21. K. M. Kadish, B. G. Maiya, and C. Araullo-McAdams, J. Phys. Chem., 45, 427 (1991).
< previous page
page_392
next page >
< previous page
page_393
next page > Page 393
22. A. S. Ito, G. C. Azzellini, S. C. Silva, O. Serra, and A. G. Szabo, Biophys. Chem., 45, 79 (1992). 23. E. J. Gibbs, I. Tinoco, Jr., M. F. Maestre, P. A. Ellinas, and R. F. Pasternack, Biochem. Biophys. Res. Commun., 157, 350 (1988). 24. P. A. Ellinas, in B.A. thesis, Solution properties and interactions with calf-thymus DNA of three derivatives of tetrakis(4-N-methylpyridy)porphyrin, Swarthmore College, 1987. 25. R. F. Pasternack and W. Qu, unpublished results. 26. R. F. Pasternack and M. A. Cobb, Biochem. Biophys. Res. Commun., 51, 507 (1973). 27. R. F. Pasternack and M. A. Cobb, J. Inorg. Nucl. Chem., 35, 4327 (1973). 28. R. F. Pasternack, E. G. Spiro, and M. Teach, J. Inorg. Nucl. Chem., 36, 599 (1974). 29. R. F. Pasternack, H. Lee, P. Malek, and C. Spencer, J. Inorg. Nucl. Chem., 39, 1865 (1977). 30. R. F. Pasternack, N. Sutin, and D. H. Turner, J. Am. Chem. Soc., 98, 1908 (1976). 31. M. Krishnamurthy and J. R. Sutton, Inorg. Chem., 24, 1943 (1985). 32. R. F. Pasternack, A. Giannetto, P. Pagano, and E. J. Gibbs, J. Am. Chem. Soc., 113, 7799 (1991). 33. R. J. Fiel and B. R. Munson, Nucl. Acids Res., 8, 2835 (1980). 34. R. F. Pasternack, E. J. Gibbs, and J. J. Villafranca, Biochemistry, 22, 2406 (1983). 35. E. J. Gibbs and R. F. Pasternack, Sem. Hematol., 26, 77 (1989). 36. R. F. Pasternack, E. J. Gibbs, and J. J. Villafranca, Biochemistry, 22, 5409 (1983). 37. M. Lin, M. Lee, K. T. Yue, and L. G. Marzilli, Inorg. Chem., 32, 3217 (1993). 38. G. Dougherty, J. Inorg. Biochem., 34, 95 (1988). 39. S. D. Bromley, B. W. Ward, and J. C. Dabrowiak, Nucl. Acids Res., 14, 9133 (1986). 40. X. Hui, N. Gresh, and B. Pullman, Nucl. Acids Res., 18, 1109 (1990). 41. M. J. Carvlin and R. J. Fiel, Nucl. Acids Res., 11, 6121 (1983).
< previous page
page_393
next page >
< previous page
page_394
next page > Page 394
42. P. E. Pieter, E. Schipper, B. Nordén, and F. Tjerneld, Chem. Phys. Lett., 70, 17 (1980). 43. R. F. Pasternack, P. Garrity, B. Ehrlich, C. B. Davis, E. J. Gibbs, G. Orloff, A. Giartosio, and C. Turano, Nucl. Acids Res., 14, 5919 (1986). 44. E. Tuite and B. Nordén, J. Am. Chem. Soc., 116, 7548 (1994). 45. D. L. Banville, L. G. Marzilli, and D. W. Wilson, Biochem. Biophys. Res. Commun., 113, 148 (1983). 46. J. M. Kelly, M. J. Murphy, D. J. McConnell, and C. OhUigin, Nucl. Acids Res, 13, 167 (1985). 47. N. E. Geacintov, V. Ibanez, M. Rougee, and R. V. Bensasson, Biochemistry, 26, 3087 (1987). 48. Y. Shen, P. Myslinski, T. Treszczanowicz, U. Liu, and J. A. Koningstein, J. Phys. Chem., 96, 7782 (1992). 49. Y. Lui, J. A. Koningstein, and Y. Yevdokimov, Can. J. Chem., 69, 1791 (1991). 50. G. Dougherty, J. R. Pilbrow, A. Skorobogaty, and T. D. Smith, J. Chem. Soc. Faraday Trans., 81, 1739 (1985). 51. S. P. Greiner, R. W. Kreilick, and L. G. Marzilli, J. Biomol. Struct. Dynam., 9, 837 (1992). 52. B. P. Hudson, J. Sou, D. J. Berger, and D. R. McMillin, J. Am. Chem. Soc., 114, 8997 (1992). 53. G. Dougherty and R. F. Pasternack, Biophys. Chem., 44, 11 (1992). 54. K. Büjte and K. Nakamoto, J. Inorg. Biochem., 39, 75 (1990). 55. K. Büjte, J. H. Schneider, J.-J. P. Kim, Y. Wang, S. Ikuta, and K. Nakamoto, J. Inorg. Biochem., 37, 119 (1989). 56. Y. Nonaka, D. S. Lu, A. Dwivedi, D. P. Strommen, and K. Nakamoto, Biopolymers, 29, 999 (1990). 57. J. H. Schneider, J. Odo, and K. Nakamoto, Nucl. Acids Res., 16, 10323 (1988). 58. K. T. Yue, M. Lui, T. A. Gray, and L. G. Marzilli, Inorg. Chem., 30, 3214 (1991). 59. R. F. Pasternack and E. J. Gibbs, ACS Symposium Series (Metal-DNA Chemistry), 402, 519 (1990). 60. M. J. Carvlin, N. Datta-Gupta, and R. J. Fiel, Biochem. Biophys. Res. Commun., 108, 66 (1982).
< previous page
page_394
next page >
< previous page
page_395
next page > Page 395
61. T. A. Gray, K. T. Yue, and L. G. Marzilli, J. Inorg. Biochem., 41, 205 (1991). 62. R. Kuroda, E. Takahashi, C. A. Austin, and L. M. Fisher, FEES Lett, 262, 293 (1990). 63. M. A. Sari, J. P. Battioni, D. Dupre, D. Mansuy, and J. B. LePecq, Biochem. Pharmacol., 37, 1861 (1988). 64. K. G. Ford, L. H. Pearl, and S. Neidle, Nucl. Acids Res., 15, 6553 (1987). 65. M. A. Sari, J. P. Battioni, C. Mansuy, and J. B. LePecq, Biochem. Biophys. Res. Commun., 141, 643 (1986). 66. M. A. Sari, J. P. Battioni, D. Dupre, D. Mansuy, and J. B. LePecq, Biochemistry, 29, 4205 (1990). 67. L. Ding, J. Bernadou, and B. Meunier, Bioconj. Chem., 2, 201 (1991). 68. R. F. Pasternack, M. Caccam, B. Keogh, T. A. Stephenson, A. P. Williams, and E. J. Gibbs, J. Am. Chem. Soc., 113, 6835 (1991). 69. A. M. Brun and A. Harriman, J. Am. Chem. Soc., 116, 10383 (1994). 70. B. Ward, A. Skorobogaty, and J. C. Dabrowiak, Biochemistry, 25, 7827 (1986). 71. G. Raner, B. Ward, and J. C. Dabrowiak, J. Coord. Chem., 19, 17 (1988). 72. U. Sehlstedt, S. K. Kim, P. Carter, J. Goodisman, J. F. Vollano, B. Nordén, and J. C. Dabrowiak, Biochemistry, 33, 417 (1994). 73. J. Bernadou, P. Gelas and B. Meunier, Tetrahedron Lett., 29, 6615 (1988). 74. K. Ford, K. R. Fox, S. Neidle, and M. J. Waring, Nucl. Acids Res., 15, 2221 (1987). 75. R. F. Pasternack, D. Sidney, P. A. Hunt, E. A. Snowden, and E. J. Gibbs, Nucl. Acids Res., 14, 3927 (1986). 76. R. E. McKinnie, J. D. Choi, J. W. Bell, E. J. Gibbs, and R. F. Pasternack, J. Inorg. Biochem., 32, 207 (1988). 77. D. L. Banville, L. G. Marzilli, J. A. Strickland, and W. D. Wilson, Biopolymers, 25, 1837 (1986). 78. L. G. Marzilli, D. L. Banville, G. Zon and W. D. Wilson, J. Am. Chem. Soc., 108, 41881 (1986). 79. R. J. Fiel, J. Biomol. Struct. Dynam., 6, 1259 (1989).
< previous page
page_395
next page >
< previous page
page_396
next page > Page 396
80. J. A. Strickland, D. L. Banville, W. D. Wilson, and L. G. Marzilli, Inorg. Chem., 26, 3398 (1987). 81. E. J. Gibbs, M. C. Maurer, J. H. Zhang, W. M. Reiff, D. T. Hill, M. Malicka-Blaszkiewicz, R. E. McKinnie, H.Q. Liu, and R. F. Pasternack, J. Inorg. Biochem., 32, 39 (1988). 82. Y. Lui and J. A. Koningstein, J. Phys. Chem., 97, 6155 (1993). 83. R. Kuroda and H. Tanka, J. Chem. Soc. Commun., 1575 (1994). 84. R. F. Pasternack, E. J. Gibbs, A. Gaudemer, A. Antebi, S. Bassner, L. De Poy, D. H. Turner, A. Williams, F. LaPlace, M. H. Lansard, C. Merienne, and M. Perreé-Fauvet, J. Am. Chem. Soc., 107, 8179 (1985). 85. Y.-X. Ci, Y.-G. Zheng, J.-K. Tie, and W.-B. Chang, Anal. Chim. Acta, 282, 695 (1993). 86. P. Mojzes, L. Chinsky, and P.-Y. Turpin, J. Phys. Chem., 97, 4841 (1993). 87. P.-Y. Turpin, L. Chinsky, A. Laigle, M. Tsuboi J. R. Kincaid, and K. Nakamoto, Photochem. Photobiol., 51, 519 (1990). 88. L. Chinsky, P.-Y. Turpin, A. H. R. Al-Obaidi, S. E. J. Bell, and R. E. Hester, J. Phys. Chem., 95, 5754 (1991). 89. G. D. Strahan, D. Lu, M. Tsuboi, and K. Nakamoto, J. Phys. Chem., 96, 6450 (1992). 90. R. R. Monaco and M. Zhao, Int. J. Quantum Chem., 46, 701 (1993). 91. K.-C. Chou and B. Mao, Biopolymers, 27, 1795 (1988). 92. G. Gaeta, Phys. Lett. A., 190, 301 (1994). 93. L. G. Marzilli, G. Pethö, M. Lin, M. S. Kim, and D. W. Dixon, J. Am. Chem. Soc., 114, 7575 (1992). 94. J. A. Strickland, L. G. Marzilli, and W. D. Wilson, Biopolymers, 29, 1307 (1990). 95. J. A. Strickland, L. G. Marzilli, K. M. Gay, and W. D. Wilson, Biochemistry, 27, 8870 (1988). 96. N. Sugimoto, K. Hasegawa, and M. Sasaki, Bull. Chem. Soc. Jpn., 63, 1641 (1990). 97. N. Sugimoto, K. Hasegawa, N. Monden, and M. Sasaki, Chem. Exp., 5, 399 (1990). 98. W. J. Birdsall, W. R. Anderson, Jr., and N. Foster, Biochim. Biophys. Acta, 1007, 176 (1989).
< previous page
page_396
next page >
< previous page
page_397
next page > Page 397
99. N. Foster, A. K. Singhal, M. W. Smith, N. G. Marcos, and K. J. Schray, Biochim. Biophys. Acta, 950, 118 (1988). 100. M. Lu, Q. Guo, R. F. Pasternack, D. J. Wink, N. C. Seeman, and N. R. Kallenbach, Biochemistry, 29, 1614 (1990). 101. J. M. Nussbaum, M. E. A. Newport, M. Mackie, and B. B. Leontis, Photochem. Photobiol., 59, 515 (1994). 102. D. Keller, C. Bustamante, M. F. Maestre, and I. Tinoco, Jr., Biopolymers, 24, 783 (1985). 103. C. L. Phillips, W. E. Mickols, M. F. Maestre and I. Tinoco, Jr., Biochemistry, 25, 7803 (1986). 104. D. Bloomquist, in BA thesis, Interaction of nickel(II), manganese(III) and zinc(II) substituted tetrakis(Nmethylpyridinium-4-yl)porphines with DNA and psi(+) aggregated DNA, Swarthmore College, 1994. 105. D. F. Bradley and M. K. Wolf, Proc. Natl. Acad. Sci. USA, 45, 944 (1959). 106. I. Tinoco, Jr., R. W. Woody, and D. F. Bradley, J. Chem. Phys., 38, 1317 (1963). 107. R. F. Pasternack, C. Bustamante, P. J. Collings, A. Giannetto, and E. J. Gibbs, J. Am. Chem. Soc., 115, 5393 (1993). 108. R. F. Pasternack, K. F. Schaefer, and P. Hambright, Inorg. Chem., 33, 2062 (1994). 109. J. C. dePaula, J. H. Roblee, and R. F. Pasternack, Biophys. J., 68, 335 (1995). 110. C. Bustamante, S. Gurrieri, R. F. Pasternack, R. Purrello, and E. Rizzarelli, Biopolymers, 34, 1099 (1994). 111. R. F. Pasternack, S. Gurrieri, R. Lanceri, and R. Purrello, Inorg. Chim. Acta, in press. 112. M. Pitié, C. Casas, J. C. Lacey, G. Pratviel, J. Bernadou, and B. Meunier, Angew. Chem. Int. Ed. Eng., 32, 557 (1993). 113. L. Mastruzzo, A. Woisard, D. D. F. Ma, E. Rizzarelli, A. Favre, and T. LeDoan, Photochem. Photobiol., 60, 316 (1994). 114. M. Perrée-Fauvet and N. Gresh, submitted.
< previous page
page_397
next page >
< previous page
page_399
next page > Page 399
13 Selective DNA Cleavage by Metalloporphyrin Derivatives Geneviève Pratviel, Jean Bernadou, and Bernard Meunier Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, F-31077 Toulouse Cedex, France
400
1. Introduction
400
2. DNA Cleavage by Nonvectorized Metalloporphyrin Complexes
400
2.1. Nuclease Activity of Photoactivatable Porphyrin Derivatives
403
2.2. Nuclease Activity of Manganese and Iron Porphyrins Activated by Oxidants
404
2.3. Detailed Mechanism of DNA Cleavage by MnIII-TMPyP/KHSO5
412
2.4. ''Pseudohydrolysis" of DNA
413
3. Selective DNA Cleavage by Vectorized Metalloporphyrins
413
3.1. Metalloporphyrin Intercalators
417
3.2. Metalloporphyrin Oligonucleotides
4. Concluding Remarks: Is It Possible to Go from Oxidative DNA Cleavage to Drugs Based on Metalloporphyrins?
< previous page
page_399
420
next page >
< previous page
page_400
next page > Page 400 421
Acknowledgments
421
Abbreviations
421
References
1 Introduction Metalloporphyrin derivatives are versatile DNA cleavers due to easy modifications of the macrocycle and of the nature of the central metal ion, which allow modulation of their DNA affinity and change of their mode of activation and mechanism of action. In this chapter we will describe: 1. The DNA cleavage by regular metalloporphyrins, i.e., porphyrin derivatives which are not tethered to a vector. Two major modes of activation, by light (Sec. 2.1) or by oxidants (Sec. 2.2), will be described. The detailed mechanism of DNA cleavage will be presented for the manganeseIII meso-tetrakis(4-N-methylpyridiniumyl)porphyrin, Mn-TMPyP, activated by potassium monopersulfate (Sec. 2.3). The possibility of using this artificial nuclease to perform the ''pseudohydrolysis" of DNA phosphodiesters will also be reported (Sec. 2.4). 2. The selective DNA cleavage by metalloporphyrins linked to a molecule able to target the nuclease activity on DNA or RNA. Two categories of vectors will be presented: intercalating agents (Section 3.1) or oligonucleotides (Section 3.2). The concluding remarks will address the following question: Is it possible to go from oxidative DNA cleavage by porphyrin derivatives to the design of molecules active as anticancer or antiviral agents? 2 DNA Cleavage by Nonvectorized Metalloporphyrin Complexes 2.1 Nuclease Activity of Photoactivatable Porphyrin Derivatives Two different categories of photoactivatable porphyrin derivatives are currently being studied as photosensitizers in the photodynamic ther-
< previous page
page_400
next page >
< previous page
page_401
next page > Page 401
apy approach: (1) derivatives of hematoporphyrin like HpD, which is a modified protoporphyrin IX with two αhydroxyethyl groups instead of two vinyl groups, (2) fully synthetic porphyrin derivatives having a noticeable affinity for DNA. The first category corresponds to porphyrin molecules having hydrophobic substituents at the periphery of the tetrapyrrole ring or substituents bearing hydroxy or carboxylic acid functions. Some of these HpD derivatives have been proposed in phototherapy of cancers in human clinical trials, such as Photofrin which is currently used in the treatment of superficial bladder cancer [14]. The toxicity of these porphyrin derivatives on tumor cells is due to their affinity for membranes [3] and may be explained, when they are activated with red light at 630 nm, by the generation of reactive singlet oxygen in a type II photochemical reaction. Current efforts are focused on the preparation of new macrocyclic conjugated molecules having a strong absorption between 600 and 700 nm, in the region corresponding to the maximum of light penetration through tissues. Phthalocyanine or pyropheophorbide derivatives are potential new photosensitizers [57]. The second category of photoactivatable porphyrins corresponds to molecules exhibiting a reasonable affinity for DNA or RNA. The paradigm molecule in this domain is the meso-tetrakis (4-N-methylpyridiniumyl)porphyrin (H2TMPyP, Fig. 1 shows metallated derivatives of this ligand). Interactions of cationic porphyrin derivatives with DNA are exhaustively described in Chap. 12, but some key features which are governing their photonuclease properties should be remembered. H2TMPyP and its metal derivatives without axial ligands (NiII-
Fig. 1. Structure of M-TMPyP pentaacetate (M = Mn or Fe).
< previous page
page_401
next page >
< previous page
page_402
next page > Page 402
TMPyP and CuII-TMPyP) behave like intercalating molecules with poly(dG-dC)2 (see [810] for review articles) when they bind externally into poly(dA-dT)2, probably in the minor groove because of its more attractive molecular electrostatic potential [816]. The affinity constants of cationic porphyrin derivatives for DNA or RNA have been determined by several methods [11,1619] and, as example, the value for H2TMPyP/calf thymus DNA (CT DNA) is 7.7 × 105 M1 [11,18]. Sari et al. suggested that only half of the porphyrin ring is necessary for DNA intercalation [11]. Kinetic data of the interaction of H2TMPyP with poly(dG-dC)2 showed that rate constants for association and dissociation are 3.7 × 105 M1 sec1 and 1.8 sec1, respectively [16]. Finally, microinjections of H2TMPyP within the cytoplasm of mouse fibroblast cells indicate, as judged by fluorescence intensity, that most of the dye is located within the cell nuclei, 2 hr after the porphyrin injection [9]. The triplet lifetime of H2TMPyP is 170 msec, instead of 2.0 msec for the corresponding zinc derivative [20]. The photonuclease activity of H2TMPyP was initially reported by Fiel et al. in 1981 [21]. In further studies by Kelly et al. [22] and Gaudemer et al. [23] it was confirmed that H2TMPyP and Zn-TMPyP are able to photolyse supercoiled plasmid DNA to open circular and linear DNA by irradiation at 436 nm, near the Soret band maxima of these two porphyrin derivatives. The efficiency of photocleavage of DNA by cationic porphyrins is not directly correlated to the number of positive charges: cis- trans-[di(4-N-methylpyridiniumyl)diphenyl]porphyrin are more efficient DNA cleavers than H2TMPyP) or [tris(4-N-methylpyridiniumyl)phenyl]porphyrin, despite a smaller DNA affinity constant (eightfold smaller than that measured for H2TMPyP) [24]. Single-strand (ss) breaks are the main damage induced on double-stranded (ds) DNA by photoactivated H2TMPyP and it occurs mainly at guanine and thymine bases. Inhibition of cleavage by sodium azide or by nitrogen bubbling (dioxygen removal) strongly suggests that DNA breaks are mediated by singlet oxygen. Alkali-labile sites are also produced and have been related to type I reactions (electron transfer reactions). Following the ''photoprobing" technique developed by Barton et al. with photoactivatable ruthenium or rhodium complexes to study the secondary and tertiary structure of nucleic acids [25], Nussbaum et al. used different isomers of H2TMPyP to study branched DNA [26]. The "three-way junction" DNAs are more sensitive to photocleavage by
< previous page
page_402
next page >
< previous page
page_403
next page > Page 403
H2TMPyP than regular DNA duplexes, the main targets being the guanine residues flanking the junction region. Recently, it was shown that phosphorusV tetraphenylporphyrin is also capable of inducing DNA cleavage via a photoreaction involving singlet oxygen and an electron transfer from DNA to the short-lived singlet excited state of the phosphorus [27]. Water-soluble sulfonated porphyrins like H2TPPS or Zn-TPPS are good photosensitizers but very poor DNA cleavers because of their weak interactions with DNA due to their anionic groups at the periphery of the porphyrin ring [21,23]. 2.2 Nuclease Activity of Manganese and Iron Porphyrins Activated by Oxidants The oxidative cleavage of DNA with a cationic water-soluble porphyrin was first demonstrated by Fiel et al. by using the ironIII meso-tetrakis-(4-N-methylpyridiniumyl)porphyrin, Fe-TMPyP (Fig. 1), activated by a reducing agent in the presence of molecular oxygen [28]. Then Dabrowiak et al. [29,30] showed that not only reducing agents like ascorbate or superoxide anion in the presence of oxygen but also oxygen atom donors like iodosylbenzene are able to activate the complex with an efficiency depending on the metal. Several other activating agents were subsequently investigated (Fig. 2): potassium monopersulfate, KHSO5, efficiently activates Mn-TMPyP, allowing one to induce catalytic ss DNA breaks at low concentrations of manganese porphyrin (2200 nM) and KHSO5 (10 µM), and for a short incubation time of 1 min [31,32].
Fig. 2. Structures of the oxygen atom donors PhIO, KHSO5, and MMPP.
< previous page
page_403
next page >
< previous page
page_404
next page > Page 404
The peracid compound magnesium monoperoxophthalate (MMPP) was shown to be efficient on cleavage of supercoiled Φχ174 DNA [33] or CT DNA [34]. Hydrogen peroxide seems to be only a poor cofactor since it activates Mn-TMPyP giving similar DNA cleavage profiles as with KHSO5, but only for concentrations 103104 fold higher [32,35]. Electrochemical methods were also used to activate MnIII and FeIII complexes of TMPyP to cause cleavage of DNA [17]. Anionic porphyrins are usually much less reactive, as has been reported for Fe-TPPS activated with DTT [28]. In the case of hemin (ferric protoporphyrin IX chloride), for which some of its effects on gene expression might be mediated by nicking of DNA, strand scissions require the presence of O2, a reducing agent, and transition metal salts, e.g., CoII, ZnII, NiII, used to favor a ternary association porphyrin/metallic cation/DNA [36,37]. Cleavage experiments with anionic manganese porphyrins activated by KHSO5 failed, even in the presence of polycations [32]. 2.3 Detailed Mechanism on DNA Cleavage by MnIII-TMPyP/KHSO5 2.3.1 Metalloporphyrin Activation and Reactive Species Involved MnIII, FeIII, and CoIII complexes of TMPyP cleaved DNA when activated by thiols or ascorbate in the presence of dioxygen, or by superoxide ion or iodosylbenzene [2830]. The most efficient was iodosylbenzene: since this oxygen atom donor is expected to react directly with the metalloporphyrin in a one-step process to yield a catalytic active species, it is thought, by analogy with cytochrome P-450 activation chemistry, that the reactive species involved in DNA breaks is a high-valent oxo intermediate [Eq. (1)]. Under treatment with ascorbate in the presence of molecular oxygen, reductive activation of O2 by manganese or iron porphyrins yields high-valent metal-oxo species [Eq. (2)]. However, the use of inhibitors and scavengers suggested H2O2 and peroxy radicals as possible intermediates; diffusible hydroxyl radicals can be excluded but hydroxyl radicals produced in close proximity to DNA target might be involved [Eq. (3)] [38]. Other oxygen atom donors were subsequently demonstrated to be
< previous page
page_404
next page >
< previous page
page_405
next page > Page 405
very efficient and easy-to-handle activating agents: both KHSO5 [30,31] and MMPP [33,34] activate Mn-TMPyP and give clean discrete sites of cleavage on DNA. The exact nature of the chemistry of DNA cleavage (see Sec. 2.3.3) supports a cytochrome P-450 route for their mechanism of action, the reactive species being a high-valent metal-oxo complex [Eq. (1)]. In the same conditions, Fe-TMPyP is less active compared to the manganese derivative: such a difference is not fully understood at the present time. With hydrogen peroxide as cofactor, breaks were as well defined as for KHSO5 and probably result from the same chemistry [35], suggesting that H2O2 is able to generate MnV-oxo species via a heterolytic cleavage of the OO bond [Eq. (1)]. However, because the reaction is much less efficient, the question of a homolytic cleavage of H2O2 giving hydroxyl radicals and MIV-OH species with an intrinsic lower reactivity [Eq. (4)] is still open. Electrochemical reduction used to activate MnIII and FeIII complexes of TMPyP in the presence of oxygen was shown to cause a slow cleavage reaction on plasmid DNA, but the mechanism remains unclear [17]. When produced in the proximity of DNA, one (at least) of the reactive species described above attacks deoxyribose moieties and gives oxidative lesions which result in breaks of the sugar phosphate DNA backbone, according to the mechanism detailed below. 2.3.2 Preliminary Results on the Mechanism of DNA Cleavage Gel electrophoresis analyses of DNA binding and cleavage specificity indicated first that cationic manganese or iron porphyrins bind into the minor groove (this class of metalloporphyrins does not intercalate be-
< previous page
page_405
next page >
< previous page
page_406
next page > Page 406
tween DNA base pairs) and cleave at all four possible nucleotide positions of DNA, strongly suggesting the deoxyribose moiety as the primary site of attack [39]. Other preliminary data were obtained by reference to the mechanism of the antitumor antibiotic bleomycin (BLM), this drug being known to cleave DNA by abstracting a hydrogen atom at the C4' position of deoxyribose. Using MnIII, FeIII, or CoIII complexes of H2TMPyP and ascorbate, superoxide ion, or iodosylbenzene as activating agents, Dabrowiak et al. [29] did not succeed in characterization of base propenals, the main characteristic products of DNA damage induced by BLM in the presence of O2. With Mn-TMPyP activated with KHSO5, similar attempts to identify these degradation products failed also, although the base propenals were shown to be stable in the experimental conditions [34]. At least these data suggest that if the hydrogen atom at C4' is the target for the high-valent manganese-oxo species as observed for metallo-BLMmediated DNA breaks, then the oxygen-dependent route is not involved. Alternative targets accessible from the minor groove are the tertiary CH bond at C1' and one of the two secondary CH bonds at C5' (pro-S H5'). The tertiary CH bond at C3' and the other CH bond at C5' (pro-R H5'), only accessible from the major groove, might be discarded. 2.3.3 Molecular Mechanism of DNA Cleavage 2.3.3.1 5-MF and FUR As Reporters of Main Cleavage Pathways First, a simplified model of DNA, polydeoxyadenylic acid or poly(dA), was shown to be readily cleaved at neutral pH by Mn-TMPyP/KHSO5. Free adenine was spontaneously released and, after heating, a rather unstable sugar degradation product identified as 5-methylene-2-furanone (5-MF) was evidenced, supporting a C1' oxidation pathway (Fig. 3, route A) [40]: hydroxylation at C1' induces spontaneous release of free adenine and is followed by two β eliminations which give 5-MF and leave the two DNA strands terminated by phosphate groups. The first β elimination creates a 5'-phosphate end and the second one is accompanied by 5-MF release. Then the study was extended to other polydeoxynucleotides, to ds copolymers, and to CT DNA [34,41]. The main results are as follows: (1) a first amount of free base is released at ambient temperature; (2) after heating, two other products can be detected, identified as 5-MF
< previous page
page_406
next page >
< previous page
page_407
next page > Page 407
Fig. 3. Proposed mechanism for the cleavage of DNA or DNA models after hydroxylation of CH bonds on the 1' and 5' positions of deoxyribose by Mn-TMPyP/KHSO5. (∆, thermal step; B, base; BE, β elimination.) and furfural (FUR) by HPLC; (3) the appearance of FUR correlates with an enhanced release of the corresponding bases during heating and provides evidence for a C5' oxidative pathway (route B in Fig. 3): hydroxylation at C5' gives a C5' aldehyde at one strand terminus and a 3'phosphate at the other at ambient temperature. Upon heating, two β eliminations result in the formation of FUR, free base release, and leave a phosphate at the 5' terminus. Both routes A and B give rise to the same phosphate termini but the sugar degradation products are different and the chronology of events is reversed: in route A, the base is spontaneously released and the DNA break follows; in route B, the strand break occurs first and the base release is only a consequence of the thermal treatment. The two monophosphate esters (at 3' and 5' ends) on both sides of the cleavage site could be characterized by 31P NMR analysis [42]. The FUR/5-MF ratio is indicative of the relative reactivities of the activated Mn-TMPyP complex toward C5' and C1' for various DNA and DNA models. The C1' target is the main hydroxylation site for CT DNA and G·C polymers when selective attack at C5' is observed for A·T polymers, which suggests that the mechanism of DNA cleavage is highly sequence-dependent. Hydroxylation at C1' also confirms that the cationic manganese porphyrin interacts with the minor groove of DNA. Subtle changes in the interaction of the cleaving reagent with
< previous page
page_407
next page >
< previous page
page_408
next page > Page 408
DNA may explain that the reactivity changes from one target to the other: the largely opened single helix of ss polydeoxynucleotides or the widened minor groove of G·C-rich ds polymers allow the more reactive tertiary C1' carbon to be oxidized. For A·T rich ds polymers, the narrow minor groove restricts the access to C1', deeply located in the groove, and then C5', which is located at the entrance of the groove, is preferred (minor groove width is 9 Å in A·T-rich sequences compared with 12 Å for B-DNA or even higher values for G·C polymers (see Ref. 43 and references therein). Until now, oxidative chemistry at C1' and C5' of deoxyribose has been demonstrated only for Mn-TMPyP, enediyne compounds, and the copper phenanthroline complex [44]. H1' abstraction has been described for oxidative degradation of DNA using bis(1,10-phenanthroline)-copperI and H2O2 [45] or enediyne antibiotics like neocarzinostatin in the presence of thiols [46], but it is only in the first case, when no reductants were used as coreactants, that the release of 5-MF could be evidenced. On the other hand, concerning a selective attack at C5', H5' abstraction represents the major mode of DNA cleavage by neocarzinostatin [46], but because of the presence of reducing agents, the end-product FUR could never be observed. 2.3.3.2 Sequence Specificity: (A·T)3 As Minimum Cleavage Site Analysis of the metalloporphyrin-mediated strand scission on a restriction fragment revealed that the minimum porphyrin cleavage site is (A·T)3 [29]. A more detailed cleavage analysis on a series of short ds oligodesoxyribonucleotides (ODNs) containing a trinucleotide sequence having only the bases adenine and thymine, i.e., three continuous A·T bp (named as A·T triplet), evidences the following main features [43,47]: 1. When a selective cleavage occurs, hydroxylation at C5' represents the initial damage on the sugar phosphodiester backbone. No significant oxidative attack at C1' could be detected. Hydroxylation at C5' leaves a 3'-phosphate and a 5'-aldehyde at the ends. As exemplified in Fig. 4 for the analysis of the CAAAGCG fragments resulting from cleavage of the ds heptamer CAAAGCG·CGCTTTG, NaBH4 reduction of the 5'-aldehyde fragment gave the corresponding 5'-OH oligomer, free bases were only released after heating, and FUR (not 5-MF) could be detected as sugar residue. All the reaction products were identified after chemical (NaBH4), biochemical (alkaline phosphatase, nuclease P1), or thermal treatment. The α,β-unsaturated aldehyde intermediate produced by the first β elimination was formally identified during the
< previous page
page_408
next page >
< previous page
page_409
next page > Page 409
Fig. 4. Cleavage of CAAAGCG in ds ODN CAAAGCG·CGCTTTG by Mn-TMPyP/KHSO5: identification of the resulting products. SB, strand break; BE, β elimination; ∆, thermal step; P1, P1 nuclease; AP, alkaline phosphatase; Pi, inorganic phosphate. *This way was shown during a poly(dA)·poly(dT) cleavage study with B = A, T.
< previous page
page_409
next page >
< previous page
page_410
next page > Page 410
course of cleavage reaction on poly(dA)·poly(dT): reduction with NaBH4 (to give the alcohol from the aldehyde) followed by hydrogenation on palladium over charcoal (to reduce the sugar double bond) produced the corresponding dideoxyribonucleosides (the base was A or T), which were identified by reference to authentic samples (Fig. 4 [48]). 2. In all short ds ODNs, the presence of an A·T triplet is necessary to observe a selective cleavage which always occurred at the sugar on the 3' side of this triplet, no matter was the base linked to this sugar. The cleavage site on one strand was always shifted to the 3' end by 4 bp relative to the site of the opposite strand (Fig. 5). Cleavage selectivity described above is due to a tight interaction between the metalloporphyrin and A·T-rich regions of DNA, the A·T triplet having to be considered as the minimal size of its preferred binding site. This selectivity is consistent with the relative affinity of Mn-TMPyP for poly[d(A-T)·d(A-T)] and poly[d(G-C)·d(G-C)] (K = 12 × 104 M1 and 0.2 × 104 M1, respectively [49]) and can be explained by electrostatic interactions of the cationic porphyrin with the more negative potential in the minor groove of A·Trich polymers compared to G·C polymers [43]. When the A·T sequence is longer than a triplet, cleavage not only affects adjacent G·C or C·G base pairs but also base pairs inside the A·T-rich sequence. As shown in Fig. 5, a duplex which contains two independent triplets gives two independent cleavage sites on each strand, and a duplex which contains four, as example, overlapping triplets gives four consecutive cleavage sites [35].
Fig. 5. Selectivity of cleavage of short ODN duplexes by Mn-TMPyP/KHSO5 depends on the presence and the location of an A·T triplet. Arrows indicate the sites of selective cleavage on the 3' side of the A·T triplets.
< previous page
page_410
next page >
< previous page
page_411
next page > Page 411
3. During PAGE analysis of cleaved 5'-32P-end-labeled ODN, the 5'-end-labeled fragments were not sensitive to thermal or various chemical treatments, in agreement with the presence of 3'-phosphate termini, and so migrated according to the corresponding fragments obtained by Maxam-Gilbert sequencing [35,43,50]. On the opposite, 3'-32P-end-labeled fragments with a 5'-aldehyde terminus (CHO) migrated differently after reduction (CH2OH), oxidation (COOH), heating (5'-P terminus and release of free base and FUR), further heating and alkaline phosphatase treatment (5'-OH terminus and loss of free base, FUR, and inorganic phosphate). Oxidation of the 5'-aldehyde to a carboxylic acid may occur in small amounts under drastic cleavage conditions, probably resulting from two consecutive attacks of the same C5' sugar position by the activated Mn-TMPyP. The initial DNA cleavage band (aldehyde terminus) always migrated slowly: for short fragments, the difference may be of several nucleotides; for longer ones, the difference is about two nucleotides. Using Maxam-Gilbert sequencing bands as references, the exact location of the cleavage site is only observed after a heating step: in the absence of thermal treatment, the exact position of breaks has to be shifted by about two nucleotides to the 3' end [35]. 2.3.3.3 About the Nature of Activated Mn-TMPyP Species The quasiabsence of additional weaker bands close to the main fragments in polyacrylamide gel electrophoresis (PAGE) analysis strongly supports a nondiffusible active species, namely, a high-valent manganese-oxoporphyrin complex similar to the iron-oxo intermediate involved in cytochrome P-450 chemistry. A recent investigation using Mn-TMPyP/KHSO5 in aqueous solution to oxidize carbamazepine, an analgesic and anticonvulsant drug, shows that through a ''redox tautomerism" mechanism [Eq. (5)] involving a coordinated water molecule on the metallo-
porphyrin catalyst, the oxidizing entity can be localized on one or the other face of the activated metalloporphyrin [51]: so half of the oxygen atom incorporated in the substrate comes from the solvent and half from KHSO5. Very recently we reported that oxidation at C1' of DNA deoxyribose results from such a mechanism, supporting an Mn-oxo as
< previous page
page_411
next page >
< previous page
page_412
next page > Page 412
the reactive species involved in a cytochrome P-450-type hydroxylation reaction during DNA cleavage [52]. 2.4 "Pseudohydrolysis" of DNA Among cleaving agents available as chemical nucleases, no real hydrolytic reagent exists. Because the cleavage of the phosphate-sugar backbone of DNA is usually mediated by oxidation of sugars, the generated fragments are ending with "nonnatural" termini, not suitable for further ligation. Interestingly, in A·T rich sequences of DNA, Mn-TMPyP/KHSO5 is cleaving DNA by hydroxylating 5'-carbons of deoxyribose units, leading to DNA fragments ending with a 5'-aldehyde nucleoside and a 3'-phosphate (Fig. 6). This oxidative cleavage can be easily reverted to the equivalent of a hydrolytic cleavage by reduction of the 5'-aldehyde back to the primary 5'-alcohol. Furthermore, this cleavage occurs on both 3'-sides of A·T triplets and is reminiscent of ds breaks induced by restriction enzymes. The generated ds DNA fragments have 3' protruding ends overlapping on three A·T base pairs (Fig. 5). The present pseudohydrolysis of DNA is only different from what natural restriction enzymes are doing by the nature of the cleaved CO bond when restriction enzymes leave a 5'-phosphate group, MnTMPyP/KHSO5 releases a 3'-phosphate ending DNA fragment. To our knowledge, no ligase able to ligate 5'-OH with 3'-phosphate ending fragments is available but DNA chemical ligation methods can be used [47]. On a simplified DNA substrate, a 35-mer duplex containing one
Fig. 6. "Pseudohydrolysis" of the phosphodiester bond in DNA mediated in two steps by the Mn-TMPyP/KHSO5 system and NaBH4. (A negative charge has been omitted for clarity on phosphodiesters and terminal phosphates.)
< previous page
page_412
next page >
< previous page
page_413
next page > Page 413
A·T triplet, we showed that the Mn-TMPyP/KHSO5 system associated to a further reduction step could be effectively used as a model of restriction enzyme. After the pseudohydrolysis step (Fig. 6), the cleaved fragments were religated by chemical ligation. To ensure that the chemically cleaved and religated strands were intact, the reformed DNA duplex was tested as a substrate for a natural restriction enzyme. No chemical alterations of both ODN strands could be detected [47]. Provided this remarkable pseudohydrolysis reactivity could be addressed to one selected A·T triplet in a 103106 bp DNA molecule, one would expect to have in hand a practical-use artificial restriction enzyme system. For an example of targeting the activity of Mn-TMPyP by an oligonucleotide, see Sec. 3.2. 3 Selective DNA Cleavage by Vectorized Metalloporphyrins As pioneered by Dervan et al. there is the possibility of targeting the activity of a DNA cleaver by attaching it to different vectors (for recent articles on selective DNA cleavers based on iron-EDTA or copperphenanthroline, see [5357] and [5861], respectively). In the present chapter, we summarize what has been done by linking porphyrin derivatives to two different type of vectors: intercalators and oligonucleotides. 3.1 Metalloporphyrin Intercalators The covalent attachment of a chelated redox active transition metal to an intercalating agent is a way to mimic BLM, the paradigm of DNA cleavers. This antitumoral antibiotic is able to create ss and ds breaks on duplex DNA in association with three cofactors: iron or copper ions, molecular oxygen, and an electron source [6265]. BLM has two distinct structural domains. One is responsible for DNA binding and the other includes five nitrogen atoms able to strongly chelate different metal ions, i.e., iron, manganese, cobalt, or copper. So a good BLM model
< previous page
page_413
next page >
< previous page
page_414
next page > Page 414
should (1) have a noticeable affinity for DNA, (2) be activated either by molecular oxygen and a reducing agent or by a single-oxygen-atom donor like potassium monopersulfate, (3) be able to cleave DNA in vitro at low concentration, and (4) be cytotoxic against tumor cells in vitro at concentrations below 1 or 2 µM (IC50 of BLM against murine leukemia cells L1210: 1.7 µM). The first BLM model was prepared by Lown et al. by attaching an iron(III) deuteroporphyrin IX to 9-chloroacridine via a polyamine linker [66]. Despite the poor water solubility of these BLM models, these molecules cleave ds DNA at 37°C in the presence of molecular oxygen and DTT at a porphyrin/bp ratio of 1:10. Only 5% of the DNA target is cleaved and no base-propenal derivatives were detected as possible DNA degradation products. The DNA cleavage is partially inhibited by catalase and superoxide dismutase, suggesting that these BLM models are producing reduced oxygen species (superoxide, hydrogen peroxide, hydroxyl radicals) rather than metal-oxo entities. These BLM models are cytotoxic against L1210 cells in culture (IC50≈ 1 µM) [66]. Hashimoto et al. used a potent mutagen resulting from glutamic acid pyrolysis as intercalator to be linked either to a derivative of ironIII [tris-o-tolyl(p-aminophenyl)]porphyrin, to a hydrophobic macrocycle complex or to iron(III)protoporphyrin IX [67]. In this latter case, two intercalators were linked to the two carboxylic acid functions of the porphyrin ring. With sodium dithionite as reducing agent, 5060% of supercoiled DNA was cleaved by these molecules at a base/porphyrin ratio of 2:1 and the analysis of DNA fragments indicates that the 5' ends have a phosphate group, while the 3' ends bear a hydroxy group or a nondefined sugar residue [67]. Since our group was involved in oxidative reactions [68,69] and especially in oxidative DNA cleavage catalyzed by metalloporphyrins (see Sec. 2), as well as in studies on cytotoxic ellipticine derivatives [70], we have synthetized hybrid ''metalloporphyrin-ellipticine" molecules. The first "metalloporphyrin-ellipticine" molecules were based on a tris(p-tolyl)porphyrin moiety. These molecules appeared as poor DNA cleavers and did not present cytotoxicity against L1210 cells in culture [71]. In order to increase the water solubility and DNA affinity of these hybrid compounds, we then used a cationic tris(methylpyridiniumyl)-porphyrin motif linked to the ellipticine intercalator (see compound 1 in Fig. 7) [72,73]. The affinity for a DNA duplex depends on the nature and length of the linker and ranges from 2.9 × 108 to 8.2 × 109 M1 for
< previous page
page_414
next page >
< previous page
page_415
next page > Page 415
Fig. 7. Examples of vectorized metalloporphyrins prepared by covalent attachment to an intercalating agent (1) or to an ODN in antisense (2) or antigene (3) strategy.
< previous page
page_415
next page >
< previous page
page_416
next page > Page 416
poly[d(A-T)] and from 1.0 to 7.0 × 108 M1 for poly[d(G-C)] [73]. By monitoring the fast degradation of the nonbonded metalloporphyrin derivative compared to the DNA-bonded molecules, it has been possible to access the intrinsic DNA affinity of the metalloporphyrin moiety of the hybrid metalloporphyrin-ellipticine molecules [49]: 12 × 104 M1 for poly[d(A-T)] and 0.2 × 104 M1 for poly[d(G-C)], confirming that the cationic metalloporphyrin entities still behave as a DNA groove binder and have a higher affinity for an A·T-rich region than for a G·C-rich region. These hybrid molecules are cytotoxic against L1210 cells in culture. The IC50 values are ranging from 0.7 to 1.4 µM for manganese derivatives and 1.6 µM for the corresponding iron complex. On the same cell line, the IC50 value of BLM is only 1.7 µM [73]. The corresponding zinc or nickel hybrid molecules are not cytotoxic. The manganese derivatives activated by potassium monopersulfate are also able in vitro to cleave supercoiled Φχ 174 DNA at low porphyrin/bp ratio (1:124) [74]. The formation of 5-MF as sugar degradation product strongly suggests that the metalloporphyrin is interacting within the minor groove of DNA [34]. Compared to manganese derivatives, the corresponding iron complexes are not as efficient when activated by potassium monopersulfate in vitro [74]. However, they are better when activated by oxygen and a reducing agent, which is the expected mode of activation after penetration of such hybrid molecules within a cell. These hybrid molecules should be considered as potential antitumor agents and, in addition, are able to inhibit the cytopathicity of HIV-1 virus in MT-4 cells at concentrations ranging from 1.4 to 17 µg/mL, i.e., at a concentration that is 2.5- to 30-fold below the cytotoxicity threshold [75]. Photoactivatable zinc porphyrin-ellipticine derivatives have also been studied [76]. Both fluorescence yield and production of singlet oxygen are lower compared to those of Zn-TMPyP, the parent porphyrin molecule, due to the quenching of the singlet state of the porphyrin moiety by the attached ellipticine. Interestingly, these photochemical properties are dramatically enhanced in the presence of DNA as a result of a conformational change. The ellipticine moiety intercalates in ds DNA and so cannot quench the zinc-porphyrin singlet state. On irradiation at 436 nm, the photocleavage of supercoiled Φχ 174 DNA is 50-fold greater than that of HpD [77]. At these wavelengths, singlet oxygen seems to be the main species responsible for DNA cleavage.
< previous page
page_416
next page >
< previous page
page_417
next page > Page 417
Since these Zn-porphyrin derivatives are only photoactivatable in the presence of DNA, they can be considered as ''molecular light switches." The concept of hybrid porphyrin-vector molecules has been extented to pyropheophorbide-polyamine molecules [78]. These hybrid molecules are able to cleave DNA by an electron transfer when activated at 690 nm, a suitable wavelength in phototherapy. Other examples of hybrid molecules were recently prepared: porphyrin-nucleoside [79], porphyrin-netropsin [80], porphyrin-acridione, porphyrin-chlorambucil [81], and porphyrin-dextran [82]. These three latter hybrid molecules are able to cleave DNA by activation with visible light. 3.2 Metalloporphyrin Oligonucleotides It is well established that ODNs can specifically recognize their complementary sequence and can function as DNA probes. Several examples of sequence-selective cleavage of DNA by cleaving agents covalently attached to ODNs have been described in the literature [83,84]. In this section, the results of selective DNA cleavage by metalloporphyrins covalently linked to ODNs will be described. 3.2.1 Hydrophobic or Anionic Metalloporphyrins Linked to Oligonucleotides First examples of hybrid molecules synthesized for sequence-selective cleavage of DNA based on metalloporphyrinsubstituted ODNs were prepared with hydrophobic derivatives of methylpyrroporphyrin XXI tethered to (dT)7 ODNs at a β-pyrrolic position of the porphyrin ring [85,86]. The corresponding ironIII compound activated by H2O2 was the most reactive. Damage induced to the ss DNA target included direct breaks and crosslinked products that were partially transformed to DNA breaks after piperidine treatment. Damage appeared to cover 10 bases around the position of the reactive moiety on the target sequence and only limited yields of target cleavage were reached (1020%). A second generation of negatively charged metalloporphyrin-substituted ODNs was synthesized and tested for DNA cleavage by Russian researchers [8789]. These hybrid molecules consist of ironIII-
< previous page
page_417
next page >
< previous page
page_418
next page > Page 418
hemin or deuterohemin derivatives attached to ODNs through a short linker and were activated for cleavage with H2O2 or reductants. Also described is a palladium-coproporphyrin I conjugate [87] that by photochemical activation led to 30% of DNA cleavage over 10 bases on the target sequence. No direct DNA breaks were observed, but alkalilabile lesions and alkali-labile crosslinks are responsible for DNA cleavage mediated by this compound [87]. IronIII hemin or deuterohemin conjugates show unexpected selective cleavage (at guanine positions in the close vicinity of the metalloporphyrin location on the target sequence) [88,89]. This guanine-selective DNA cleavage is due to complete transformation of crosslinked products upon piperidine treatment. About 3050% of cleavage of ss DNA target is achieved. A more efficient hybrid molecule is ironIII meso-tetra(4-carboxyphenyl)porphyrin when coupled to an ODN by the mean of a three-methylene amino linker; in the presence of a reductant, 60% of direct DNA cleavage is mediated at very low concentration and for a reagent/target ratio of 1:1 [90]. The mechanism of DNA damage may involve either diffusing oxygenated reactive species and then the damage spans over several bases or a nondiffusible oxidative species of the metalloporphyrin. Oxidation of bases or sugars can lead to direct breaks, to alkali-labile lesions, and to the formation of crosslinked products that can be converted (partially or totally) to DNA breaks after alkali treatment. 3.2.2 Cationic Metalloporphyrins Linked to Oligonucleotides Differing from hydrophobic or anionic metalloporphyrins, cationic metalloporphyrins show a high affinity for nucleic acids. Thus they appear as reasonable candidates to be coupled to ODNs. The metalloporphyrin precursor (Fig. 7) is covalently attached to the 5' end of the ODN via a tether of variable nature and length [91]. The porphyrin moiety is carrying only three cationic pyridiniumyl groups instead of four in the parent compound (Mn-TMPyP). Two ODN sequences have been chosen as vectors; one (ODN1) is complementary to the initiation codon region of the tat gene of HIV-1 genome, and the other (ODN2) is designed to target one triple-helix possible site in the rev and pol genes of HIV-1 (compounds 2 and 3, respectively, in Fig. 7).
< previous page
page_418
next page >
< previous page
page_419
next page > Page 419
3.2.2.1 Selective ss DNA Cleavage (Antisense Strategy) The ODN1-manganese porphyrin hybrid was shown to recognize and cut selectively the ss target sequence, even in the presence of a large excess of random DNA (either ss or ds DNA). The cleavage of the 35-mer target labeled on the 5' end by 32P occurs at the predicted zone, in the vicinity of the metalloporphyrin site of linkage [92]. While ''free" metalloporphyrin Mn-TMPyP was reaching the labeled 35-mer target only at 1 µM concentration, the vectorized metalloporphyrin was cleaving it at 10 nM, which corresponds, under experimental conditions, to a ratio of 3 molecules of cleaver per molecule of target. Half of the target was cleaved. Since the cleavage pattern shows an intense smear near the junction of ss and ds regions of the duplex, the mechanism in this case must not be restricted to the specific 5' attack previously observed for the "free" Mn-TMPyP. In additional experiments with A·T triplets located in the vicinity of the anchorage site of the cleaving reagent, the increase of the linker length was shown to give more freedom to the cleaver moiety, allowing it to interact with the ds ODN in a way that is reminiscent of the free compound but with lower efficiency; concomitant smears were still observed on the upper ss part of the target DNA [93]. Compared with the results of other types of metalloporphyrin-ODNs conjugates, the better cleaving reactions were performed when very short linkers were used [49,86,88,89]. Even if interactions with DNA prove to be different in the case of hydrophobic, anionic, or cationic porphyrins, all of these cleaving species are prone to self-degradation if the interaction (and reaction) with target DNA is not optimized [49,89]. So a careful design of the tether of the conjugate is required depending on the cleaving agent and on what is known about its interaction with the target. In conclusion, without question the attachment of a cationic metalloporphyrin to an ODN vector allowed selective cleavage of the complementary sequence of DNA. However, a real high efficiency of cleavage (100% of degradation of target or catalytic activity) as well as site selectivity of cleavage inside the target sequence is still not reached with this cleaver or any other. 3.2.2.2 Selective ds DNA Cleavage (Triple-Helix Strategy) Covalent attachment of a cationic metalloporphyrin to a triple-helix forming ODN allows one to adress the cleaving reactivity directly onto ds DNA (for a review of the triple-helix strategy, see [94]). We chose ODN2 as
< previous page
page_419
next page >
< previous page
page_420
next page > Page 420
vector and attached the cationic Mn-porphyrin to its 5' end by the means of two types of linker (Fig. 7). When the spacer was a hydrophobic alkyldiamine, the yield of cleavage of the target duplex in experiments carried out in the presence of KHSO5 was very low. But when the cleaving reagent was attached to ODN2 through a spermine linker (compound 3, Fig. 7 [95]), the yield of cleavage reached 70%. Spermine is known to increase the stability of triplex structure [96]. In the present case, the Tm of the 5'-spermine conjugate is 42°C [95] compared to 33°C for the alkylamine one. This high value of Tm associated to the high efficiency of ds DNA cleavage under such ''physiological" conditions (low salt concentration, 37°C) make these conjugates good candidates for the antigene strategy [94]. 4 Concluding Remarks: Is It Possible to Go from Oxidative DNA Cleavage to Drugs Based on Metalloporphyrins? The use of BLM as anticancer drug reminds us that oxidative DNA cleavage is a reasonable strategy of designing new drugs active in the treatment of cancers. Metalloporphyrin derivatives are versatile DNA cleavers with a nuclease activity which can be modulated by changing the central metal or modifying the macrocyclic ligand. The mechanism of action of these DNA cleavers is rather well understood at the present time and the question about the possibility of their use as active entities in the design of new antitumor or antiviral agents is open. One key point in the development of such molecules is the targeting of their nuclease activity. In the case of anticancer drugs, it is not an easy task to design an efficient vector to target the cytotoxic and nuclease activity of cationic manganese porphyrins to tumor cells. Such a step is still the weak point in the so-called rational design of new antitumor agents. When ODNs are used as vectors to target the nuclease activity of these DNA or RNA cleavers to viral DNA or viral mRNA, it is still very difficult to overcome the key problems of cell membrane crossing and rapid diffusion of modified ODNs to their targets. If such difficulties can be solved in the future, then we shall remember that metalloporphyrins are efficient DNA cleavers.
< previous page
page_420
next page >
< previous page
page_421
next page > Page 421
Acknowledgments The authors are grateful for the contributions of coworkers whose names appear in the reference list. Financial support was provided by CNRS, ARC, ANRS, GENSET, and Region Midi-Pyrénées. Abbreviations
BLM
bleomycin
CT DNA
calf thymus DNA
DTT
dithiothreitol
ds
double strand
EDTA
ethylenediamine-N,N,N',N'-tetraacetic acid
FUR
furfural
HIV
human immunodeficiency virus
HpD
hematoporphyrin derivative
HPLC
high-performance liquid chromatography
IC50
50% inhibiting concentration
5-MF
5-methylene-2-furanone
MMPP
magnesium monoperoxophthalate
ODN
oligodesoxyribonucleotide
PAGE
polyacrylamide gel electrophoresis
PhIO
iodosylbenzene
ss
single strand
Tm
thermal denaturation
TMPyP
meso-tetrakis(4-N-methylpyridiniumyl)porphyrin
TPPS
tetrasodium meso-tetrakis(p-sulfonatophenyl)porphyrin
References 1. D. Dolphin, Can. J. Chem., 72, 1005 (1994). 2. Y.-K. Ho, J. R. Missert, and T. J. Dougherty, Photochem. Photobiol., 54, 83 (1991). 3. C. J. Gomer, N. Rucker, and A. L. Murphree, Cancer Res., 48, 4539 (1988).
< previous page
page_421
next page >
< previous page
page_422
next page > Page 422
4. G. Bock and S. Harnett, eds., Photosensitizing Compounds: Their Chemistry, Biology and Clinical Use (Ciba Foundation Symposium 146), John Wiley and Sons, Chichester, 1989. 5. N. L. Oleinick, A. R. Antunez, M. E. Clay, B. D. Rihter, and M. E. Kenney, Photochem. Photobiol., 57, 242 (1993). 6. D. A. Bellnier, B. W. Henderson, R. K. Pandey, W. R. Potter, and T. J. Dougherty, J. Photochem. Photobiol. B: Biol., 20, 55 (1993). 7. E. Kvam and J. Moan, Photochem. Photobiol., 52, 769 (1990). 8. R. J. Fiel, J. Biomol. Struct. Dynam., 6, 1259 (1989). 9. E. J. Gibbs, M. C. Maurer, J. H. Zhang, W. M. Reiff, D. T. Hill, M. Malicka-Blaszkiewicz, R. E. McKinnie, H.-Q. Liu, and R. F. Pasternack, J. Inorg. Biochem., 32, 39 (1988). 10. L. G. Marzilli, New J. Chem., 14, 409 (1990). 11. M. A. Sari, J. P. Battioni, D. Dupré, D. Mansuy, and J. B. Le Pecq, Biochemistry, 29, 4205 (1990). 12. L. G. Marzilli, G. Pethö, M. Lin, M. S. Kim, and D. W. Dixon, J. Am. Chem. Soc., 114, 7575 (1992). 13. R. F. Pasternack, C. Bustamante, P. J. Collings, A. Giannetto, and E. J. Gibbs, J. Am. Chem. Soc., 115, 5393 (1993). 14. U. Sehlstedt, S. K. Kim, P. Carter, J. Goodisman, J. F. Vollano, B. Nordén, and J. C. Dabrowiak, Biochemistry, 33, 417 (1994). 15. X. Hui, N. Gresh, and B. Pullman, Nucl. Acids Res., 18, 1109 (1990). 16. R. F. Pasternack, E. J. Gibbs, and J. J. Villafranca, Biochemistry, 22, 5409 (1983). 17. M. Rodriguez, T. Kodadek, M. Torres, and A. Bard, Bioconj. Chem., 1, 123 (1990). 18. R. F. Pasternack, E. J. Gibbs, and J. J. Villafranca, Biochemistry, 22, 2406 (1983). 19. N. Foster, A. K. Singhal, M. W. Smith, N. G. Marcos, and K. J. Schray, Biochim. Biophys. Acta, 950, 118 (1988). 20. K. Kalyanasundaram, Inorg. Chem., 23, 2453 (1984). 21. R. J. Fiel, N. Datta-Gupta, E. H. Mark, and J. C. Howard, Cancer Res., 41, 3543 (1981). 22. J. M. Kelly, M. J. Murphy, D. J. McConnell, and C. OhUigin, Nucl. Acids Res., 13, 167 (1985).
< previous page
page_422
next page >
< previous page
page_423
next page > Page 423
23. D. Praseuth, A. Gaudemer, J.-B. Verlhac, I. Kraljic, I. Sissoëff, and E. Guillé, Photochem., Photobiol., 44, 717 (1986). 24. B. R. Munson and R. J. Fiel, Nucl. Acid Res., 20, 1315 (1992). 25. A. M. Pyle and J. K. Barton, Prog. Inorg. Chem., 38, 413 (1990). 26. J. M. Nussbaum, M. E. A. Newport, M. Mackie, and N. B. Leontis, Photochem., Photobiol., 59, 515 (1994). 27. T. Saito, M. Kitamura, M. Tanaka, M. Morimoto, H. Segawa, and T. Shimidzu, Nucleosides and Nucleotides, 13, 1607 (1994). 28. R. J. Fiel, T. A. Beerman, E. H. Mark, and N. Datta-Gupta, Biochem. Biophys. Res. Commun., 107, 1067 (1982). 29. B. Ward, A. Skorobogaty, and J. C. Dabrowiak, Biochemistry, 25, 6875 (1986). 30. J. C. Dabrowiak, B. Ward, and J. Goodisman, Biochemistry, 28, 3314 (1989). 31. E. Fouquet, G. Pratviel, J. Bernadou, and B. Meunier, J. Chem. Soc., Chem. Commun., 1169 (1987). 32. J. Bernadou, G. Pratviel, F. Bennis, M. Girardet, and B. Meunier, Biochemistry, 28, 7268 (1989). 33. G. Pratviel, J. Bernadou, M. Ricci, and B. Meunier, Biochem. Biophys. Res. Commun., 160, 1212 (1989). 34. G. Pratviel, M. Pitié, J. Bernadou, and B. Meunier, Nucl. Acids Res., 19, 6283 (1991). 35. M. Pitié, G. Pratviel, J. Bernadou, and B. Meunier in The Activation of Dioxygen and Homogeneous Catalytic Oxidation (D. H. R. Barton, A. E. Martell, and D. T. Sawyer, eds.), Plenum Press, New York, 1993, pp. 333346. 36. R. L. Aft and G. C. Mueller, J. Biol. Chem., 258, 12069 (1983). 37. H. Sakurai, M. Shibuya, C. Shimizu, S. Akimoto, M. Maeda, and K. Kawasaki, Biochem. Biophys. Res. Commun., 136, 645 (1986). 38. R. W. Byrnes, R. J. Fiel, and N. Datta-Gupta, Chem.-Biol. Interact., 67, 225 (1988). 39. B. Ward, A. Skorobogaty, and J. C. Dabrowiak, Biochemistry, 25, 7827 (1986). 40. Bernadou, B. Lauretta, G. Pratviel, and B. Meunier, C. R. Acad. Sci. Paris, 309 III, 409 (1989). 41. G. Pratviel, M. Pitié, J. Bernadou, and B. Meunier, Angew. Chem. Int. Ed. Engl., 30, 702 (1991).
< previous page
page_423
next page >
< previous page
page_424
next page > Page 424
42. G. Gasmi, M. Pasdeloup, G. Pratviel, M. Pitié, J. Bernadou, and B. Meunier, Nucl. Acids Res., 19, 2835 (1991). 43. M. Pitié, G. Pratviel, J. Bernadou, and B. Meunier, Proc. Natl. Acad. Sci. USA, 89, 3967 (1992). 44. G. Pratviel, J. Bernadou, and B. Meunier, Angew. Chem. Int. Ed. Engl., 34, 746 (1995). 45. T. E. Goyne and D. S. Sigman, J. Am. Chem. Soc., 109, 2846 (1987). 46. I. H. Goldberg, Acc. Chem. Res., 24, 191 (1991). 47. G. Pratviel, V. Duarte, J. Bernadou, and B. Meunier, J. Am. Chem. Soc., 115, 7939 (1993). 48. G. Pratviel, M. Pitié, C. Périgaud, G. Gosselin, J. Bernadou, and B. Meunier, J. Chem. Soc. Chem. Commun., 149 (1993). 49. L. Ding, J. Bernadou, and B. Meunier, Bioconj. Chem., 2, 201 (1991). 50. R. B. Van Atta, J. Bernadou, B. Meunier, and S. M. Hecht, Biochemistry, 29, 4783 (1990). 51. J. Bernadou, A.-S. Fabiano, A. Robert, and B. Meunier, J. Am. Chem. Soc., 116, 9375 (1994). 52. M. Pitié, J. Bernadou, and B. Meunier, J. Am. Chem. Soc., 117, 2935 (1995). 53. R. P. Herzberg and P. B. Dervan, J. Am. Chem. Soc., 104, 313 (1982). 54. A. S. Boutorin, V. V. Vlassov, S. A. Kazakov, I. V. Kutiavin, and M. A. Podyminogin, FEBS Lett., 172, 43 (1984). 55. R. S. Youngquist, and P. B. Dervan, Proc. Natl. Acad. Sci. USA, 82, 2565 (1985). 56. H. E. Moser and P. B. Dervan, Science, 238, 645 (1987). 57. H. Han, A. Shepartz, M. Pelligrini, and P. B. Dervan, Biochemistry, 33, 9831 (1994). 58. D. S. Sigman, D. R. Graham, V. D'Aurora, and A. M. Stern, J. Biol. Chem., 254, 12269 (1979). 59. D. S. Sigman, T. W. Bruice, A. Mazumder, and C. L. Sutton, Acc. Chem. Res., 26, 98 (1993). 60. P. S. Pendergast, Y. W. Ebright, and R. H. Ebright, Science, 265, 959 (1994). 61. J. F. Christophe, T. Saison-Behmoaras, C. Barbier, M. Chassignol, N. T. Thuong, and C. Hélène, Proc. Natl. Acad. Sci. USA, 86, 9702 (1989).
< previous page
page_424
next page >
< previous page
page_425
next page > Page 425
62. S. M. Hecht, Acc. Chem. Res., 19, 383 (1986). 63. J. Stubbe, and J. Kozarich, Chem. Rev., 87, 1107 (1987). 64. J. W. Sam, X. J. Tang, and J. Peisach, J. Am. Chem. Soc., 116, 5250 (1994). 65. G. Pratviel, J. Bernadou, and B. Meunier, Biochem. Pharmacol., 38, 133 (1989). 66. J. W. Lown, S. M. Sondhi, C.-W. Ong, A. Skorobogaty, H. Kishikawa, and J. C. Dabrowiak, Biochemistry, 25, 5111 (1986). 67. Y. Hashimoto, H. Iijima, Y. Nozaki, and K. Shudo, Biochemistry, 25, 5103 (1986). 68. B. Meunier, E. Guilmet, M.-E. De Carvalho, and R. Poilblanc, J. Am. Chem. Soc., 106, 6668 (1984). 69. B. Meunier, Chem. Rev., 92, 1411 (1992). 70. G. Meunier, D. de Montauzon, J. Bernadou, G. Grassy, M. Bonnafous, S. Cros, and B. Meunier, Mol. Pharmacol., 33, 93 (1988). 71. G. Etemad-Moghadam, L. Ding, F. Tadj, and B. Meunier, Tetrahedron, 45, 2641 (1989). 72. L. Ding, G. Etemad-Moghadam, S. Cros, C. Auclair, and B. Meunier, J. Chem. Soc., Chem. Commun., 1711 (1989). 73. L. Ding, G. Etemad-Moghadam, S. Cros, C. Auclair, and B. Meunier, J. Med. Chem., 34, 900 (1991). 74. L. Ding, G. Etemad-Moghadam, and B. Meunier, Biochemistry, 29, 7868 (1990). 75. L. Ding, J. Balzarini, D. Schols, B. Meunier, and E. De Clercq, Biochem. Pharmacol., 44, 1675 (1992). 76. S. J. Milder, L. Ding, G. Etemad-Moghadam, B. Meunier, and N. Paillous, J. Chem. Soc., Chem. Commun. 1131 (1990). 77. C. Sentagne, B. Meunier, and N. Paillous, J. Photochem. Photobiol. B: Biol., 16, 47 (1992). 78. S. Mansouri, A. Gossauer, B. Meunier, and N. Paillous, New J. Chem., 18, 745 (1994). 79. H. Li, and L. Czuchajowski, Tetrahedron Lett., 35, 1629 (1994). 80. G. Anneheim-Herbelin, M. Perrée-Fauvet, A. Gaudemer, P. Helissey, S. Giorgi-Renault, and N. Gresh, Tetrahedron Lett., 34, 7263 (1993).
< previous page
page_425
next page >
< previous page
page_426
next page > Page 426
81. G. Mehta, T. Sambaiah, B. G. Maiya, M. Sirish, and A. Dattagupta, Tetrahedron Lett., 35, 4201 (1994). 82. O. Nakajima, H. Mizoguchi, Y. Hashimoto, and S. Iwasaki, J. Am. Chem. Soc., 114, 9203 (1992). 83. P. B. Dervan, Nature, 359, 87 (1992). 84. D. S. Sigman, C.-h. B. Chen, and M. B. Gorin, Nature, 363, 474 (1993). 85. T. Le Doan, L. Perrouault, C. Hélène, M. Chassignol, and N. T. Thuong, Biochemistry, 25, 6736 (1986). 86. T. Le Doan, L. Perrouault, M. Chassignol, N. T. Thuong, and C. Hélène, Nucl. Acids Res., 15, 8643 (1987). 87. O. S. Fedorova, A. P. Savitskii, K. G. Shoikhet, and G. V. Ponomarev, FEBS Lett., 259, 335 (1990). 88. E. I. Frolova, E. M. Ivanova, V. F. Zarytova, T. V. Abramova, and V. V. Vlassov, FEBS Lett., 269, 101 (1990). 89. E. I. Frolova, O. S. Fedorova, and D. G. Knorre, Biochimie, 75, 5 (1993). 90. J. F. Ramalho Ortigao, A. Rück, K. C. Gupta, R. Rösch, R. Steiner, and H. Seliger, Biochimie, 75, 29 (1993). 91. C. Casas, C. J. Lacey, and B. Meunier, Bioconj. Chem., 4, 366 (1993). 92. M. Pitié, C. Casas, C. J. Lacey, G. Pratviel, J. Bernadou, and B. Meunier, Angew. Chem. Int. Ed. Engl., 32, 557 (1993). 93. G. Pratviel, P. Bigey, J. Bernadou, and B. Meunier, in Metal and Genetics (B. Sarkar, ed.), Marcel Dekker, New York, 1995, pp. 153171. 94. N. T. Thuong, and C. Hélène, Angew. Chem. Int. Ed. Engl., 32, 666 (1993). 95. P. Bigey, G. Pratviel, and B. Meunier, J. Chem. Soc. Chem. Commun., 181 (1995). 96. C.-H. Tung, K. J. Breslauer, and S. Stein, Nucl. Acids Res., 21, 5489 (1993).
< previous page
page_426
next page >
< previous page
page_427
next page > Page 427
14 Synthetic Metallopeptides As Probes of Protein-DNA Interactions Eric C. Long, Paula Denney Eason, and Qi Liang Department of Chemistry, Indiana University-Purdue University Indianapolis, 402 N. Blackford St., Indianapolis, IN 46202-3274, USA
428
1. Introduction
2. Oligopeptides Containing a Pendant Metal-Binding Domain
429
2.1. DNA Binding Domains of Proteins
2.1.1. Unnatural or Derivatized Amino Acids
2.1.2. Naturally Occurring Amino Acids and Sequences
430
432
439
2.2. Minor Groove-Binding Oligopeptides
2.3. Other Metal-Binding Oligopeptide Systems
440
443
3. Low Molecular Weight Metallopeptides
444
3.1. Metallo-Gly-Gly-His
3.2. Orientation of Metallopeptides on the DNA Helix
446
447
4. Conclusions and Future Prospects
448
Abbreviations
< previous page
429
page_427
next page >
< previous page
page_428
next page > Page 428 449
References
1 Introduction Oligopeptides with the ability to bind reactive transition metal ions are finding interesting applications in the study of protein-nucleic acid interactions and the exploration of DNA molecular recognition. The utility of metallopeptides, both synthetic and natural, is derived in part from several features including their relative ease of synthesis and their substantially smaller size in comparison to an entire protein (which facilitates experimental investigation). Indeed, metallopeptides can, like other intensely studied metal complexes that bind nucleic acids [13], make use of the ability of a metal ion to impart structure and reactivity to an otherwise unreactive organic ligand. Presently, while difficult to exactly categorize given the increasing number of examples appearing in the literature, most metallopeptides used in the study of nucleic acid-ligand interactions can be classified into two general forms: (1) those that employ metal ions as a ''pendant" segment to impart redox activity to a preexisting DNA binding ligand or recognition motif, and (2) low molecular weight peptides that bind transition metal ions and, as a stand-alone species, have the ability to bind/modify a nucleic acid substrate. The two classifications listed above are thus distinguished by the fundamental role(s) played by the metal ion involved. In the first classification, the metal chelating moiety does not contribute to the binding interaction between the oligopeptide and the DNA substrate. Thus, the reactive properties of metal ions are exploited simply to convert a portion of a protein (e.g., a DNA binding domain) or oligopeptide into a species that not only binds but has the wherewithal to induce oxidative strand scission of a nucleic acid substrate. This form of modification enables one to examine the location(s) and orientation(s) of a ligand bound to DNA in solution through affinity cleavage techniques [4], thus deriving crucial information in the absence of high-resolution data from crystallographic or nuclear magnetic resonance (NMR) methods. In contrast to the transformation of a protein or oligopeptide into a site-specific nuclease, the metal ion employed by peptides of the second
< previous page
page_428
next page >
< previous page
page_429
next page > Page 429
classification is central to the overall structure formed. Thus, with a low molecular weight peptide (e.g., a tripeptide), the metal ion creates a structurally organized species that has the ability to recognize and bind specific features of a nucleic acid. Recent examples suggest that some low molecular weight metallopeptides can interact with a nucleic acid selectively using, quite possibly, the same palette of weak molecular forces available to a protein (while also permitting their study through a host of techniques designed for the elucidation of small molecule-DNA interactions). Importantly, such systems may eventually bridge the gap between the study of protein-DNA and drug-DNA interactions and lead to new insights into the principles of nucleic acid recognition. This chapter surveys the ongoing use of metallopeptides in the study of protein-DNA interactions including strategies for their synthesis and applications to biological problems. In addition, the potential contributions of metallopeptides to the existing body of knowledge concerning small molecule and drug-DNA interactions are also examined. This chapter is written with the intent of acquainting the reader with the strategies currently available for the development and exploitation of metallopeptides with the hope of spurring the design of new ligand systems and further applications in the area of nucleic acids. 2 Oligopeptides Containing a Pendant Metal-Binding Domain 2.1 DNA Binding Domains of Proteins In the absence of high-resolution data (i.e., crystallography or NMR), affinity cleavage techniques [4] have assisted in determining the site selectivity and orientation of proteins bound to the DNA helix in solution. These techniques rely on the ability to selectively modify a protein with a reactive metal ion that ultimately catalyzes DNA strand scission through the generation of reactive oxygen species. In most cases present in the literature to date, the subject of investigation has been an oligopeptide that encompasses only the DNA binding domain of a protein (e.g., a helixturn-helix motif), thus facilitating its total chemical synthesis, modification for metal binding, and subsequent determination of selectivity. Currently, several strategies exist for the modifica-
< previous page
page_429
next page >
< previous page
page_430
next page > Page 430
tion of such domains involving the use of unnatural amino acid residues or peptide sequences of naturally occurring amino acids. 2.1.1 Unnatural or Derivatized Amino Acids The addition of a metal-binding segment to an oligopeptide with the preexisting ability to associate with nucleic acids can be accomplished through the incorporation of synthetic metal-binding amino acids. Several examples of the use of artificial amino acids or derivatized amino acid side chains exist in the literature [512] and many rely on the ability to incorporate a derivative of EDTA-Fe at a select position; EDTA-Fe, when reductively activated in the presence of dioxygen, generates diffusible oxidizing species capable of degrading nucleic acids. To date, methods for the synthetic incorporation of EDTA (e.g., tribenzyl-EDTA-γ-aminobutanoic acid, or BEG, and tricyclohexyl ester of EDTA, or TCE) at the amino termini and near the carboxy termini of oligopeptides using solid phase methods [5] have been developed (Figs. 1 and 2). These methods have been applied in the design of metallopeptides useful to our understanding of the activities of several DNA-binding proteins including Hin recombinase, which contains a helix-turn-helix motif [6-8], GCN4 [9,10], a member of the basic region-leucine zipper protein structural motif (b-ZIP) family of yeast transcriptional activators, and the Lac repressor [11]. In addition to the derivatization of wholly synthetic peptides, methods for a similar metallation of whole proteins have also been successfully developed in which surface-accessible cysteine residues (either present in the native sequence or selectively placed through site-directed mutagenesis) can be modified through the use of S-(2pyridylthio)cysteaminyl-EDTA (Fig. 3). Studies employing this means of metallation have facilitated an examination of the DNA binding selectivity of the catabolite gene activator protein (CAP) and Cro [12]. Alternatively, accessible cysteine residues of a protein of interest can also be modified with 5-(iodoacetamido)-1,10-phenanthroline which, in the presence of Cu(II) and appropriate activating conditions [13,14], also creates a targeted chemical nuclease [15]. In each of the above cases, once an oligopeptide or protein derivative of choice is synthesized and purified, DNA binding selectivity can be examined. For example, reactions that contain an EDTA-Fe deri-
< previous page
page_430
next page >
< previous page
page_431
next page > Page 431
Fig. 1. Scheme for the attachment of EDTA (in the form of TCE) to the amino terminus (left) or to the group of lysine at the second amino acid residue from the carboxy terminus (right) of synthetic oligopeptides. (Reproduced from Ref. 5 by permission of the American Chemical Society.)
< previous page
page_431
next page >
< previous page
page_432
next page > Page 432
Fig. 2. Orientation of Fe-EDTA-modified oligopeptides derived from the DNA binding helix-turn-helix motif of Hin(136186) on the DNA duplex. (Reproduced from Ref. 5 by permission of the American Chemical Society.) vatized oligopeptide/protein require a radiolabeled DNA substrate, dithiothreitol or ascorbate (for metal-catalyzed reductive activation of dioxygen), and appropriate buffer conditions. Using these conditions and knowledge of the patterns of DNA modification produced by diffusible radical oxidants generated by EDTA-Fe (Fig. 4), locations of binding can be mapped to single-nucleotide resolution [4]. 2.1.2 Naturally Occurring Amino Acids and Sequences Along with the development of oligopeptides that employ synthetically derived amino acids, recent experiments have also exploited naturally
< previous page
page_432
next page >
< previous page
page_433
next page > Page 433
Fig. 3. Modification of surface-accessible cysteine residues of a protein by S-(2-pyridylthio)cysteaminyl-EDTA. (Reproduced from Ref. 12 by permission of the American Chemical Society.) occurring peptide sequences that exhibit a high affinity for redox-active metal ions [1619]. Unlike the methods discussed in the previous section, this form of oligopeptide modification also has the potential advantage of allowing the protein/oligopeptide of interest to be modified through mutagenesis and recombinant methods of protein synthesis. Examples of this means of creating a metallopeptide include the recent derivatization of the amino terminus of oligopeptides by the simple metal binding tripeptide glycylglycylhistidine (GGH). This tripeptide, which mimics the amino terminal, square planar Cu(II) chelating domain of serum albumins [20,21], exhibits a high affinity for Cu(II) and Ni(II) (KD ~ 10161017) and forms a 1:1 complex mediated by the histidine imidazole nitrogen, two deprotonated amide nitrogens, and the terminal α-amine. When GGH is incorporated through solid phase peptide synthesis at the amino terminus of an oligopeptide, it can be activated in the presence of either Cu(II) or Ni(II) [involving Cu(II) + ascorbate and peroxide, or Ni(II) + monoperoxyphthalic acid or oxone (KHSO5)] to induce the cleavage of a nucleic acid substrate. Oligopeptide modifica-
< previous page
page_433
next page >
< previous page
page_434
next page > Page 434
Fig. 4. Cleavage patterns produced by a diffusible radical oxidant generated by a ligand bound in the major or minor grooves of the DNA helix. (Reproduced from Ref. 5 by permission of the American Chemical Society.)
tion through the synthetic addition of GGH has successfully altered Hin recombinase for affinity cleavage (Fig. 5) [1618]. Importantly, these studies indicate that, when bound to Cu(II) or Ni(II), a viable DNA cleaving agent is created with characteristics that suggest the operation of a nondiffusible oxidizing species. Oligopeptide modification by GGH thus creates an oxidizing system complementary to EDTA-Fe derivatization that can identify the precise location of the GGH appendage through cleavage experiments. Moreover, GGH modification does not require the preparation of specialized amino acids for metal chela-
< previous page
page_434
next page >
< previous page
page_435
next page > Page 435
Fig. 5. Incorporation of GGH at the amino terminus of the helix-turn-helix motif of Hin recombinase. Arrows indicate positions of cleavage induced by the metallopeptide. (Reproduced from Ref. 18 by permission of the American Chemical Society.)
< previous page
page_435
next page >
< previous page
page_436
next page > Page 436
tion nor does it require synthetic protocols beyond those normally employed in automated or manual solid phase peptide synthesis [1618]. As noted previously, one of the chief advantages of using a system of naturally occurring amino acids in the derivatization of an oligopeptide involves the potential biosynthetic incorporation of the metal binding domain through genetic alteration of the naturally occurring protein. Recently, such a system involving the derivatization of a zinc finger motif with amino terminal GGH was achieved [19]. Importantly, this biosynthetic protocol also extended modification to a naturally occurring metalloprotein and established that selective binding of Zn(II) to the finger motif and Ni(II) to the appended GGH domain can be accomplished within the same protein. While oligopeptide modification with GGH has several distinct advantages, one unfortunate disadvantage of this system stems from the necessary placement of the GGH domain at the amino terminus of an oligopeptide. This limitation prevents the application of GGH to a similar mapping of additional locations along the polypeptide strand, particularly disappointing given the nondiffusible nature of the oxidant generated by this domain. In an attempt to circumvent this limitation of GGH, a tripeptide linkage based on (δ)-Orn-Gly-His was recently developed [22] that permits the synthetic incorporation of a GGH-like unit at either the carboxy terminal or interior regions of an oligopeptide chain. This tripeptide unit, which can be readily incorporated into an oligopeptide chain through conventional solid phase techniques (Fig. 6), was shown to preserve the metal binding, electronic, and DNA cleavage properties of GGH [22,23]. As shown (Fig. 6), (δ)-Orn-Gly-His connects to the amino terminus of a peptide chain through the δ-amino group of an Orn residue, permitting the α-amino group of the same residue to participate in metal binding, e.g., with Ni(II) or Cu(II), in a fashion analogous to that of the amino terminal glycine in GGH. Further studies of the (δ)-Orn-Gly-His sequence also indicate that the chirality of the His residue can influence the overall structure of the chain; incorporation of a D-His residue in place of an L-His residue results in the conversion of the oligopeptide chain from one in which the amino terminus and carboxy terminus diverge from the bound metal at distinct angles from one another to a chain that can be predominantly linear (Fig. 7). These differences in chain structure, which can be
< previous page
page_436
next page >
< previous page
page_437
next page > Page 437
Fig. 6. Stepwise solid phase peptide synthesis of ornithine-based metal-binding oligopeptides: (a) coupling of Boc-Orn(Fmoc) via Boc-benzyl method; (b) two cycles of Boc-benzyl peptide synthesis; (c) thiophenol/DMF; (d) piperidine/DMF; (e) TFA/CH2Cl2; (f) NH3/CH3OH; (g) Ni(OAc)2/10 mM sodium cacodylate.
< previous page
page_437
next page >
< previous page
page_438
next page > Page 438
Fig. 7. Molecular models of peptides containing (δ)-Orn-Gly-D-His (top) and (δ)-Orn-Gly-L-His metal binding sequences. achieved through a very simple substitution, may find applications in probing the exact locations or orientations of ligand-DNA binding (i.e., the oligopeptide chain can be customized to suit a particular structural environment or potentially direct the oxidizing potential of the bound metal ion to a select residue or location within the substrate; a similar effect may also operate when amino terminal GGD-H is employed [18]). The development of (δ)-Orn-Gly-His thus permits the oxidative properties of GGH to be incorporated at a variety of sterically permissible locations within an oligopeptide strand and, possibly, several locations simultaneously. Of possibly even greater importance, (δ)-Orn-Gly-
< previous page
page_438
next page >
< previous page
page_439
next page > Page 439
His takes advantage of established synthetic protocols and readily available starting materials making this modification accessible to researchers not directly involved in peptide synthesis. 2.2 Minor Groove-Binding Oligopeptides Along with the development of metallopeptides containing elements of secondary or supersecondary protein structure, similar strategies for the incorporation of reactive transition metal ions have been employed in the examination of the DNA binding selectivity of relatively small peptides, e.g., analogs of distamycin or netropsin that bind in the minor groove of the DNA helix [24,25]. Like the examples discussed above, the metal-bound appendage again serves to outfit the ligand of interest with the potential to induce DNA strand scission. In several cases examined, the peptide of interest has been either distamycin linked directly to EDTA-Fe [26,27] or similarly modified oligo-N-methylpyrrolecarboxamide analogs of this minor groove binder (Fig. 8) [2833]. Like the DNA binding domains discussed above, these metal-derivatized oligopeptides have found use in determining the sequence selectivity of a particular distamycin analog through affinity cleavage experiments. More recently, similar metallopeptide systems have also been employed in the determination of the orientation of peptides in heterodimeric complexes [34,35]. In addition to the use of EDTA-Fe derivatization, naturally occurring peptide sequences have also been employed in the modification of groove-binding oligopeptides in a fashion similar to their applications discussed previously. In one example [36], an oligopeptide consisting of 14 naturally occurring L-amino acids encompassing one unit of the proposed DNA binding motif found in the carboxy terminal domain of RNA polymerase II [37] was equipped with GGH at its amino terminus. This relatively short oligopeptide was designed to contain two distinct, independently functioning domains, one that bound a reactive Cu(II) via the amino terminal GGH and another that contained the DNA binding motif that folds into two overlapping β-turns. Experiments with this peptide indicated that each domain was capable of functioning independently; the GGH was capable of binding Cu(II) and initiating the oxidative strand scission of DNA while the other was capable of
< previous page
page_439
next page >
< previous page
page_440
next page > Page 440
Fig. 8. Minor groove-binding metallopeptide analogs of distamycin (DE-Fe, left) and netropsin (2-ImNE and 2-PyNE, right). (Reproduced from Refs. 26 and 34 by permission of the American Chemical Society.) folding into a tandem β-turn structure and binding to the target DNA helix. This example thus demonstrated that GGH also has applications in the development of reactive oligopeptides of a much lower molecular weight than previously examined. 2.3 Other Metal-Binding Oligopeptide Systems In contrast to the metallopeptides discussed above, several other examples of metal-derivatized oligopeptides not easily classified as to their type have also appeared recently in the literature. In these examples, the metal is not a passive moiety but actually facilitates the binding of the entire metal-peptide conjugate with a DNA substrate. In one example [38], the tripeptide Gly-His-Lys (GHK), which like GGH also binds Cu(II) [3941], was covalently linked to the Nmethylpyrrolecarboxamide portion of a netropsin-9-(4-glycylanilino)acridine
< previous page
page_440
next page >
< previous page
page_441
next page > Page 441
Fig. 9. Structure of the Gly-His-Lys~netropsin-acridine hybrid molecule. (Reproduced from Ref. 38 by permission of the American Chemical Society.) hybrid molecule (Fig. 9). In this particular case, it was found that the GHK moiety contributed positively to the interaction of the entire molecule with the DNA backbone; the binding of the GHK-containing hybrid was found to be 20-fold greater than the netropsin-acridine portion alone due in part to the positively charged lysine side chain. Overall, studies of this hybrid molecule revealed that (1) the Cu(II)-GHK portion of the structure was entirely compatible with minor groove binding and (2) DNA cleavage induced by the metallopeptide portion indicated the potential for two orientations of the entire structure on the DNA duplex. Similarly, a family of metal-peptide complexes was recently synthesized [42] by coupling of a series of short oligopeptides derived from the α3 helix of the DNA-binding phage P22 repressor with the octahedral metallointercalator Rh(phi)2(phen)3+ through a modified phen residue [phen', 5-(aminoglutaryl)-1,10phenanthroline] (Fig. 10). Unlike most of the systems discussed previously, this series of modified oligopeptides was synthesized to determine whether the side chain functionalities of a DNA recognition peptide could be used to contribute to the activity of the metal complex which alone is capable of sequence-selective DNA photocleavage [4346]. These modified oligopeptides were observed to bind selectively to the DNA major groove in a fashion dependent on the nature of the side chain functional groups present in the peptide sequence. A single Glu
< previous page
page_441
next page >
< previous page
page_442
next page > Page 442
Fig. 10. Structure of a Rh(phi)2(phen')-peptide conjugate. The sequence of the appended peptide moiety is given in the single-letter code. (Reproduced from Ref. 42 by permission of the American Chemical Society.) residue at position 10 of the helix-forming moiety was found to be capable of driving a selective interaction of the hybrid with 5'-CCA-3' sites contained within DNA restriction fragments; substitutions at other positions within the peptide did not alter this selectivity. Importantly, this form of metal-peptide conjugate creates a new strategy for the design of sequence-selective DNA-binding agents and also provides a useful tool through which to study larger DNAbinding proteins. In an additional example [47,48], a metallo-bZIP protein derived from GCN4 [49] was constructed in which the native dimerization domain (a leucine ''zipper") was replaced by a series of metal complexes that systematically altered the orientation of the basic region helical "arms" that form contacts with the major groove of a target DNA site (Fig. 11). In this case, the metal employed was not utilized as a cleavage functionality but, given its defined geometry, was used as a template to explore structure-function relationships within this molecule. Results from these studies indicated that both the overall affinity and the specificity of peptide binding to DNA could be altered through small
< previous page
page_442
next page >
< previous page
page_443
next page > Page 443
Fig. 11. Schematic representation of metallo-bZIP proteins derived from GCN4. (Reproduced from Ref. 48 by permission of the author.) changes in the stereochemistry of the central metal complex. In addition, this study demonstrated the ability to exploit a metal complex to control the overall structural orientation of an oligopeptide, possibly leading to a general approach that could be used in the examination of other protein-DNA systems or proteins in general. 3 Low Molecular Weight Metallopeptides Along with the ability to act as reactive appendages, low molecular weight metallopeptides without an attached DNA binding domain can adopt structures that interact directly with the DNA helix. Unlike many of the examples discussed so far, these metallopeptides more
< previous page
page_443
next page >
< previous page
page_444
next page > Page 444
closely resemble the host of complexes currently employed in the study of nucleic acids [13] while potentially including chemical functionalities and structural features found in antitumor natural products and proteins that interact with DNA. These features make metallopeptides attractive in the design of molecules useful in testing principles of DNA molecular recognition and may also provide models of drug- and protein-DNA binding. 3.1 Metallo-Gly-Gly-His As discussed in a previous section, amino terminal GGH represents the consensus sequence for the square planar, Cu(II) binding and transport domain of serum albumins [20,21]. This tripeptide binds Cu(II) in a 1:1 complex (KD ~ 10161017; pH range 6.511) [20,21] and forms similar complexes with Ni(II) [5052] or Co(III) [53]. While most work to date has focused on the cleavage of DNA by metallo-GGH as part of an established DNA-binding ligand [1619,36], several studies also have determined that the metallo-GGH tripeptide alone is capable of inducing DNA strand scission when appropriately activated. In an examination of the DNA cleavage activity of Cu(II)-GGH, strand scission of a supercoiled substrate was found to depend on the presence of ascorbate and, quite likely, endogenously generated hydrogen peroxide [54,55]. Alternatively, recent evidence suggests that the Ni(II) complex of GGH is also capable of generating diffusible oxygen radicals in the presence of reduced dioxygen species [56,57]. In a subsequent investigation of Ni(II)-GGH, DNA cleavage by this metallopeptide was explored in the presence of the activating agent KHSO5 (oxone) [58]; the results of this study indicated that, relative to other Ni(II) complexes, Ni(II)-GGH exhibited little activity, due quite possibly to the overall negative charge of this complex (containing a free carboxylate terminus) which may have lowered its affinity for the polyanionic DNA backbone. In light of the above findings, results from our own laboratory [59] have determined that Ni(II)-GGH, when synthesized as its respective carboxamide [resulting in a Ni(II) complex of neutral charge; Fig. 12] and activated with oxone, mediates substantial DNA damage that occurs with some degree of sequence selectivity. As shown in Fig. 13,
< previous page
page_444
next page >
< previous page
page_445
next page > Page 445
Fig. 12. Structure of Ni(II)-Gly-Gly-His-CONH2. which is representative of a much larger dataset, the major positions of Ni(II)-GGH-induced cleavage appear to occur within or adjacent to mixed AT-rich regions of DNA. Importantly, these data patterns also revealed that (1) cleavage is not selective for any particular nucleobase (i.e., the complex does not cleave ''G" residues exclusively) and (2) given the single sites of cleavage observed, a diffusible oxidant was probably
Fig. 13. Ni(II)-GGH cleavage of a 32P-end-labeled DNA strand derived from the 167-bp EcoRI → Rsa I restriction fragment of pBR322 (bars represent the nucleotide positions of strand scission with lengths proportional to cleavage intensity).
< previous page
page_445
next page >
< previous page
page_446
next page > Page 446
not produced in the reaction. In addition to these findings, the total dataset obtained with this metallopeptide also revealed a consistent 3'-asymmetrical pattern of DNA modification that suggests binding of the peptide in the minor groove of the helix; the authentic minor groove binding ligand distamycin can also be footprinted by this metallopeptide, further supporting the notion of a minor groove binding mechanism. 3.2 Orientation of Metallopeptides on the DNA Helix The groove and site selectivity displayed by Ni(II)-GGH is supported indirectly by electron paramagnetic resonance (EPR) studies of the orientation of similar tripeptides on DNA fibers [60]. In this study, which included Cu(II)-GlyHis-Lys and Cu(II)-Gly-His-Gly, the mean planes of these square planar complexes were found to orient parallel to the DNA fiber axis in single, well-oriented configurations (as apparently dictated by the -Gly-His portion of the structure). Such a defined binding orientation implies a directed means of interacting with the DNA helix and that factors other than simple electrostatic surface binding are at work. Indeed, the lysine residue of Cu(II)-Gly-His-Lys was determined not to be an influential factor in the binding orientation of the peptides studied. Importantly, these studies suggest that the overall structure of a metallopeptide determines its orientation on the DNA strand and, like other metal complexes [13], appropriately configured metallopeptides may be able to sense features of DNA structure. Interestingly, in a comparative study [60], the antitumor agent bleomycin [61,62], which has a peptide-derived domain responsible for Cu or Fe binding and bears a striking resemblance to GGH metallopeptides (Fig. 14), was found to adopt a similar, distinct orientation on DNA fibers (albeit at an angle of 65° with respect to the fiber axis). Like Ni(II)-GGH, metallobleomycins have the ability to induce selective DNA strand scission (predominantly at the pyrimidine of 5'-GC and 5'-GT sites) through the operation of a nondiffusible oxidant [61,62]. In an additional parallel to the chemistry of metallo-GGH, the metal binding domain of bleomycin has recently been utilized to convert Hin recombinase into a site-specific nuclease; the cleavage chemistry here, however, was attributed to the production of a diffusible radical oxidant [63].
< previous page
page_446
next page >
< previous page
page_447
next page > Page 447
Fig. 14. Proposed structure of the peptide-derived metal binding domain of Fe-bleomycin (R = bithiazole and C-terminal region of bleomycin). Metallo-GGH and metallobleomycin, therefore, while dissimilar in many structural and mechanistic respects, both have the wherewithal to interact with the DNA helix and induce selective strand scission. These similarities suggest that appropriately structured metallopeptides might be developed which can interact with controlled DNA selectivity. In addition, given the apparent nondiffusible nature of the chemistry of Ni(II)-GGH, reagents selective for a particular atom of a target nucleotide might be created to permit control of the chemistry of DNA strand scission. These possibilities are especially promising in light of recent findings which suggest that the metal binding domain of bleomycin contributes significantly to the site selectivity of this antitumor agent [64,65]. 4 Conclusions and Future Prospects Metallopeptides have contributed significantly to the design of reagents useful in the study of DNA-ligand interactions. Their ability to impart reactivity to an otherwise unreactive organic structure have allowed the development of affinity cleavage reagents capable of investigating the binding of proteins, secondary structural motifs, and oligopeptides. In addition, varied strategies for their construction have exploited features of synthetic chemistry, techniques of protein modification, and recombinant methods of protein synthesis thus bringing together sev-
< previous page
page_447
next page >
< previous page
page_448
next page > Page 448
eral aspects of science in order to probe nucleic acid interactions. No doubt, as recent examples suggest, the increasing use of metal-modified oligopeptides in the study of nucleic acids will produce structural and chemical applications that go well beyond those discussed herein. Along with their ability to modify existing DNA binding motifs, low molecular weight metallopeptides have also been demonstrated to produce viable DNA strand scission reagents. Given the similarities that exist between simple metallopeptides and selective DNA binding agents such as bleomycin, the question remains as to whether or not metallo-GGH or similar metallopeptides can be used to develop selective binding or modifying agents through rational alterations in amino acid composition and chirality [59]. The process of developing such metallopeptides and assessing whether any generalizable principles govern their binding to DNA may shed light on features of DNA molecular recognition not currently probed by conventional metal complexes. Abbreviations
Ac
acetyl
BEG
tribenzyl-EDTA-γ-aminobutanoic acid
bZIP
basic region-leucine zipper protein structural motif
CAP
catabolite gene activator protein
DCC
dicyclohexylcarbodiimide
DCM
dichloromethane
DE
distamycin
DMF
dimethylformamide
Dnp
dinitrophenyl
EDTA
ethylenediaminetetraacetic acid
EPR
electron paramagnetic resonance
Fmoc
9-fluorenylmethyloxycarbonyl
GCN4
yeast transcriptional activator protein
GGH
glycylglycylhistidine (Gly-Gly-His)
GHK
glycylhistidyllysine (Gly-His-Lys)
HOBt
N-hydroxybenzotriazole
2-ImNE
1-methylimidazole-2-carboxamide-netropsin-EDTA-Fe(II)
NMR
nuclear magnetic resonance
< previous page
page_448
next page >
< previous page
page_449
next page > Page 449
phen'
5-(aminoglutaryl)-1,10-phenanthroline
phi
phenanthrenequinone diimine
2-PyNE
pyridine-2-carboxamide-netropsin-EDTA-Fe(II)
tBOC
tert-butyloxycarbonyl
TCE
tricyclohexyl ester of EDTA
TFA
trifluoroacetic acid
References 1. D. S. Sigman, A. Mazumder, and D. M. Perrin, Chem. Rev., 93, 2295 (1993). 2. A. M. Pyle and J. K. Barton, Prog. Inorg. Chem., 38, 413 (1990). 3. E. C. Long, J. Inorg. Organomet. Pol., 3, 3 (1993). 4. P. B. Dervan, Meth. Enzymol., 208, 497 (1991). 5. J. P. Sluka, J. H. Griffin, D. P. Mack, and P. B. Dervan, J. Am. Chem. Soc., 112, 6369 (1990). 6. J. P. Sluka, S. J. Horvath, M. F. Bruist, M. I. Simon, and P. B. Dervan, Science, 238, 1129 (1987). 7. D. P. Mack, J. P. Sluka, J. A. Shin, J. H. Griffin, M. I. Simon, and P. B. Dervan, Biochemistry, 29, 6561 (1990). 8. J. P. Sluka, S. J. Horvath, A. C. Glasgow, M. I. Simon, and P. B. Dervan, Biochemistry, 29, 6551 (1990). 9. M. G. Oakley and P. B. Dervan, Science, 248, 847 (1990). 10. M. G. Oakley, M. Mrksich, and P. B. Dervan, Biochemistry, 31, 10969 (1992). 11. J. A. Shin, R. H. Ebright, and P. B. Dervan, Nucl. Acids Res., 19, 5233 (1991). 12. Y. W. Ebright, Y. Chen, P. S. Pendergrast, and E. H. Ebright, Biochemistry, 31, 10664 (1992). 13. T. W. Bruice, J. Wise, and D. S. Sigman, Biochemistry, 29, 2185 (1990). 14. T. W. Bruice, J. Wise, D. S. E. Rosser, and D. S. Sigman, J. Am. Chem. Soc., 113, 5446 (1991). 15. D. S. Sigman, T. W. Bruice, A. Mazumder, and C. L. Sutton, Acc. Chem. Res., 26, 98 (1993).
< previous page
page_449
next page >
< previous page
page_450
next page > Page 450
16. D. P. Mack, B. L. Iverson, and P. B. Dervan, J. Am. Chem. Soc., 110, 7572 (1988). 17. D. P. Mack and P. B. Dervan, J. Am. Chem. Soc., 112, 4604 (1990). 18. D. P. Mack and P. B. Dervan, Biochemistry, 31, 9399 (1992). 19. M. Nagaoka, M. Hagihara, J. Kuwahara, and Y. Sugiura, J. Am. Chem. Soc., 116, 4085 (1994). 20. N. Camerman, A. Camerman, and B. Sarkar, Can. J. Chem., 54, 1309 (1976). 21. S.-J. Lau, T. P. A. Kruck, and B. Sarkar, J. Biol. Chem., 249, 5878 (1974). 22. D. F. Shullenberger, P. D. Eason, and E. C. Long, J. Am. Chem. Soc., 115, 11038 (1993). 23. E. C. Long, P. D. Eason, and D. F. Shullenberger in Metal-Containing Polymeric Materials, Plenum Press, New York, in press. 24. M. L. Kopka, C. Yoon, D. Goodsell, P. Pjura, and R. E. Dickerson, Proc. Natl. Acad. Sci. USA, 82, 1376 (1985). 25. C. Zimmer and U. Wahnert, Prog. Biophys. Mol. Biol., 47, 31 (1986). 26. P. G. Schultz, J. S. Taylor, and P. B. Dervan, J. Am. Chem. Soc., 104, 6861 (1982). 27. J. S. Taylor, P. G. Schultz, and P. B. Dervan, Tetrahedron, 40, 457 (1984). 28. R. S. Youngquist and P. B. Dervan, J. Am. Chem. Soc., 107, 5528 (1985). 29. R. S. Youngquist and P. B. Dervan, Proc. Natl. Acad. Sci. USA, 82, 2565 (1985). 30. P. B. Dervan, Science, 232, 464 (1986). 31. J. H. Griffin and P. B. Dervan, J. Am. Chem. Soc., 108, 5008 (1986). 32. J. H. Griffin and P. B. Dervan, J. Am. Chem. Soc., 109, 6840 (1987). 33. W. S. Wade and P. B. Dervan, J. Am. Chem. Soc., 109, 1574 (1987). 34. W. S. Wade, M. Mrksich, and P. B. Dervan, J. Am. Chem. Soc., 114, 8783 (1992). 35. M. Mrksich and P. B. Dervan, J. Am. Chem. Soc., 115, 2572 (1993). 36. D. F. Shullenberger and E. C. Long, Bioorg. Med. Chem. Lett., 3, 333 (1993).
< previous page
page_450
next page >
< previous page
page_451
next page > Page 451
37. X. Huang, D. F. Shullenberger, and E. C. Long, Biochem. Biophys. Res. Commun., 198, 712 (1994). 38. C. Bailly, J.-S. Sun, P. Colson, C. Houssier, C. Hélène, M. J. Waring, and J.-P. Henichart, Bioconj. Chem., 3, 100 (1992). 39. L. Pickart, J. H. Freedman, W. J. Loker, J. Peisach, C. M. Perkins, R. E. Stenkamp, and B. Weinstein, Nature, 288, 715 (1980). 40. J. H. Freedman, L. Pickart, B. Weinstein, W. B. Mims, and J. Peisach, Biochemistry, 21, 4540 (1982). 41. J. P. Laussac, R. Haran, and B. Sarkar, Biochem. J., 209, 533 (1983). 42. N. Y. Sardesai, K. Zimmerman, and J. K. Barton, J. Am. Chem. Soc., 116, 7502 (1994). 43. A. M. Pyle, E. C. Long, and J. K. Barton, J. Am. Chem. Soc., 111, 4520 (1989). 44. S. S. David and J. K. Barton, J. Am. Chem. Soc., 115, 2984 (1993). 45. A. H. Krotz, L. Y. Kuo, T. P. Shields, and J. K. Barton, J. Am. Chem. Soc., 115, 3877 (1993). 46. A. Sitlani, E. C. Long, A. M. Pyle, and J. K. Barton, J. Am. Chem. Soc., 114, 2303 (1992). 47. B. Cuenoud and A. Schepartz, Science, 259, 510 (1993). 48. B. Cuenoud and A. Schepartz, Proc. Natl. Acad. Sci. USA, 90, 1154 (1993). 49. C. R. Vinson, P. B. Sigler, and S. L. McKnight, Science, 246, 911 (1989). 50. F. P. Bossu and D. W. Margerum, Inorg. Chem., 16, 1210 (1977). 51. C. E. Bannister, J. M. T. Raycheba, and D. W. Margerum, Inorg. Chem., 21, 1106 (1982). 52. T. Sakurai and A. Nakahara, Inorg. Chem., 19, 847 (1980). 53. C. J. Hawkins and J. Martin, Inorg. Chem., 22, 3879 (1983). 54. S.-H. Chiou, J. Biochem., 94, 1259 (1983). 55. S.-H. Chiou, W.-C. Chang, Y.-S. Jou, H.-M. M. Chung, and T.-B. Lo, J. Biochem., 98, 1723 (1985). 56. S. Inoue and S. Kawanishi, Biochem. Biophys. Res. Commun., 159, 445 (1989). 57. N. Cotelle, E. Tremolieres, J. L. Bernier, J. P. Catteau, and J. P. Henichart, J. Inorg. Biochem., 46, 7 (1992).
< previous page
page_451
next page >
< previous page
page_452
next page > Page 452
58. X. Chen, S. E. Rokita, and C. J. Burrows, J. Am. Chem. Soc., 113, 5884 (1991). 59. Q. Liang, P. D. Eason, and E. C. Long, J. Am. Chem. Soc., 117, 9625 (1995). 60. M. Chikira, T. Sato, W. E. Antholine, and D. H. Petering, J. Biol. Chem., 266, 2859 (1991). 61. S. M. Hecht, Acc. Chem. Res., 19, 383 (1986). 62. J. Stubbe and J. W. Kozarich, Chem. Rev., 87, 1107 (1987). 63. M. G. Oakley, K. D. Turnbull, and P. B. Dervan, Bioconj. Chem., 5, 242 (1994). 64. B. J. Carter, V. S. Murty, K. S. Reddy, S.-N. Wang, and S. M. Hecht, J. Biol. Chem., 265, 4193 (1990). 65. S. A. Kane, A. Natrajan, and S. M. Hecht, J. Biol. Chem., 269, 10899 (1994).
< previous page
page_452
next page >
< previous page
page_453
next page > Page 453
15 Targeting of Nucleic Acids by Iron Complexes Alexandra Draganescu and Thomas D. Tullius Department of Chemistry, The Johns Hopkins University, Baltimore, MD 21218, USA
454
1. Introduction
2. The Fenton Reaction Generates the Hydroxyl Radical
2.1. Chemistry of the Hydroxyl Radical with Nucleic Acids
456
457
2.2. Experimental Strategy
3. The Hydroxyl Radical As a Probe of Nucleic Acid Structure
460
461
3.1. DNA Structure
461
3.1.1. The Helical Twist of DNA
461
3.1.2. Bent DNA
464
3.1.3. DNA Junctions
468
3.2. Protein-DNA Complexes
468
3.2.1. Prokaryotic Repressors
470
3.2.2. Homeodomains
471
3.2.3. Zinc Fingers
< previous page
456
page_453
next page >
< previous page
page_454
next page > Page 454 473
3.2.4. The Nucleosome
475
3.2.5. Transcription Complexes
476
3.2.6. Other Protein-DNA Complexes
477
3.3. Drug-DNA Complexes
3.4. RNA Structure and RNA-Protein Complexes
478
479
4. Protein-Tethered Iron Complexes
479
5. Conclusions
479
Abbreviations
480
References
1 Introduction A wide variety of methods have been developed for the study of nucleic acid structure. Nuclear magnetic resonance (NMR) and X-ray crystallography have been used to examine DNA and RNA structure, as well as complexes of DNA with proteins and drugs. However, both of these methods require large amounts of material and are carried out under conditions which do not closely approximate a physiological environment. Alternatively, experiments involving enzymatic or chemical modification of DNA or RNA allow for the study of structure at relatively low concentration of the nucleic acid molecule and under physiological conditions. The principle behind these experiments is based on the ability of the enzyme or chemical reagent to modify the nucleic acid in some detectable way, usually resulting in cleavage of the phosphodiester backbone. The cleavage pattern is conveniently observed by electrophoresis of the products of the reaction on a denaturing polyacrylamide gel, using the methods developed for DNA sequencing. Single-nucleotide resolution, using picomole or femtomole quantities of the nucleic acid molecule, is typically obtained in such experiments. When a drug or protein molecule is bound to the nucleic acid molecule, or when the RNA or DNA molecule adopts an unusual structure, the chemical or enzymatic probe no longer has access to all possible modification sites. By comparing the cleavage patterns produced in the presence and absence of bound ligand or fully folded structure, valuable information is ob-
< previous page
page_454
next page >
< previous page
page_455
next page > Page 455
tained about the structure of the nucleic acid molecule, or the location of the bound ligand and its mode of interaction. The first reagents used for such experiments were enzymes. This technique is still common today and has been applied to a multitude of systems. The enzymes used most often in these experiments are deoxyribonuclease I (DNase I) and exonuclease III. The DNase I footprinting method was first used by Galas and Schmitz [1] for the study of lac repressor protein bound to DNA and independently by Ptashne and coworkers [2] in their work on the bacteriophage λ repressor. Noll had found a similar effect of bound protein on the DNase I cleavage pattern of DNA in work on the nucleosome [3]. DNase I is advantageous for such studies because it cleaves endonucleolytically throughout a DNA molecule, although not every phosphodiester is cut by the enzyme. However, if a protein molecule is bound to the DNA at a specific site DNase I is unable to cleave there and a blank spot (the ''footprint") appears in the cleavage pattern. Exonuclease III digests the DNA in a DNA-protein complex exonucleolytically starting from the 3' end, until its progress is halted by the bound protein [46]. While enzyme-based footprinting methods are useful in defining where a ligand (protein or drug molecule) is bound to DNA, detailed structural information on the complex is difficult to obtain due to the size of the enzyme and its sequence preferences for cleavage of the nucleic acid. For higher resolution studies, chemical and photochemical probes have been devised [7,8]. Chemical probes consist either of organic compounds or of metal complexes that modify and ultimately lead to cleavage of the nucleic acid molecule. While chemical probes provide a higher degree of structural detail compared to enzymes, they also have drawbacks. Many metal complexes intercalate into, or bind in the grooves of, the nucleic acid molecule and therefore might disrupt the structure being examined. Most of these reagents exhibit some sequence preference or specificity, and therefore information for all nucleotides within a particular nucleic acid molecule is not available. An alternative technique, lacking the disadvantages above, utilizes the Fenton reaction [9] to produce a small, uncharged, highly reactive, non-sequence-specific chemical probe which can provide detailed information about nucleic acid structure [10,11]. The probe is the hydroxyl radical. In this chapter we first discuss the production of the
< previous page
page_455
next page >
< previous page
page_456
next page > Page 456
hydroxyl radical in solution and its chemistry with nucleic acids. We then go on to describe the wide variety of systems that have been studied using the hydroxyl radical footprinting technique. 2 The Fenton Reaction Generates the Hydroxyl Radical 2.1 Chemistry of the Hydroxyl Radical with Nucleic Acids The hydroxyl radical can be produced by a variant of the Fenton reaction [9] in which iron(II), chelated by ethylenediaminetetraacetic acid (EDTA), reacts with hydrogen peroxide [12]:
The reaction yields an oxidized iron complex, the hydroxide ion, and the hydroxyl radical. The reaction is carried out in the presence of ascorbate ion, which reduces the Fe(III) product to Fe(II), creating a reaction system that is catalytic in iron. Chelation of iron by a negatively charged ligand prevents its binding to polyanionic DNA or RNA. The non-sequence-specific cleavage pattern that is produced by the reaction of DNA with the products of Eq. (1) implies that the cleaving agent is small, diffusible, and reactive [1315]. Cleavage of the nucleic acid molecule appears to result mainly from initial radical attack on the ribose or deoxyribose moiety [16,17]. Subsequent chemistry leads to elimination of a single nucleoside (base and attached deoxyribose), leading to formation of a single-stranded gap in the nucleic acid sequence. Mechanistic studies [13] reveal that the highly reactive radical may be abstracting hydrogen from multiple positions on the sugar and strand scission may occur via several pathways. Preliminary analysis of the products [13,18] of hydroxyl radical cleavage of DNA, and comparison with products resulting from single-hydrogen-atom abstraction from DNA by reagents such as Fe-bleomycin [19], are consistent with this hypothesis. Although Fe-bleomycin uses a different reactive species to abstract hydrogen from the deoxyribose moiety, the subsequent chemistry leading to strand scission is likely to
< previous page
page_456
next page >
< previous page
page_457
next page > Page 457
be similar. A thorough examination of the mechanistic details of the reaction of the hydroxyl radical with DNA is ongoing in our laboratory. 2.2 Experimental Strategy Protection and interference are two experimental approaches to studying nucleic acid structure with the hydroxyl radical. In a protection (or footprinting) experiment, the cleavage pattern of B-form DNA or unfolded RNA (the control) is compared to that of a DNA-ligand complex, an unusual DNA structure, or a folded RNA. Cleavage of the nucleic acid molecule is diminished at the location of the ligand or the unusual structure. A general scheme for a hydroxyl radical footprinting experiment is shown in Fig. 1. Radiolabeled double-stranded DNA is first incubated with the ligand of choice and allowed to form a complex. To the binding reaction are added [Fe(EDTA)]2, sodium ascorbate, and hydrogen peroxide. The reaction is allowed to proceed for 12 min and is then quenched by the addition of thiourea. The gapped DNA is isolated, denatured, and electrophoresed on a denaturing polyacrylamide gel. The gel is imaged on X-ray film or an imaging phosphor screen and the resulting image analyzed by scanning densitometry. Detailed experimental procedures and considerations have been published [15,2022]. To obtain information about particular nucleotides within a binding site which are important for protein binding, an interference experiment is used. In an interference experiment the DNA first is chemically modified, and then the protein is added and allowed to bind to the modified nucleic acid molecule. Complexes are separated from free DNA by nondenaturing gel electrophoresis. Both bound and free DNA are excised from the gel matrix, treated further to induce backbone cleavage if necessary, and then electrophoresed on a denaturing polyacrylamide gel. The results are visualized and analyzed as discussed previously for the footprinting experiment. The missing nucleoside experiment [23] is an interference experiment that takes advantage of the unique properties of the hydroxyl radical as a DNA-modifying agent. A diagram of the experiment is shown in Fig. 2. A singly radiolabeled DNA substrate is treated with the hydroxyl radical such that each DNA molecule contains a maximum of
< previous page
page_457
next page >
< previous page
page_458
next page > Page 458
Fig. 1. A hydroxyl radical footprinting experiment. DNA is cleaved by the hydroxyl radical in the presence (left) and absence (not shown) of protein. The cleavage products are separated by denaturing polyacrylamide gel electrophoresis (right). Since the reaction conditions are adjusted so that each DNA molecule is cleaved no more than once, the majority of DNA (≥70%) remains uncut. Intact DNA is represented by the intense band at the top of the gel. The non-sequence-specific cleavage of DNA by the hydroxyl radical results in a ladder of bands of approximately equal intensity in the absence of ligand (gel, left-hand lane). The presence of bound ligand causes decreased cleavage at certain positions, resulting in a footprint (gel, right-hand lane).
< previous page
page_458
next page >
< previous page
page_459
next page > Page 459
Fig. 2. A missing nucleoside experiment. Radiolabeled DNA is treated with the hydroxyl radical under the same reaction conditions as the footprinting experiment (see Fig. 1). The resulting sample of gapped DNA is incubated with the protein, followed by nondenaturing polyacrylamide gel electrophoresis to separate protein-bound from unbound DNA. Bands containing bound and unbound DNA are excised from the gel, and the DNA is extracted and electrophoresed on a denaturing gel. An even ladder of bands is observed in a control sample to which no ligand was added (lane marked "Free"). The bound lane contains bands representing all missing nucleoside positions which do not affect ligand binding. Important contacts are absent from this lane, but appear in the unbound lane.
< previous page
page_459
next page >
< previous page
page_460
next page > Page 460
one missing nucleoside, randomly positioned throughout the DNA. Protein is then added to the collection of gapped DNA molecules and allowed to bind. If a gap is present at a position which is unimportant for complex formation, the protein will bind to that DNA molecule and the complex will migrate more slowly on a native polyacrylamide gel than free DNA. If, on the other hand, a gap is present at a nucleoside that is crucial for complex formation, the protein can no longer bind and the gapped DNA molecule migrates as free DNA on the gel. Both bands, bound and free DNA, are excised from the gel and the DNA is isolated. Electrophoresis of the two DNA samples on a denaturing gel then identifies the nucleosides that are important for complex formation by their absence from the bound sample and corresponding overrepresentation in the unbound sample. The advantage of the hydroxyl radical-based missing nucleoside experiment over its predecessor, the missing contact experiment [24], is that modification at each nucleoside within a DNA molecule is accomplished in a single reaction rather than the multiple Maxam-Gilbert sequencing reactions required by the missing contact method. The advantage of the missing nucleoside experiment over another experimental approach, mutagenesis, which aims to obtain similar information, is that one samples many ''mutants" in a single experiment. Only individual nucleosides of a base pair are altered, providing base- (and not just base pair-) specific information on protein-DNA contacts. 3 The Hydroxyl Radical As a Probe of Nucleic Acid Structure Since the introduction of the hydroxyl radical as a sensitive probe to study nucleic acid structure [10,11], both the footprinting and missing nucleoside experiments have been applied to a variety of biological systems. Any nucleic acid sequence may be probed by the hydroxyl radical since it nonspecifically cleaves single- and double-stranded DNA and RNA [11,25]. However, there is some evidence that when single-stranded DNA is treated with the hydroxyl radical, a substantial amount of cleavage may result from hydrogen abstraction from the bases rather than from the deoxyribose moieties [26]. While it is reason-
< previous page
page_460
next page >
< previous page
page_461
next page > Page 461
able that more base damage would occur in single-stranded DNA, experimental evidence from our laboratory [27] and others [25] has not shown large differences in cleavage efficiency between single- and double-stranded DNA or RNA. 3.1 DNA Structure 3.1.1 The Helical Twist of DNA Due to the small, diffusible, and highly reactive nature of the hydroxyl radical, it seemed reasonable that this probe could be used to measure the helical periodicity of DNA in solution [11]. This was accomplished by immobilizing the target DNA molecules on a solid support, such as calcium phosphate microcrystals, followed by treatment with the hydroxyl radical. While cleavage of free DNA in solution results in approximately equal cleavage at every position along the sequence, bound DNA is periodically protected from radical attack at positions where it contacts the inorganic support. The cleavage pattern thus appears sinusoidal, directly reflecting the helical nature of duplex DNA. By counting the number of bases between sites of strong protection, the number of base pairs per turn of the DNA helix (or the helical periodicity) can be measured [11,28]. 3.1.2 Bent DNA A striking example of natural bent DNA is a fragment of the kinetoplast minicircle from the trypanosomatid Crithidia fasciculata, which contains 18 repeats of A46 which are approximately phased with the DNA helical repeat. These short runs of adenine, which are associated with DNA bending, are called A tracts. The hydroxyl radical cleavage experiment was used to investigate the structural details of this highly curved DNA [29]. The hydroxyl radical cleavage pattern of the kinetoplast DNA fragment was found to be strongly modulated, in remarkable contrast to the smooth cleavage pattern previously seen for mixed-sequence, noncurved DNA. In Fig. 3 we compare the cleavage pattern of a synthetic bent DNA sequence, containing repeats of T5A5, with the cleavage pattern of a straight DNA molecule having repeats of T4A4N2 [30]. The
< previous page
page_461
next page >
< previous page
page_462
next page > Page 462
Fig. 3. Hydroxyl radical cleavage pattern of a straight TnAn sequence (T4A4N2) (top panel) and a bent DNA sequence (T5A5) (bottom panel). Whereas cleavage of the straight DNA molecule is relatively even at all nucleotide positions, the bent DNA sequence exhibits decreased cleavage within the A tract, giving rise to a sinusoidal cleavage pattern that is characteristic of bent DNA molecules.
< previous page
page_462
next page >
< previous page
page_463
next page > Page 463
strong modulation of the cleavage pattern of bent DNA is readily apparent. A closer look at the cleavage pattern of bent DNA reveals that the intensity of cleavage decreases smoothly from the 5' to the 3' end of each A tract and then increases to a normal level in the intervening segments between A tracts (Fig. 3). This periodic decrease and increase in cleavage gives the pattern the appearance of a sine wave. Examination of the cleavage pattern of both strands of bent kinetoplast DNA [29] revealed that there was a two-nucleotide offset in the 3' direction between nucleotides with comparable cleavage intensity levels on the two strands. The combined presence of the strand offset and decreased cleavage is indicative of a markedly narrow minor groove in the adenine tracts. This unusual cleavage pattern also was observed when a single A5 tract within a mixed sequence was examined [29]. When the cleavage was performed at elevated temperature, the A5 tract was cleaved with the same efficiency as the surrounding sequence. It therefore appears that hydroxyl radical is a sensitive probe of unusual DNA structure. A detailed study of several A-tract sequences yielded results which led to a new model for DNA bending [30]. In these experiments the hydroxyl radical cleavage patterns of sequences of the form 5'TnAn3' were found to be different from those of simple An (A-tract) sequences. It was found that, depending on the value of n, a TnAn sequence may adopt either a B-DNA or an A-tract (bent) conformation. An example of two TnAn sequences, one of which is bent and the other straight, is shown in Fig. 3. Consideration of the data for several TnAn sequences led to the proposal that within a T2A2 or T3A3 segment, or, in other words, in the region immediately surrounding the 5'TA3' step in the center of a TnAn sequence, the structure of the DNA is B form. Outside that core region the DNA adopts the unusual A-tract structure, providing the value of n is sufficiently large. To test the model, the sequence T7A7N7 was examined. A four-nucleotide DNA sequence was substituted for the T2A2 sequence at the center of each of two T7A7N7 repeats in a synthetic DNA molecule, so that the sequence of the repeating unit became T5N4A5N7. Despite the difference in sequence, the hydroxyl radical cleavage pattern and the anomalous mobility of the fragment on a nondenaturing gel were identical for the two DNA molecules. Fourier analysis of the cleavage pat-
< previous page
page_463
next page >
< previous page
page_464
next page > Page 464
terns revealed the existence of an unusual structure which occurred with a periodicity of 10.5 nucleotides. This finding suggests that periodic repetition of an unusual structure in phase with the DNA helical repeat leads to the pronounced bending observed in DNAs with repeats of T7A7N7. 3.1.3 DNA Junctions During recombination DNA adopts an unusual structure, the four-stranded Holliday junction [31]. Unfortunately, the sequence symmetry of naturally occurring junctions [32,33] makes them difficult to study because the position of strand crossover is not well defined. A solution to this problem came with the design and synthesis of immobile junctions [34] which made available stable DNA junctions for structural studies. The first hydroxyl radical cleavage experiments on the Holliday junction [35] addressed the symmetry of the fourway junction. Several models had been proposed previously for how the arms of the four-way junction are disposed in space, including structures with tetrahedral, fourfold, or twofold symmetry. The hydroxyl radical cleavage pattern of an immobile model four-way junction demonstrated definitively that the structure was at most twofold symmetrical, immediately ruling out proposals for structures having D4h or Td symmetry. The most surprising result of these experiments was the demonstration that there was predominantly a single crossover isomer of the junction present in solution. That is, two of the strands remained helical, while the other two strands crossed over from one helical arm to the other at the junction. This result, the first evidence of stacking preferences in DNA junctions, could have implications for the mechanism of recombination. Guo et al. studied the structure of a three-arm DNA junction and its behavior in the presence of divalent cations [36]. In the absence of Mg2+ the three arms were conformationally equivalent. In the presence of Mg2+ each of the three arms exhibited a unique hydroxyl radical cleavage pattern. Analysis of the cleavage pattern led to the proposal that magnesium ion stabilizes a twofold-symmetrical structure for the three-way junction, with one pair of arms preferentially stacking. The junction would thus resemble a T rather than the Y shape it adopts in the absence of Mg2+. Gel mobility experiments provided additional evidence in support of this model.
< previous page
page_464
next page >
< previous page
page_465
next page > Page 465
Holliday junctions in vivo are believed to exist with parallel helical arms [37], while synthetic junctions seem to prefer antiparallel structures [38,39]. (The terms parallel and antiparallel refer to the relative orientations of the two stacked helices that make up the arms of the junction. A junction is parallel if the strands in each arm that remain helical upon junction formation have the same 5'-to-3' orientation. Parallel junctions are X-shaped; antiparallel junctions resemble the letter ''H." These two types of four-way junction are illustrated at the center of Fig. 4a.) To investigate these structures, four conformationally defined four-way junctions were designed and probed with the hydroxyl radical [40]. These junctions were designed in such a way that two of the four strands of a standard immobile junction were connected by a short stretch of thymidines (either six or nine residues). In principle, which strands were connected would determine whether the junction could form a parallel or antiparallel structure (Fig. 4a, periphery). The proof that this strategy was successful in producing both kinds of junction came from hydroxyl radical cleavage experiments. The hydroxyl radical cleavage patterns showed directly which strands were crossing over in each of the four conformationally restricted junctions. It was no surprise to find from these experiments that antiparallel junctions could be formed, since several previous studies had shown that this configuration represents the "ground state" for the immobile four-way junction in solution. However, even though parallel junctions had long been postulated to be involved in natural recombination reactions, such a configuration of a four-way junction had not previously been detected. Nonetheless the hydroxyl radical cleavage experiments gave unambiguous evidence that DNA can fold into a parallel four-way junction, given sufficient conformational restraint. Through the construction of two monomobile junctions and their comparison to immobile structures, Chen et al. also sought to understand what causes junctions to adopt their unique conformation [41]. The only difference between the two monomobile junctions studied was that one contained G·C and the other A·T base pairs at the branch point. Hydroxyl radical cleavage experiments revealed that for each junction migratory conformers were in equilibrium and the arms of the junction were stacked into two helical domains. Surprisingly, each of the junctions exhibited a clear preference as to which strands formed the cross-
< previous page
page_465
next page >
< previous page
page_466
next page > Page 466
Fig. 4. (a) Four possible twofold symmetrical structural isomers of an immobile four-way junction and the four tethered junctions designed to study these isomers. The letters P and A denote parallel and antiparallel orientation of the helical strands. E or O indicate whether the even- or oddnumbered strands are involved in the crossover. Heavy lines denote odd-numbered strands; the short thymidine tethers are indicated by arcs joining two arms of the junction.
< previous page
page_466
next page >
< previous page
page_467
next page > Page 467
(b) Densitometric scans of hydroxyl radical cleavage patterns of duplex DNA, an antiparallel junction with even strands involved in the crossover (JAE), an antiparallel junction with the odd strands crossing over (JAO), a parallel junction with the odd strands crossing over (JPO), and a parallel junction with evenstrand crossover (JPE). Nucleotides are numbered 5' to 3' such that the branch point always occurs between residues 8 and 9. (Reproduced with permission from Ref. 40.) over. When A·T base pairs were present at the branch point, strands 1 and 3 crossed over, but when G·C pairs were substituted, strands 2 and 4 were the crossover strands. This shows that the identity of crossover strands and the resulting folded structure are dependent on the sequence at the branch point, which may have implications for the activities of resolving enzymes which act on DNA junctions. Footprinting studies of RuvC resolvase bound to a synthetic Holliday junction revealed some sequence specificity of the enzyme [42]. The
< previous page
page_467
next page >
< previous page
page_468
next page > Page 468
junction contained a 12-bp-long homologous region which permitted limited branch migration. The enzyme bound to and distorted the junction, such that an unusual hypersensitivity to hydroxyl radical was observed within the twofoldsymmetrical site at the center of the homologous sequence. The same hypersensitivity also was observed upon RuvC binding to an immobile junction. The lack of an observable footprint implies that the protein may be binding to the DNA and causing the backbone to distort, ultimately leading to a wider minor groove. In contrast, when the hydroxyl radical was used to probe the complex of another resolvase, T4 endonuclease VII, with DNA, a clear footprint was observed [43]. The protection was symmetrical on two diametrically opposed strands and coincided with the sites of enzyme cleavage. The authors concluded that structural distortion of DNA at the branch point of the junction likely is critical for recognition by this enzyme. 3.2 Protein-DNA Complexes 3.2.1 Prokaryotic Repressors Many prokaryotic repressor proteins bind to duplex DNA through a domain composed of an α helix followed by a turn and another α helix, a motif of protein structure dubbed the helix-turn-helix. These proteins regulate transcription by binding as dimers to highly homologous sets of DNA sequences. Hydroxyl radical footprinting experiments have led to models for how repressors recognize and bind to DNA with such high specificity. Two proteins of this family which have undergone footprinting and missing nucleoside analyses are the bacteriophage λ repressor and cro proteins complexed to the operator site OR1[10,21,23]. The hydroxyl radical footprints of the two protein-DNA complexes were similar to each other, revealing two regions of protection separated by a 10- to 11-bp-long region left unprotected [10]. The footprinting patterns directly reflect the dyad symmetry found by crystallography for both the cro-and repressor-DNA complexes. Insight into the mechanism of ''image generation" in hydroxyl radical footprinting was gained by performing solvent accessibility calculations on the X-ray cocrystal structure of the λ repressor-DNA complex [21]. The impressive correspondence of the variation of accessible
< previous page
page_468
next page >
< previous page
page_469
next page > Page 469
surface area within the complex, with the hydroxyl radical footprint, provides strong evidence that the hydroxyl radical footprint gives a quantitative experimental picture of the shape of a DNA-protein complex in solution. Nucleoside-specific interactions between the λ repressor or cro protein and the OR1 binding site were identified using the missing nucleoside experiment [23]. OR1 is one of the six similar but nonidentical binding sites (or operators) in the genome of the bacteriophage λ. The λ operators are 17 bp in length and have approximate inversion symmetry of their sequence. One of the half-sites in each operator is identical (or very nearly so) in all six sites (the consensus half-site), while the other half-site in each diverges from the consensus. The different hierarchies of affinity of cro and repressor for these sites (e.g., repressor binds to OR1 with high affinity, while cro binds better to OR3) forms the basis of the ''genetic switch" which governs the life cycle of phage λ. For both cro and repressor it was found that the removal of nucleosides within the consensus half of the OR1 site most severely affected protein binding. Missing nucleosides in the nonconsensus half-site also were found to affect binding, but not to quite the same extent. Differences observed in the interference signals for λ repressor and cro were proposed to stem from the fact that cro protein, in contrast to λ repressor, does not contain an extended Nterminal arm in addition to the helix-turn-helix motif. The cocrystal structure shows that the arm of λ repressor contacts base pairs around the binding site dyad. The results of the missing nucleoside experiment for λ repressor are consistent with these arm-DNA interactions observed by crystallography at the center of the operator. There are no interference signals in this region for the cro-OR1 complex. Koudelka and coworkers used hydroxyl radical footprinting for detailed studies of the bacteriophage 434 repressor bound to operator sites OR1 and OR3 [44], as well as P22 repressor bound to its operator [45]. Their aim was to better understand how noncontacted bases within a binding site contribute to the affinity of a repressor for DNA. In the case of 434 repressor, the effects of single-base-pair changes in the DNA sequence were monitored by footprinting. The observed differences between footprints led these authors to propose that a single-base-pair change can give rise to a global change in the structure of free DNA and the protein-DNA complex. (As an aside, the data were less
< previous page
page_469
next page >
< previous page
page_470
next page > Page 470
clear when the complexes were probed with the reagent bis(1,10-phenanthroline)copper(I)). Results from similar experiments performed on the P22 repressor-operator complex [45] led to the proposal of a binding model similar to that of 434 repressor. 3.2.2 Homeodomains Homeodomain-containing proteins are instrumental in the development of a multitude of organisms [46]. The homeodomain is a highly conserved domain of 60 amino acids which has some sequence homology to the bacterial repressors. The homeodomain-containing proteins thus compose a family of DNA-binding proteins that utilizes the helix-turn-helix motif. Many of these proteins prefer to bind to A/T-rich sequences, often containing a 5'TAAT3' site. However there are examples to the contrary. In particular, the MAT α2 homeodomain binds DNA as a dimer but does not require a 5'TAAT3' site. MAT α2 protects three regions from hydroxyl radical cleavage within each halfsite [47]. Two of the protected regions are across the major groove from each other, and the third occurs across the adjacent minor groove. This protection pattern is consistent with the cocrystal structure of MAT α2 [48], which shows the second helix of the helix-turn-helix motif binding to the major groove and an N-terminal arm reaching into the adjacent minor groove. Very similar footprints are obtained when three other homeodomains, Engrailed, Deformed, and Ultrabithorax, are complexed with DNA [49]. All three of these proteins prefer a DNA binding site which contains the 5'TAAT3' sequence, and they bind as monomers. A logical conclusion that might be made from these footprinting and structural results is that all homeodomains interact with DNA in a similar manner, due to the high degree of homology in the protein and DNA sequences. However, recent missing nucleoside experiments have shown that this is not always true [49]. Three homeodomains gave different interference signals in the minor groove region of the same DNA binding site. This result was interesting because it has long been thought that proteins are unable to distinguish an A·T from a T·A base pair in the minor groove [50]. It appears from these missing nucleoside data that the N-terminal arm of the homeodomain, which is thought to contact the TAAT sequence in the minor groove, is able to interact selectively
< previous page
page_470
next page >
< previous page
page_471
next page > Page 471
with A and T nucleosides in that groove. Mutagenesis experiments, on the other hand, had indicated that the α helix of the homeodomain that is bound in the major groove controls DNA recognition [51]. A model that accommodates both the missing nucleoside and mutagenesis data postulates that the interactions of the homeodomain N-terminal arm in the minor groove dictate the orientation of the α helix which binds in the adjacent major groove (Fig. 5) [49]. 3.2.3 Zinc Fingers Proteins containing zinc fingers interact differently with DNA than do helix-turn-helix proteins. This structural difference is evident in their hydroxyl radical footprints.
Fig. 5. Model for DNA recognition by a homeodomain. The idealized homeodomain-DNA complex shown is based on the cocrystal structure of an engrailed homeodomain-DNA complex [48]. The shaded base pairs near the N-terminal arm of the protein (bottom) indicate the nucleotide positions which experience differential interactions depending on the identity of the homeodomain. These differential interactions of the homeodomain arm lead to different orientations of helix 3 (the ''recognition helix") in the major groove (center) and thereby mediate differential recognition of the shaded base pairs at the top of the major groove. (Reproduced with permission from Ref. 49.)
< previous page
page_471
next page >
< previous page
page_472
next page > Page 472
An extensively studied zinc finger protein, by both hydroxyl radical footprinting and missing nucleoside analysis, is transcription factor IIIA (TFIIIA) from Xenopus. TFIIIA contains nine zinc fingers, in each of which the metal is coordinated by two cysteines and two histidines. Hydroxyl radical footprinting experiments revealed three regions of protection in the 5S ribosomal RNA gene which TFIIIA regulates. The first and third footprints each were continuous for one helical turn, while the central footprint showed protection on one side of the DNA [52,53]. Footprints of deletion mutants of TFIIIA revealed the position of each zinc finger within the 50-bp-long binding site [54]. Missing nucleoside experiments with a TFIIIA-DNA complex [53] revealed three groups of interference signals, each spanning approximately one helical turn. The effect of missing nucleosides on binding appeared to be stronger on one strand than the other. The three regions of the binding site giving missing nucleoside signals coincided with the three protected regions found in the hydroxyl radical footprinting ex-
Fig. 6. Model of the TFIIIA-5S DNA complex, based on hydroxyl radical footprinting and missing nucleoside analysis. The approximate position of each of the nine zinc fingers of TFIIIA is shown on a DNA helix representing the intragenic control region of the 5S ribosomal RNA gene of Xenopus. The orientations of fingers 13 and 79 relative to DNA were modeled on the cocrystal structure of the Zif 268-DNA complex [56]. The amino and carboxyl termini of the protein are marked, as are the DNA coding (C) and noncoding (NC) strands. The letters C, IE, and A indicate the positions of the three ''boxes" of the binding site defined by mutagenesis experiments. The numbers above the DNA denote nucleotide positions within the 5S intragenic control region. (Reproduced with permission from Ref. 53.)
< previous page
page_472
next page >
< previous page
page_473
next page > Page 473
periments. The missing nucleoside signals in the center of the binding site exhibited an offset between the two strands indicative of a protein lying nearly parallel to the DNA helical axis. There were no such strand-to-strand offsets for the missing nucleoside patterns at the ends of the binding site. These data were used to construct a model of the TFIIIA-DNA complex in which fingers 13 and 79 bind DNA by wrapping around the major groove, while fingers 46 lay on one side of the DNA (Fig. 6). Although the footprinting data are consistent with a different model which is based on the NMR structure of the ZFY protein [55], the combined protection and interference data best fit the model shown in Fig. 6. The Fpg protein contains a single zinc binding site in which the metal is coordinated by four cysteines. Fpg functions as an excision repair enzyme and it therefore prefers to bind to abasic sites in DNA. Footprinting studies of Fpg protein bound to double-stranded DNA containing a simulated abasic site showed that the protein protects a region of five nucleotides centered around the damaged base [57]. The footprint was apparent only on the strand containing the abasic site. However, the protein appeared to need a double-stranded substrate for recognition because it could not bind single-stranded DNA. 3.2.4 The Nucleosome In eukaryotic nuclei, DNA is compacted into chromosomes. An initial step in the compaction process is the formation of the nucleosome, in which approximately 160 base pairs of DNA wrap in two superhelical turns around a core of eight histone proteins. In order to better understand the structure of DNA in the nucleosome, footprinting experiments were carried out [28,5861]. The hydroxyl radical footprint of a single nucleosome formed on the 5S ribosomal RNA gene of Xenopus demonstrated that the helical twist of the DNA is not constant throughout the core particle and moreover is different from the helical twist of the same DNA free in solution [28]. These experiments showed that the helical twist is 10.7 bp per turn for three helical turns of DNA around the center of the nucleosome. The helical twist is different, 10.05 bp per turn, in the remainder of the nucleosomal DNA. In contrast, the helical periodicity of the same DNA molecule bound to a calcium phosphate precipitate is 10.5 bp per turn. The hydroxyl radical cleavage pattern
< previous page
page_473
next page >
< previous page
page_474
next page > Page 474
also showed that some of the variation in minor groove width observed in free 5S DNA was accentuated upon histone binding. The footprinting data were consistent with the notion that DNA bends toward and compresses the minor groove when the nucleosome is formed. Hydroxyl radical footprinting experiments showed that the 5S nucleosome is remarkably stable over a wide range of temperatures and salt concentrations. The structure of the 5S nucleosome was examined at sodium chloride concentrations ranging from 0 to 800 mM, and for temperatures from 0 to 75°C [59,60]. No major changes in the structure of the complex were seen, particularly at the center of the nucleosome. At the high end of this temperature range some dissociation of the DNA from the nucleosome was observed, a sort of ''fraying" of the ends of the DNA away from the histone octamer. Above 75°C, a hydroxyl radical cleavage pattern identical to that of free DNA was found, showing that the DNA had finally dissociated from the protein core. It is useful to note from these experiments that the hydroxyl radical can be used to study DNA structure under extreme conditions of salt and temperature that would render an enzyme (like DNase I) inactive. While spectroscopic methods (e.g., ultraviolet-visible, circular dichroism) can be used for this sort of experiment, footprinting has the advantage that nucleotide-specific structural information is obtained. So, while spectroscopic measurements might show that 30% of the DNA in the nucleosome has dissociated at some elevated temperature, the hydroxyl radical footprint shows precisely which segments of the nucleosome have "melted" away from the histone core and which are still bound. In vivo the histone proteins in nucleosomes become acetylated, but it was not clear how this modification affects nucleosome structure. Hydroxyl radical footprinting studies have demonstrated that acetylation does not change the helical repeat of bound DNA or the existing protein-DNA contacts [58]. It does appear, however, that there is a reduction in the number of times the DNA wraps around the histone core, and therefore it was proposed that the role of acetylation is to change the geometry of the core particles. Some proteins, of which HMG-14/-17 are examples, appear to specifically recognize and bind to nucleosomes. Hydroxyl radical footprinting studies of the interactions of HMG-14/-17 with a nucleosome [62] led to the conclusion that these proteins contact DNA at its entry and exit points from the nucleosome, as well as in the major grooves adjacent to
< previous page
page_474
next page >
< previous page
page_475
next page > Page 475
the nucleosomal dyad axis. Since there are two molecules of HMG-14/-17 bound to each nucleosome, their N termini bind to DNA 2030 bp from the ends of the nucleosome. The proteins then loop under the bound DNA and emerge near the dyad, on the surface of the central DNA strand. Binding of HMG-14/-17 does not appear to require rearrangement of the structure of the nucleosome core particle. 3.2.5 Transcription Complexes High-resolution structural studies of transcription complexes have been limited due to the large size of such complexes. Solution studies, using hydroxyl radical footprinting, have revealed structural details of transcriptional complexes poised at different stages of transcription. During initiation of transcription, the footprint of E. coli RNA polymerase bound to the promoter changes as the complex is transformed from the closed complex, through an intermediate, to the open complex [63]. In the closed complex, the polymerase protects approximately 50 bp of DNA with a three-nucleotide offset between strands. This footprint shows directly that the polymerase protects one side of the DNA template. The region that is protected extends to 74 bp in the intermediate and open complexes, and includes the region of the promoter that melts upon initiation of transcription. The additional region of protection is continuous on both strands and implies that the polymerase wraps around the nucleic acid template in this region. There are some differences between the footprints of the intermediate and open complexes. A nine-nucleotide-long segment on the coding strand centered around the 4 position (relative to the transcription start site) is fully protected in the intermediate complex but is accessible in the open complex. The degree of accessibility changes with temperature and may reflect conformational changes in the complex due to strand separation. From these experiments it appears that the ''transcription bubble" is at least nine nucleotides in length. The authors proposed a model which shows the DNA template bending as the polymerase wraps around it. The observed difference in hydroxyl radical accessibility in the open complex implies that the DNA may be able to rotate within the complex with RNA polymerase to relieve stress caused by strand separation upstream. Footprinting experiments have elucidated how the KorB protein,
< previous page
page_475
next page >
< previous page
page_476
next page > Page 476
which interacts with transcription complexes, functions to repress transcription. KorB interacts differently with its various DNA binding sites [64,65]. When bound to certain sequences it deforms the DNA backbone, and sites hypersensitive to hydroxyl radical are produced [65]. KorB binds to a promoter simultaneously with RNA polymerase, leading to the hypothesis that KorB represses transcription by preventing the transition from a closed to an open complex [64]. Other footprinting experiments have sought to identify pause sites during transcription elongation and to determine how the structure of the elongation complex differs from the structure of the initiation complexes [66]. For example, the region of DNA that is protected in the hydroxyl radical footprint of an elongation complex is only about 35 bp long and is centered around the pause site. This change in footprint signals major structural differences between initiation and elongation complexes. Hydroxyl radical probing of the enzyme HIV reverse transcriptase bound to template-primer DNA complexes at several stages during initiation and elongation revealed that the polymerase interacts with approximately 18 bases of template and 15 bases of primer [67]. The footprints were similar for several different complexes in which the polymerase was made to pause at different points. Therefore the interactions of HIV reverse transcriptase with its nucleic acid substrate are likely independent of sequence. When reverse transcriptase-associated RNase H is present, a conformational change is induced in the template strand which makes the sugars more accessible to hydroxyl radical. A similar phenomenon is thought to occur when E. coli RNA polymerase is bound to DNA in the open complex and enhanced cleavage of the transcription bubble is observed [63]. This increase in accessibility of the nucleic acid backbone was proposed to be indicative of DNA unwinding and strand separation. 3.2.6 Other Protein-DNA Complexes Many other protein-DNA complexes have been studied using hydroxyl radical footprinting [6876]. An unusual example is the enzyme DNA gyrase [77]. The footprint of gyrase is quite large, 128 bp, and remarkably periodic. When the enzyme was induced to cause a single round of
< previous page
page_476
next page >
< previous page
page_477
next page > Page 477
DNA supercoiling, a different footprint was seen. This conformational change in the gyrase-DNA complex has implications for the mechanism of gyrase-mediated DNA supercoiling. 3.3 Drug-DNA Complexes Hydroxyl radical footprinting has proven useful not only for studying large molecular complexes, such as proteins bound to DNA, but also as a sensitive probe of small molecules, like drugs, bound to DNA. One example is calicheamicin, a potent antibiotic that binds in the minor groove of DNA and makes double-stranded cuts. This natural product consists of an aryl-linked oligosaccharide chain, which imparts DNA binding specificity, and an enediyne moiety which effects DNA cleavage. The drug prefers to bind to pyrimidine-rich sequences, protecting four to five nucleotides on both strands from attack by the hydroxyl radical [78]. The footprint of calicheamicin is consistent with the drug binding in an extended configuration along the DNA minor groove. The NMR structure of the calicheamicin-DNA complex closely matches the footprint-derived model [79]. The complex of distamycin with DNA also has been studied extensively by hydroxyl radical footprinting [8082]. This drug binds in the DNA minor groove, typically in A/T-rich sequences. Footprinting of the complex of distamycin with the 5S ribosomal RNA gene [80] showed that the drug binds to the same DNA sequences which bend when this DNA molecule wraps around the histone octamer to form a nucleosome. Distamycin also binds to Atract DNA, protecting approximately 7 bp from radical attack, including the sequence A5 [81]. At low temperature it appears that the A tract adopts two different conformations, but distamycin binding inhibits the transition between these conformations. The high sensitivity of the hydroxyl radical as a structural probe was apparent when the footprints produced by netropsin and distamycin were compared [82]. Both drugs bind to similar DNA sequences but give different patterns of protection. The binding of two intercalators, actinomycin [80] and nogalamycin [83], to DNA has been studied by hydroxyl radical footprinting. Actinomycin was found to protect the binding site only to a small degree
< previous page
page_477
next page >
< previous page
page_478
next page > Page 478
from radical attack, although it did distort the DNA upon binding, leading to hypersensitivity to the hydroxyl radical [80]. 3.4 RNA Structure and RNA-Protein Complexes The tertiary structures of a fragment of the Tetrahymena thermophila ribozyme and several of its mutants have been probed by the hydroxyl radical to determine the role of the various structural elements of the ribozyme [84,85] in producing the folded structure. A part of the catalytic core of the ribozyme (called P3P7) was found to be destabilized upon the removal of sequences near the 3' terminus of the RNA. These deletions, however, did not affect folding of the P4P6 region within the core. Three mutations were found which did destabilize the folding of P4P6 and which require a higher concentration of Mg2+ for appropriate folding of P3P7. Therefore the presence of the P4P6 structure and the 3' peripheral elements are required for the stable formation of the catalytic core [84]. There are small structural features within the P4P6 element which affect the stability of this domain [85]. For example, the existence of a GAAA tetraloop is crucial. A single-base change within this loop alters the local hydroxyl radical protection pattern as well as the protection pattern observed in another region of the P4P6 element. Mutations in two other regions within the P4P6 element, an A-rich bulge and a base pair in segment P6a, also were found to affect folding. It was concluded from these studies that remote elements in the secondary structure of the RNA are linked upon folding into the tertiary structure. The structure of the guanosine binding site in the Tetrahymena ribozyme was elucidated by tethering Fe(II)-EDTA (or DTPA) to GMP [86]. Affinity cleavage of the ribozyme by the metal-GMP complex required the presence of Mg2+, implying that the RNA molecule needs to be properly folded to act as a ribozyme. The observed sites of cleavage by the metal-based affinity probe are consistent with the Michel-Westhof model for the tertiary structure of the active site of the Tetrahymena ribozyme. Hydroxyl radical footprinting has been used to examine the complex of tRNA and mRNA with the ribosome [87,88]. Specific regions of the two RNAs were protected to different degrees upon complexation.
< previous page
page_478
next page >
< previous page
page_479
next page > Page 479
These experiments showed which segments of RNA interact with the ribosome and how bound tRNA affects an mRNA-ribosome complex. 4 Protein-Tethered Iron Complexes Iron complexes have been tethered to several proteins to study the mode of binding of that protein to DNA [89,90], to study the interaction of its catalytic domain with the nucleic acid substrate [91], or to create sequence-specific cleavage reagents [92]. The tethered metal complexes produce hydroxyl radicals which cleave DNA only at those backbone positions which are proximal to the complex [89,91]. These sites of cleavage, then, indicate which part of the protein is located near the affected positions. 5 Conclusions The hydroxyl radical, produced by combining [Fe(EDTA)]2 with H2O2, has proved to be a powerful probe for studying nucleic acid structure and ligand-nucleic acid complexes. Because of the lack of sequence specificity and the small size of the hydroxyl radical, detailed structural information about the system being studied has been obtained. The work described in this chapter illustrates how the hydroxyl radical has been used to study a wide variety of nucleic acid structures and complexes since its introduction as a probe of DNA a decade ago. Abbreviations
cro
control of repressor and other things
DTPA
diethylenetriaminepentaacetic acid
EDTA
ethylenediaminetetraacetic acid
Fpg
formamidopyrimidine-DNA-glycosylase (fapy-DNA- glycosylase)
GMP
guanosine 5'-monophosphate
HIV
human immunodeficiency virus
< previous page
page_479
next page >
< previous page
page_480
next page > Page 480
HMG
high-mobility group
KorB
kil-override B
NMR
nuclear magnetic resonance
RuvC
designated name of the protein (not an abbreviation)
TFIIIA
transcription factor IIIA
References 1. D. J. Galas and A. Schmitz, Nucl. Acids Res., 5, 3157 (1978). 2. A. D. Johnson, B. J. Meyer, and M. Ptashne, Proc. Natl. Acad. Sci. USA, 76, 5061 (1979). 3. M. Noll, Nucl. Acids Res., 1, 1573 (1974). 4. D. C. Straney and D. M. Crothers, J. Mol. Biol., 193, 279 (1987). 5. D. C. Straney and D. M. Crothers, Cell, 51, 699 (1987). 6. C. Wu, Nature, 317, 84 (1985). 7. P. E. Nielsen, J. Mol. Recog., 3, 1 (1990). 8. D. S. Sigman, T. W. Bruice, A. Mazumder, and C. L. Sutton, Acc. Chem. Res., 26, 98 (1993). 9. H. J. H. Fenton, J. Chem. Soc., 65, 899 (1894). 10. T. D. Tullius and B. A. Dombroski, Proc. Nat. Acad. Sci. USA, 83, 5469 (1986). 11. T. D. Tullius and B. A. Dombroski, Science, 230, 679 (1985). 12. S. Udenfriend, C. T. Clark, J. Axelrod, and B. B. Brodie, J. Biol. Chem., 208, 731 (1954). 13. W. K. Pogozelski, Mechanistic Studies of Hydroxyl Radical-Induced DNA Cleavage, PhD Dissertation, The Johns Hopkins University, 1994. 14. W. K. Pogozelski, T. J. McNeese, and T. D. Tullius, J. Am. Chem. Soc., 117, C428 (1995). 15. J. J. Hayes, L. Kam, and T. D. Tullius, Meth. Enzymol., 186, 545 (1990). 16. J. C. Wu, J. Kozarich, and J. Stubbe, J. Biol. Chem., 258, 4694 (1983). 17. R. P. Hertzberg and P. B. Dervan, Biochemistry, 23, 3934 (1984).
< previous page
page_480
next page >
< previous page
page_481
next page > Page 481
18. G. E. Shafer, M. A. Price, and T. D. Tullius, Electrophoresis, 10, 397 (1989). 19. J. Stubbe and J. W. Kozarich, Chem. Rev., 87, 1107 (1987). 20. J. S. Bashkin and T. D. Tullius in Footprinting of Nucleic AcidProtein Complexes (A. Revzin, ed.), Academic Press, San Diego, 1993, p. 75. 21. W. J. Dixon, J. J. Hayes, J. R. Levin, M. F. Weidner, B. A. Dombroski, and T. D. Tullius, Meth. Enzymol., 208, 380 (1991). 22. M. A. Price and T. D. Tullius, Meth. Enzymol., 212, 194 (1992). 23. J. J. Hayes and T. D. Tullius, Biochemistry, 28, 9521 (1989). 24. A. Brunelle and R. F. Schleif, Proc. Natl. Acad. Sci. USA, 84, 6673 (1987). 25. D. W. Celander and T. R. Cech, Biochemistry, 29, 1355 (1990). 26. R. V. Prigodich and C. T. Martin, Biochemistry, 29, 8017 (1990). 27. W. K. Pogozelski, unpublished results. 28. J. J. Hayes, T. D. Tullius, and A. P. Wolffe, Proc. Natl. Acad. Sci. USA, 87, 7405 (1990). 29. A. M. Burkhoff and T. D. Tullius, Cell, 48, 935 (1987). 30. M. A. Price and T. D. Tullius, Biochemistry, 32, 127 (1993). 31. R. Holliday, Genet. Res., 5, 282 (1964). 32. S. J. Kim, P. Sharp and N. Davidson, Proc. Natl. Acad. Sci. USA, 69, 1948 (1972). 33. B. J. Thompson, M. N. Camien and R. C. Warner, Proc. Natl. Acad. Sci. USA, 73, 2299 (1976). 34. N. R. Kallenbach, R.-I. Ma, and N. C. Seeman, Nature, 305, 829 (1983). 35. M. E. A. Churchill, T. D. Tullius, N. R. Kallenbach, and N. C. Seeman, Proc. Natl. Acad. Sci. USA, 85, 4653 (1988). 36. Q. Guo, M. Lu, M. E. A. Churchill, T. D. Tullius, and N. R. Kallenbach, Biochemistry, 29, 10927 (1990). 37. M. S. Meselson and C. M. Radding, Proc. Natl. Acad. Sci. USA, 72, 358 (1975). 38. J. P. Cooper and P. J. Hagerman, J. Mol. Biol., 198, 711 (1987). 39. D. R. Duckett, A. I. H. Murchie, S. Diekmann, E. von Kitzing, B. Kemper and D. M. J. Lilley, Cell, 55, 79 (1988).
< previous page
page_481
next page >
< previous page
page_482
next page > Page 482
40. A. S. Kimball, Q. Guo, M. Lu, R. P. Cunningham, N. R. Kallenbach, N. C. Seeman, and T. D. Tullius, J. Biol. Chem., 265, 6544 (1990). 41. J.-H. Chen, M. E. A. Churchill, T. D. Tullius, N. R. Kallenbach, and N. C. Seeman, Biochemistry, 27, 6032 (1988). 42. R. J. Bennett, H. J. Dunderdale, and S. C. West, Cell, 74, 1021 (1993). 43. C. A. Parsons, B. Kemper, and S. C. West, J. Biol. Chem., 265, 9285 (1990). 44. A. C. Bell and G. B. Koudelka, J. Mol. Biol., 234, 542 (1993). 45. L. Wu and G. B. Koudelka, J. Biol. Chem., 268, 18975 (1993). 46. M. P. Scott, J. W. Tamkun, and G. W. Hartzell, Biochim. Biophys. Acta, 989, 25 (1989). 47. R. T. Sauer, D. L. Smith, and A. D. Johnson, Genes Dev., 2, 807 (1988). 48. C. Wolberger, A. K. Vershon, B. Liu, A. D. Johnson, and C. O. Pabo, Cell, 67, 517 (1991). 49. A. Draganescu, J. R. Levin and T. D. Tullius, J. Mol. Biol., 250, 595 (1995). 50. N. C. Seeman, J. M. Rosenberg, and A. Rich, Proc. Natl. Acad. Sci. USA, 73, 804 (1976). 51. S. D. Hanes and R. Brent, Science, 251, 426 (1991). 52. M. E. A. Churchill, T. D. Tullius, and A. Klug, Proc. Natl. Acad. Sci. USA, 87, 5528 (1990). 53. J. J. Hayes and T. D. Tullius, J. Mol. Biol., 227, 407 (1992). 54. K. E. Vrana, M. E. A. Churchill, T. D. Tullius, and D. D. Brown, Mol. Cell. Biol., 8, 1684 (1988). 55. M. Kochoyan, T. F. Havel, D. T. Nguyen, C. E. Dahl, H. T. Keutmann, and M. A. Weiss, Biochemistry, 30, 3371 (1991). 56. N. P. Pavletich and C. O. Pabo, Science, 252, 809 (1991). 57. J. Tchou, M. L. Michaels, J. H. Miller, and A. P. Grollman, J. Biol. Chem., 268, 26738 (1993). 58. W. R. Bauer, J. J. Hayes, J. H. White, and A. P. Wolffe, J. Mol. Biol., 236, 685 (1994). 59. H. L. Puhl and M. J. Behe, J. Mol. Biol., 229, 827 (1993). 60. J. S. Bashkin, J. J. Hayes, T. D. Tullius, and A. P. Wolffe, Biochemistry, 32, 1895 (1993).
< previous page
page_482
next page >
< previous page
page_483
next page > Page 483
61. B. D. Thrall, D. B. Mann, M. J. Smerdon, and D. L. Springer, Biochemistry, 33, 2210 (1994). 62. P. J. Alfonso, M. P. Crippa, J. J. Hayes, and M. Bustin, J. Mol. Biol., 236, 189 (1994). 63. P. Schickor, W. Metzger, W. Werel, H. Lederer, and H. Heumann, EMBO J., 9, 2215 (1990). 64. D. R. Williams, V. M. Motallebi, and C. M. Thomas, Nucl. Acids Res., 21, 1141 (1993). 65. D. Balzer, G. Ziegelin, W. Pansegrau, V. Kruft, and E. Lanka, Nucl. Acids Res., 20, 1851 (1992). 66. M. T. Yang and J. F. Gardner, Nucl. Acids Res., 19, 1671 (1991). 67. W. Metzger, T. Hermann, O. Schatz, S. F. J. LeGrice, and H. Heumann, Proc. Natl. Acad. Sci. USA, 90, 5909 (1993). 68. S. H. Shanblatt and A. Revzin, J. Biol. Chem., 262, 11422 (1987). 69. C. Buchman, P. Skroch, W. Dixon, T. D. Tullius, and M. Karin, Mol. Cell. Biol., 10, 4778 (1990). 70. J.-C. Cortay, D. Nègre, M. Scarabel, T. M. Ramseier, N. B. Vartak, J. Reizer, M. H. Saier, Jr., and A. J. Cozzone, J. Biol. Chem., 269, 14885 (1994). 71. D. Lang and T. Stamminger, Nucl. Acids Res., 22, 3331 (1994). 72. I. L. Fink and E. Morkin, J. Biol. Chem., 265, 11233 (1990). 73. V. deLorenzo, F. Giovannini, M. Herrero, and J. B. Neilands, J. Mol. Biol., 203, 875 (1988). 74. A. Scheler and W. Hillen, Mol. Microbiol., 13, 505 (1994). 75. M. Gazeau, F. Delort, M. Fromant, P. Dessen, S. Blanquet, and P. Plateau, J. Mol. Biol., 241, 378 (1994). 76. W. J. Dixon, L. E. Theill, M. Karin, and T. D. Tullius, in preparation. 77. G. Orphanides and A. Maxwell, Nucl. Acids Res., 22, 1567 (1994). 78. S. C. Mah, C. A. Townsend, and T. D. Tullius, Biochemistry, 33, 614 (1994). 79. L. G. Paloma, J. A. Smith, W. J. Chazin, and K. C. Nicolau, J. Am. Chem. Soc., 116, 3697 (1994). 80. M. E. A. Churchill, J. J. Hayes, and T. D. Tullius, Biochemistry, 29, 6043 (1990). 81. J. G. McCarthy and A. Rich, Nucl. Acids Res., 19, 3421 (1991).
< previous page
page_483
next page >
< previous page
page_484
next page > Page 484
82. J. Portugal and M. J. Waring, FEBS Lett., 225, 195 (1987). 83. K. R. Fox, Anti-Cancer Drug Design, 3, 157 (1988). 84. B. Laggerbauer, F. L. Murphy, and T. R. Cech, EMBO J., 13, 2669 (1994). 85. F. L. Murphy and T. R. Cech, J. Mol. Biol., 236, 49 (1994). 86. J.-F. Wang and T. R. Cech, Science, 256, 526 (1992). 87. A. Huttenhofer and H. F. Noller, EMBO J., 13, 3892 (1994). 88. A. Huttenhofer and H. F. Noller, Proc. Natl. Acad. Sci. USA, 89, 7851 (1992). 89. D. P. Mack, J. P. Sluka, J. A. Shin, J. H. Griffin, M. I. Simon, and P. B. Dervan, Biochemistry, 29, 6561 (1990). 90. B. Cuenoud and A. Schepartz, Science, 259, 510 (1993). 91. J. M. Mazzarelli, M. R. Ermácora, R. O. Fox, and N. D. F. Grindley, Biochemistry, 32, 2979 (1993). 92. Y. W. Ebright, Y. Chen, R. D. Ludescher, and R. H. Ebright, Bio-conj. Chem., 4, 219 (1993).
< previous page
page_484
next page >
< previous page
page_485
next page > Page 485
16 Nucleic Acid Chemistry of the Cuprous Complexes of 1,10-Phenanthroline and Derivatives David S. Sigman*, Ralf Landgraf, David M. Perrin, and Lori Pearson Department of Biological Chemistry, Department of Chemistry and Biochemistry, and Molecular Biology Institute, University of CaliforniaLos Angeles, Los Angeles, CA 90095-1570, USA
486
1. Introduction
487
2. DNase Activity of 1,10-Phenanthroline Copper
487
2.1. Discovery
489
2.2. Mechanism of 1,10-Phenanthroline Inhibition of DNA Polymerase
490
2.3. Chemical Mechanism of Scission
3. Comparison of the Cleavage Chemistry of the Untargeted 2:1 Complex and Targeted 1:1 Complex
3.1. Diffusibility of Reactive Oxidative Species
3.2. Ligand Dissociation and Hydrogen Peroxide Reactivity
491
491
493
495
3.3. Mechanism of C-1 H Attack
*Address correspondence to this author at the Molecular Biology Institute.
< previous page
page_485
next page >
< previous page
page_486
next page > Page 486 498
4. Binding Specificity of Tetrahedral (OP)2Cu+ and Derivatives for Free DNA
498
4.1. Reaction Preference of (OP)2Cu+ for B-DNA
501
4.2. Specificity of Single-Strand Bulges and Loops in RNA
503
4.3. Specificity for Holiday Junctions
503
4.4. Reactivity with Transcriptionally Active Open Complexes
4.5. Tetrahedral Cuprous Complexes of Neocuproine As Global Inhibitors of Transcription
504
509
5. Conclusion
509
Abbreviations
510
References
1 Introduction The cuprous complex of 1,10-phenanthroline, with its coreactant hydrogen peroxide, oxidatively degrades DNA and RNA by attacking the (deoxy)ribose group [1]. The Specificity of the 2:1 1,10-phenanthroline-cuprous complex reflects the nucleic acid binding properties of this hydrophobic tetrahedral cation [2]. If 1,10-phenanthroline is covalently tethered to a ligand with high affinity for DNA or RNA, cleavage is observed with the 1:1 complex but the specificity primarily reflects the binding affinity of the carrier ligand [1,3]. Untethered to a targeting ligand, the 1:1 complex of phenanthroline-copper is ineffective at cleaving DNA or RNA [4]. This chapter focuses on the binding specificity and reactivity of the 2:1 1,10-phenanthroline-cuprous complex [(OP)2Cu+] (Fig. 1). This tetrahedral complex has proven to have remarkable preference for different nucleic acid structures as well as for stressed DNA structures formed in enzymatic reactions. As will be indicated below, strong evidence that the reaction specificity is a reflection of the binding of the tetrahedral hydrophobic cuprous chelate has been provided by the discovery that the isosteric redox-inert 2:1 2,9-dimethyl-1,10-phenanthroline-cuprous complex interacts at sites which are strongly cleaved by (OP)2Cu+ [57].
< previous page
page_486
next page >
< previous page
page_487
next page > Page 487
Fig. 1. Structure of the 2:1 1,10-phenanthrolinecuprous complex. 2 DNase Activity of 1,10-Phenanthroline-Copper 2.1 Discovery The chemical nuclease activity was discovered during the course of experiments initially carried out to probe the mechanistic role of a zinc ion which had been reported to be an essential component of the active site of E. coli DNA polymerase I, viral and mammalian DNA polymerases, terminal deoxynucleotide transferase, and E. coli RNA polymerase [811]. The initial suggestion that a tightly bound metal ion was essential for catalysis was based on the inhibition of all of these enzyme systems by 1,10-phenanthroline. Subsequent analytical work claimed that zinc ion was present in approximately stoichiometric amounts to E. coli DNA polymerase [10,12,13]. The postulated role for the zinc ion in catalysis was that it activated the 3'-OH group for nucleophilic attack on the αphosphorus of the incoming nucleotide triphosphate [14]. This suggestion was supported by a bioorganic model reaction which demonstrated that the zinc ion complex of 1,10-phenanthroline-2-carbinol was smoothly phosphorylated by ATP via an intermediate complex composed of 1 mol each of zinc ion, 1,10-phenanthroline-2carbinol, and ATP (Fig. 2) [15]. The reaction was enhanced by magnesium and calcium ions, which presumably facilitated pyrophosphate bond cleavage by either coordinating the β,γ-phosphates of ATP and neutralizing the negative charge of
< previous page
page_487
next page >
< previous page
page_488
next page > Page 488
Fig. 2. Zinc ion-catalyzed phosphorylation of 1,10-phenanthroline-2-carbinol. (top) Net reaction. (bottom) Postulated reaction mechanism. the phosphate or by coordinating the α,β-phosphate and promoting incipient monomeric metaphosphate formation. Recent structural and genetic studies of the 3'-exonuclease activity of Pol I by Steitz and colleagues showed that the concerted action of two metal ions is important in this essential proofreading activity [16,17]. According to this mechanistic proposal, one metal ion promotes the formation of hydroxide ion through coordination and another binds to the 3'-hydroxyl group of the penultimate deoxyribose to facilitate its departure as a leaving group. The simple bioorganic system involving 1,10-phenanthroline-2-carbinol provides a precedent for this proposed mechanism by demonstrating that two metal ions functioning in concert, one acting to increase the concentration of a nucleophile and another promoting the departure of a leaving group, can achieve a reaction in aqueous solution that is not observable in the absence of a catalyst.
Initial studies in our laboratory confirmed that 1,10-phenanthroline inhibited E. coli DNA polymerase but the conditions required for
< previous page
page_488
next page >
< previous page
page_489
next page > Page 489
observing the inhibition were bizarre [18,19]. The demonstration of inhibition required the presence of thiols even though control activity did not depend on thiol. Moreover, for several thiols a bell-shaped dependence on thiol concentration was observed. Low concentrations potentiated cleavage; high concentrations abolished it. The shape of the bell curve also depended on thiol structure. For dithiothreitol, the dependence was very steep while for mercaptoethanol it was very gradual. Subsequent work revealed that copper ion was also essential for inhibition, which was much more extensive if the 1,10-phenanthroline was preincubated with the primer/template poly dA-T [19]. This collection of baroque findings was finally rationalized when it was recognized that the 1,10-phenanthroline inhibition of E. coli DNA polymerase had nothing to do with a tightly bound zinc ion. In fact, subsequent studies have demonstrated that E. coli DNA polymerase does not contain zinc ion although this cation can bind to the 3',5'exonuclease catalytic site discussed above [2022]. 2.2 Mechanism of 1,10-Phenanthroline Inhibition of DNA Polymerase The inhibition of polymerase activity can now be entirely attributed to the nucleolytic activity of 1,10-phenanthrolinecopper which proceeds by the kinetic pathway indicated in Fig. 3. The following series of reactions were responsible. 1,10-Phenanthroline, upon addition to a solution containing the primer template poly dA-T, chelated trace amounts of cupric ion present in the incubation mixture which was contributed by the water and various buffer components. We estimate that with doubly deionized water and the high grade of analytical reagents available, the concentration of trace levels of cupric ion is approximately 0.5 to 1.0 µM unless special precautions are taken. The 1,10phenanthroline-cupric complex is then reduced by thiol that is present in the assay mixtures. The cuprous complex is then reoxidized by ambient oxygen to generate hydrogen peroxide via a superoxide intermediate. As long as thiol is present, two essential components of the reaction mixture are present, i.e., 2:1 1,10-phenanthroline-cuprous complex and hydrogen peroxide. The product of the scission reaction proved to be inhibitors of the enzymatic activity [23]. This scheme rationalizes a variety of disparate observations associated with the inhibition of DNA polymerase by 1,10-phenanthroline.
< previous page
page_489
next page >
< previous page
page_490
next page > Page 490
Fig. 3. Kinetic mechanism of the chemical nuclease activity of (OP)2Cu+. For example, the bell-shaped dependence of the inhibition on thiol concentration can be explained in the following way. The ascending limb of the bell shape is attributable to the reducing potential of the thiol; the descending limb is observed with thiols which are effective bidentate ligands that can quench the reaction by successfully competing with 1,10-phenanthroline for limiting amounts of copper. Thiols, which are poor bidentate ligands (e.g., 3mercaptopropionic acid) and cannot compete with 1,10-phenanthroline, do not block scission at high concentrations and exhibit apparent saturation behavior.
2.3 Chemical Mechanism of Scission As would be required by the proposed scheme, other reducing agents are capable of activating scission. In some applications, they are prefer-
< previous page
page_490
next page >
< previous page
page_491
next page > Page 491
able because they are poor ligands and avoid the bell-shaped concentration dependencies alluded to above. Ascorbic acid is a good reductant [2427]. In addition, superoxide generated by xanthine/xanthine oxidase also potentiates the reaction [28]. Superoxide reduces the cupric ion by a one-electron reduction and generates hydrogen peroxide by its spontaneous dismutation. Superoxide can also be generated in a multistep pathway starting with NADH and hydrogen peroxide [29]. The absolute requirement for hydrogen peroxide has been demonstrated by complete inhibition of the cleavage reaction by catalase [30,31]. Under anaerobic conditions, scission is observed only if exogenous hydrogen peroxide is added [4]. The stable products of the DNA cleavage reaction have been established. They include the free bases, 5'- and 3'phosphorylated ends and the oxidized deoxyribose product, and 5-methylene-2-furanone (5-MF) [32,33]. These products indicate that the preferred site of attack is the C-1 hydrogen of the deoxyribose moiety. The reaction scheme outlined in Fig. 4 is consistent with the experiments that led to the discovery of the chemical nuclease activity in the first place. The reason that polymerases are inhibited by 1,10-phenanthroline is that the 3'-phosphorylated ends (as well as minor amounts of 3'-phosphoglycolates) are unproductive for polymerases. As earlier studies with micrococcal nuclease have demonstrated, they are dead-end inhibitors which tie up catalytically active enzyme [34,35]. Direct evidence for this contention is that alkaline phosphatase, which removes the 3'-phosphate termini, relieves inhibition as does exonuclease III, which removes not only the 3'-phosphates but the 3'-phosphoglycolates as well [23]. 3 Comparison of the Cleavage Chemistry of the Untargeted 2:1 Complex and Targeted 1:1 Complex 3.1 Diffusibility of Reactive Oxidative Species Two features of the underlying chemistry of the nucleolytic activity are unclear. The first is the nature of the oxidative species responsible for the initial attack on the deoxyribose; the second is the precise structure of the noncovalent intermediate formed between the tetrahedral coordination complex and its nucleic acid targets. Several lines of evidence indicate that the reactive intermediate is not a freely diffusible hy-
< previous page
page_491
next page >
< previous page
page_492
next page > Page 492
Fig. 4. Reaction mechanism for the cleavage of the phosphodiester backbone by (OP)2Cu+.
< previous page
page_492
next page >
< previous page
page_493
next page > Page 493
droxyl radical. First, ferrous EDTA, which does proceed by a diffusible hydroxyl radical, gives equal yields of the 3'monophosphate termini and 3'-phosphoglycolates. The nuclease activity of 1,10-phenanthrolinecopper does not. Only minor amounts of the 3'-glycolate, diagnostic of C-4 oxidative attack, are observed. Second, the analysis of the redox-active species by physical organic techniques indicated that the redox species formed by (OP)2Cu+ is at least 104 times less reactive than hydroxyl radicals formed by pulse radiolysis. Additional studies have indicated that Cu+ and H2O2, although formally capable of Fenton chemistry, also do not generate diffusible hydroxyl radicals [36,37]. Finally, the cleavage reaction demonstrates sequence-dependent scission when B-DNA is used as a substrate. This strongly implies that a binding event is essential in the overall reaction. If reactive species are generated from solution, scission would be equivalent at all sequence positions as is observed with ferrous EDTA. The importance of a noncovalent intermediate is also consistent with the lack of nucleolytic activity of the cuprous complex of negatively charged phenanthrolines such as the 5'-succinylamido-1,10-phenanthroline [38,39]. Their redox properties are comparable to nucleolytic-competent chelates but they do not form stable noncovalent intermediates because of the electrostatic repulsion with the anionic phosphodiester backbone. Although freely diffusible hydroxyl radicals appear not to be involved in the scission chemistry, the precise chemical structure of the proximal oxidative species responsible for cleavage is not known. This question must be considered in the light of our studies which have demonstrated that if 1,10-phenanthroline is linked to a DNA ligand (e.g., protein or nucleic acid), the 1:1 copper complex efficiently cleaves DNA by a reaction mechanism which depends on both reducing agent and hydrogen peroxide [3]. The possibility that the same mechanism of phosphodiester backbone scission is operative with untargeted (OP)2Cu+ and the 1:1 complex linked to a high-affinity ligand is very appealing. 3.2 Ligand Dissociation and Hydrogen Peroxide Reactivity Recent studies on the stability of the 2:12,9-dimethyl-1,10-phenanthrolinecuprous complex [(NC)2Cu+] in the presence of competitive ligands have supported the likelihood of a common pathway. Under aerobic conditions, the absorption spectrum of (NC)2Cu+ is stable on the time
< previous page
page_493
next page >
< previous page
page_494
next page > Page 494
scale of a week at physiological pH and temperature. In contrast, the redox-active (OP)2Cu+ is rapidly oxidized to the cupric complex via a superoxide intermediate [28]. Since superoxide generated by xanthine-xanthine oxidase can reduce the (OP)2Cu2+ and activate DNA scission [28], the position of this equilibrium may be poised near unity in aqueous solution and neutral pHs. Upon the addition of 5 mM EDTA, the absorption spectrum characteristic of (NC)2Cu+ disappears within 10 min. The rate at which the absorption spectrum is lost depends on the concentration of the challenging EDTA. Both the (OP)2Cu+ and OP-Cu+ are thermodynamically and kinetically unstable with respect to 5 mM EDTA. The rate of loss of the near-UV absorption signals characteristic of copper chelation by 1,10phenanthroline is too fast to measure using conventional mixing techniques. Monodentate ligands such as chloride, acetate, bicarbonate, nitrite, sulfate, and phosphate labilize (NC)2Cu+ with respect to EDTA. Chloride ion is the most effective anion at accelerating the destruction of (NC)2Cu+ while inorganic phosphate is the least effective. A possible explanation for these findings is that the monodentate ligand binds to the 1:1 complex formed by the dissociation of one of the 2,9-dimethyl-1,10-phenanthrolines. Anions which form the most stable mixed species may destabilize (NC)2Cu+ because this mixed species is vulnerable to losing its copper ion to EDTA (Fig. 5). Support for this mechanism has been obtained by examining the kinetic effect of excess neocuproine added to a reaction mixture containing chloride and EDTA. Since the excess neocuproine competitively blocks the chloride-induced lability, these kinetic results are consistent with a reaction mechanism in which EDTA attacks a copper ligated to one but not two neocuproines. Since the reaction becomes zero order with respect to EDTA at high concentrations, neocuproine dissociation is likely the EDTA-independent step which becomes rate limiting under these conditions.
Fig. 5. Mechanism of the anion-catalyzed lability of (NC)2Cu+ to EDTA.
< previous page
page_494
next page >
< previous page
page_495
next page > Page 495
3.3 Mechanism of C-1 H Attack The relationship of these studies to the chemical nuclease activities of the 2:1 and 1:1 complexes of 1,10phenanthroline and copper is that they suggest a plausible mechanism for the oxidation of the cuprous complexes by hydrogen peroxide. If hydrogen peroxide acts in a manner analogous to that of the anionic ligands of Fig. 5, the kinetic pathway of Fig. 6 could account for the similarity of the scission chemistry of the untargeted (OP)2Cu+ and the targeted 1:1 OP-Cu+ since they are funneled through the same intermediate. Analogous with the anion-catalyzed labilization of (NC)2Cu+, the scheme suggests that one OP ligand can dissociate from the DNA-bound (OP)2Cu+ complex. Just like chloride ion, hydrogen peroxide, acting as a monodentate ligand, can form a mixed complex composed of OP, Cu+, and hydrogen peroxide. A chemically identical intermediate is possible for the 1:1 complex. The two methods of DNA scission therefore converge to a common intermediate composed of 1 mol of OP, 1 mol of copper ion, and 1 mol of H2O2. Two reactive oxidative species may then be envisioned which do not involve a hydroxyl radical. In the first, a copperoxene species is generated which can abstract a hydrogen atom from the C-1 position and then recombine to form the C1-OH species of Fig. 4. Alternatively, the copper oxene could abstract the hydrogen atom to generate a cupric ion bound hydroxide ion. The cupric ion could then carry out the one-electron oxidation of the C-1-based radical to produce a carbonium ion which in turn would react with water to produce the C1-OH derivative. These possibilities will be distinguished by identifying the atom source of the oxygen in the 5-MF. If the oxygen is derived from water, the mechanism on the left is not excluded. If the oxygen is derived from hydrogen peroxide, the mechanism on the right is not excluded. This kinetic scheme is consistent with a variety of experimental observations that have been made by using (OP)2Cu+ as a footprinting reagent. One issue is the concentration of hydrogen peroxide that should be used to accomplish the conditions of single-hit kinetics essential for footprinting. Generally, our protocol has proved most reliable if the hydrogen peroxide is generated in situ by the (OP)2Cu2+ -catalyzed oxidation of 3-mercaptopropionic acid. If hydrogen peroxide is added at concentrations comparable to the thiol, the DNA is overdigested and no
< previous page
page_495
next page >
< previous page
page_496
next page > Page 496
Fig. 6. Possible mechanisms for the formation of 5-methylene-2-furanone during the cleavage chemistry. (left) Water is the source of C-1 oxygen. (right) Hydrogen peroxide is the source of C-1 oxygen.
< previous page
page_496
next page >
< previous page
page_497
next page > Page 497
useful footprinting data are obtained. In situ generation of hydrogen peroxide appears to be experimentally useful because it makes the oxidation of the cuprous complex rate limiting and allows regulation of the extent of cleavage by alteration of the incubation times. The initial reduction of the cupric complex will not be rate limiting because the thiol is added at millimolar levels and the reduction of cupric complexes generally proceeds at the diffusioncontrolled limit. The use of 3-mercaptopropionic acid as the reductant creates a complexity because the copper ion can be sequestered as the mercaptide complex and inhibit the nucleolytic reaction [27]. Ascorbic acid does not create these problems but the ready autoxidation of ascorbic acid to generate hydrogen peroxide creates a different set of experimental difficulties, namely, a substantial nonspecific background scission rate. The rates of the site-specific scission of DNA using 1,10-phenanthroline-copper tethered to different targeting ligands cannot be directly compared to the untargeted scission achieved with the (OP)2Cu+. Nevertheless both reducing equivalents and hydrogen peroxide are essential reactants for the cleavage chemistry even though only trace amounts of either may be necessary. In one series of experiments, cleavage was observed in a DNA-trp-OP chimera isolated by a mobility shift assay simply upon the addition of copper ion [31]. The possibility that cupric iondependent hydrolysis was observed was excluded by the demonstration that the cleavage reaction could be inhibited by the incubation of the reaction mixture with catalase. Apparently there are sufficient reducing equivalents in the acrylamide matrix to generate hydrogen peroxide and power the cleavage reaction. However, in many examples of targeted scission, addition of hydrogen peroxide does not increase the extent of the reaction. A possible explanation for this observation is that the rate-limiting step for targeted scission is the reduction of cupric ion. Since the concentration of scission reagent ranges from 1 to 100 nM, the concentration of the chelated copper may be very low because the log of the stability constant for cupric binding to 1,10-phenanthroline ranges from 7.5 to 8.5 depending on the reaction conditions and phenanthroline structure. In contrast in untargeted scission, the concentration of chelated copper may be 1000-fold greater since the concentration of copper ion ranges from 1 to 10 µM and that of OP is usually about 50 µM. Added hydrogen peroxide increases this reaction because its reduction, rather than copper ion reduction, is rate limiting.
< previous page
page_497
next page >
< previous page
page_498
next page > Page 498
4 Binding Specificity of Tetrahedral (OP)2Cu+ and Derivatives for Free DNA 4.1 Reaction Preference of (OP)2Cu+ for B-DNA Despite the fact that the 2:1 complex may react with hydrogen peroxide following the dissociation of 1 mol of OP, the untargeted tetrahedral complex has shown a remarkable series of reaction preferences. One of the first discovered is its preference for B-DNA. A-DNA is cleaved at a substantially slower rate. Z-DNA is unreactive and singlestranded DNA is uncleaved if it is blocked from forming secondary structure by chemical modification with glyoxal [38,40,41]. Since all the chemical linkages of the B-DNA are the same, the inescapable explanation for the reactivity preference is that B-DNA provides a more favorable binding site for the tetrahedral cuprous complex (Fig. 7, left). The minor groove in A-DNA is shallow with a less negative electrostatic potential because the phosphodiester backbone is oriented toward the major
Fig. 7. Docking model of the binding of (OP)2Cu+ to B-DNA and A-DNA. (left) B-DNA. (right) A-DNA.
< previous page
page_498
next page >
< previous page
page_499
next page > Page 499
groove (Fig. 7, right). No favorable binding is possible near the redox-sensitive minor groove in the left-handed ZDNA structure since there is no concave binding site for the chelate. The preference of (OP)2Cu+ for single-stranded DNA relative to double-stranded B-DNA has been demonstrated in a graphic manner by the following experiment. Two complementary DNA strands were prepared whose ends permitted the formation of a hairpin structure containing single-stranded DNA loops [40]. The sequence also allowed the annealing of the two strands to form a normal B-DNA duplex (Fig. 8).
Fig. 8. Relative reaction preference of (OP)2Cu+ for single-stranded and double-stranded DNA. Solid dots represent strong sites of scission by (OP)2Cu+; open circles, weak sites of scission; no mark, negligible cleavage.
< previous page
page_499
next page >
< previous page
page_500
next page > Page 500
In the hairpins the duplex regions were more reactive than the single-stranded regions. However, when the two complementary hairpins were hybridized, these sequences, now in a B-DNA duplex, were readily cleaved. Since the chemical stability of the DNA is obviously unchanged, this dramatic shift in reactivity preference must be associated with formation of effective binding sites near the oxidatively sensitive deoxyribose. Visualization of the noncovalent intermediate formed between B-DNA and (OP)2Cu+ provides some insight into the unique reactivity of this helical form of DNA. As noted above, the relatively narrow minor groove provides suitable docking sites for (OP)2Cu+. Based on the reactivity of the cuprous chelates of a series of 1,10-phenanthroline derivatives, the 5-substituents generally do not interfere with the reactivity of the chemical nuclease. This observation tends to exclude the intercalation of the middle ring of the phenanthroline into the double-helical DNA. This conclusion is contradicted by equilibrium measurements of Rill and colleagues who obtained experimental evidence for intercalation [27]. Although these complexes may form, they may not be on the reaction pathway. In contrast, substituents at the symmetrical 2 and 9 positions completely block cleavage. Furthermore, methyl groups at the equivalent 3 and 8 positions severely inhibit scission. Are these specificity data consistent with the previous discussion, which suggests that one of the phenanthroline ligands may dissociate prior to the ligation of hydrogen peroxide and subsequent redox chemistry? Molecular modeling is certainly consistent with this view because it is difficult to dock a tetrahedral chelate to the minor groove of B-DNA. Although one of the rings can comfortably reside within the minor groove, the second phenanthroline does not enter the minor groove without steric hindrance. Indeed, the substituents at the 2 and 3 positions of this phenanthroline block the copper from approaching within 34 Å of the C-1 hydrogen. Dissociation of this second phenanthroline relieves this steric constraint and would allow the close approach of the copper to the oxidatively sensitive bond. However, if this reaction mechanism is operative, why is the 1:1 complex an ineffective chemical nuclease unless it is tethered to a high-affinity DNA ligand? One possibility is that the 1:1 OP-Cu+ complex, like its close structural relative ethidium bromide, is an effective intercalating agent and is difficult to reduce to the cuprous complex once bound to the DNA [42,43]. The reason that the
< previous page
page_500
next page >
< previous page
page_501
next page > Page 501
2:1 complex is an efficient chemical nuclease may be that the second phenanthroline blocks the sequestration of the 1:1 OP-Cu+ complex within the duplex DNA. Comparable factors may contribute to the lack of reactivity of (NC)2Cu+. Since its stability constant is 3 orders of magnitude greater than that of (OP)2Cu+ [44], ligand dissociation will be unfavorable and hydrogen peroxide will not be accessible to the cuprous ion. In considering the docking model of (OP)2Cu+ into the minor groove of BDNA, an additional factor in restricting the activity of (NC)2Cu+ may be the pronounced steric hindrance evident between the 2- and 9-methyl groups with the phosphodiester backbone which blocks the close approach of the copper to the C-1 hydrogen. The latter argument can also explain the relatively sluggish nucleolytic activity with B-DNA of the cuprous complex of 3,4,7,8-tetramethyl-1,10-phenanthroline despite the ability of this complex to efficiently catalyze the oxidation of thiols. The methyl groups at the symmetrical 4 and 8 positions also clash with the phosphodiester backbone. 4.2 Specificity of Single-Strand Bulges and Loops in RNA The reactivity of (OP)2Cu+ with RNA provides another example of the binding specificity of this tetrahedral chelate. Based on the previously cited studies with DNA, it would have been expected that (OP)2Cu+ would cleave the double-stranded region of RNA facilely and be unreactive with the loop regions. The results obtained were exactly the opposite [4547]. The single-stranded or nonconventional duplex structures have proved to be the preferred sites of cleavage (Fig. 9). This reaction preference is so reliable that it can be used to audit secondary structure maps of RNA. Such an experiment has been done with the HIV TAR, i.e., the transactivator domain of the human immunodeficiency virus. Cleavage is observed in the wild-type RNA at the regions known from genetic and structural studies to be single stranded [46]. If mutations are inserted in the stem structure which disrupt the helical region of the RNA, new cleavage sites appear. If compensatory mutations are made which restore the integrity of the double helix, these new sites of cleavage are lost.
< previous page
page_501
next page >
< previous page
page_502
next page > Page 502
Fig. 9. Preferred cleavage sites of (OP)2Cu+ in tRNA. Arrows indicate sites of preferential cleavage by (OP)2Cu+; crosses indicate preferred sites of reactivity by MPE-Fe; stars represent sites protected from cleavage by ferrous EDTA. These findings would appear to be inherently contradictory to those obtained with DNA. However, upon closer examination, it should be recognized that the reaction preference for the single-stranded regions of RNA reflects its rate of reaction relative to that with duplex A-structure RNA. In contrast, the reaction of single-stranded DNA with (OP)2Cu+ competes with the much faster reaction of the duplex B-DNA structure. Reference to molecular models, in addition to emphasizing that the tetrahedral complex does not have a good steric fit to the minor groove of the A helix, also indicates that the phosphodiester backbone of
< previous page
page_502
next page >
< previous page
page_503
next page > Page 503
this helical structure is oriented to the major groove. Therefore, the favorable electrostatic field which enhances the affinity of cationic ligands in the case of B-DNA is not present in the minor groove of A-DNA. 4.3 Specificity for Holiday Junctions (OP)2Cu+ has exhibited an unexpected reaction preference for two other DNA structures. The first is for Holiday junctions [48]. These four-stranded structures play an essential role in recombination. (OP)2Cu+ cleaves these structures efficiently presumably because favorable binding sites for the tetrahedral coordination complexes are created [48]. The second unexpected specificity that has been uncovered by our ongoing studies of the untargeted reaction is that (OP)2Cu+ is extremely efficient in cleaving kinetically competent open complexes formed at the initiation of transcription [49]. 4.4 Reactivity with Transcriptionally Active Open Complexes This unusual reactivity was initially recognized in using (OP)2Cu+ as a footprinting reagent in studying the binding of RNA polymerase to the lac UV-5 promoter [49]. The hyperreactivity of (OP)2Cu+ to the open complex formed with E. coli RNA polymerase and the lac UV-5 promoter appears as a series of intense cleavage sites on the template strand just upstream of the start of transcription. The identical cleavage pattern is obtained with the several variants of the lac promoter [49]. The sequence positions cleaved correspond to domains which have been demonstrated to be single stranded using base-specific modification reagents such as dimethyl sulfate [50]. Both prokaryotic and eukaryotic transcription units exhibit this hyperreactivity [5154]. Convincing evidence that these hyperreactive sites arise from the binding of the tetrahedral chelate to the open complex is provided by the confirmation of a prediction intrinsic to the reaction paradigm. Since binding dictates scission, the reaction mechanism demands that coordination complexes which are isosteric to (OP)2Cu+, such as (NC)2Cu+, should inhibit transcription. Indeed, this is exactly what is observed.
< previous page
page_503
next page >
< previous page
page_504
next page > Page 504
(NC)2Cu+ is a global inhibitor of transcription [5]. It inhibits RNA polymerases from simple viruses, prokaryotes, and eukaryotes [7]. Even transcription systems in which hyperreactive sites are not observed are strongly inhibited by this redox-inactive coordination complex. 4.5 Tetrahedral Cuprous Complexes of Neocuproine As Global Inhibitors of Transcription The interaction of (NC)2Cu+ with the open complex formed from E. coli RNA polymerase and the lac UV-5 promoter has been demonstrated in a number of ways [5]. The most direct test of the paradigm is that the (NC)2Cu+ can block the scission of the open complex by (OP)2Cu+ (Fig. 10). This inhibition of the scission reaction is not limited to the coordination complex prepared from the unsubstituted 1,10-phenanthroline. Other phenanthrolines cleave the open complex of this transcription system very efficiently. For example, the cuprous complexes prepared from 5-phenyl- and 4-phenyl-1,10-phenanthroline are especially efficient in cleaving the open complex. The open complex-specific scission is blocked by (NC)2Cu+ as well as the 2,9-dimethyl derivatives of these other redox-active chelates. The data summarized in Fig. 10 are representative of the patterns that have been observed with different pairs of redox-active and redox-inactive chelates. In all cases, the neocuproine derivatives are effective inhibitors of abortive initiation and the synthesis of full-length transcript. Indeed the inhibition of the synthesis of the full-length transcript is more efficient than the inhibition of the initial steps of transcription. In contrast, the 3:1 complex of 1,10phenanthroline and ferrous ion is not inhibitory (Fig. 11). A possible reason for the more potent inhibition of full-length transcription by the cuprous complex of 5-phenyl-2,9dimethyl-1,10-phenanthroline has been provided by monitoring the progression of polymerase using the cuprous complex of 5-phenyl-1,10-phenanthroline as a function of the addition of an incomplete complement of nucleotides to E. coli RNA polymerase and the lac UV-5 operator [55]. With only ATP and UTP or ApA, UTP, and GTP hypersensitive sites are observed upstream from the point of the synthesis of the phosphodiester bond. However, when ATP, UTP, and GTP are added, so that transcription stalls
< previous page
page_504
next page >
< previous page
page_505
next page > Page 505
because of the absence of CTP, a transcript of sequence AAUUGUGAG is obtained. Hypersensitive cleavage sites are no longer observed at positions upstream of the position of phosphodiester bond formation. Instead, cleavage sites are observed at positions 8 and 9, precisely at the site of the catalytic event. This site is protected from cleavage by the addition of the corresponding neocuproine complex. This result suggests that the single-stranded DNA which trails the migrating transcription complex forms a heteroduplex with the newly synthesized RNA. Short ribooligonucleotides do not form stable heteroduplexes and dissociate. The single-stranded DNA remains a reactive target for (OP)2Cu+ and its derivatives. As the polymerase proceeds downstream, the longer RNAs form stable heteroduplexes which are poorer substrates for the chemical nuclease. The only single-stranded DNA present in the elongation complex which forms after the dissociation of the σ subunit is at the leading edge of the transcription bubble. This remains accessible to the chemical nuclease, as evidenced from the downstream cleavage sites. The ability of the corresponding neocuproine complex to block scission indicates that this binding site also has an affinity for its redox-inactive isostere. The lower IC50 observed in full-length transcription must be due to the tight binding of the chelate to this site. Presently it is not known if the cleavage within the open or transcriptionally active complex proceeds via C-1 or C-4 oxidative attack. Either is possible because the products that are formed are too large to allow resolution of 3'phosphoglycolate termini from the 3'-monophosphate esters by electrophoretic methods. Since the sites of cleavage may be entirely governed by the orientation of the chelate within the open complex, deuterium substitution at either the C-1 or C-4 positions should help resolve this ambiguity [56,57]. Most likely the cleavage of the template strand within the open complex proceeds by the same detailed kinetic mechanism as the cleavage of free B-DNA. In that case, one phenanthroline moiety would dissociate prior to ligation of hydrogen peroxide to the cuprous ion. Since the open complex is an asymmetrical environment, two possible 1:1 complexes will be formed upon dissociation of one phenanthroline. The different orientations of these chelates within the open complex could give rise to the multiple sites of cleavage that are then observed.
< previous page
page_505
next page >
< previous page
page_506
next page > Page 506
Fig. 10. (a): Protection of scission of the open complex of the cuprous complex of 4-φ-1,10-phenanthroline by the cuprous complex of 4-φ-2,9-dimethyl-1, 10-phenanthroline. Lane 1: Maxam-Gilbert G+A sequencing ladder. Lanes 2 and 3: DNase footprint in the absence and presence of RNA polymerase, respectively. Lane 4: 4φOP-Cu+ digestion of the lac UV-5 promoter. Lanes 511: 4φOP-Cu+ footprint of the open complex in the presence of increasing concentrations of (4φNC)2Cu+: 1, 3, 10, 30, 60, 100 µM. (φ = phenyl). An extremely interesting feature of the biochemical action of the neocuproine complexes is their cytotoxicity. They kill a wide variety of cell types at lower concentrations than they inhibit the transcription units that we have assayed. For this reason it cannot be assumed that the cytotoxicity necessarily relates to the ability of the chelates to inhibit transcription. A possible mechanism of cytotoxicity is that the hydrophobic cation is actively taken up by mitochondria [58,59]. The ensuing perturbation of calcium homeostasis may lead to activation of endonucleases which may be the proximal cause of cell death. However, ongoing screening of the biochemical activity of the neocuproine complexes has revealed that they also inhibit the integrase of HIV [60].
< previous page
page_506
next page >
< previous page
page_507
next page > Page 507
Fig. 10. (b): Protection of scission of the open complex of the cuprous complex of 5-φ-1,10-phenanthroline by the cuprous complex of 5-φ-2,9-dimethyl-11,10-phenanthroline. Lane 1: Maxam-Gilbert G+A sequencing ladder. Lanes 2 and 3: DNase footprint in the absence and presence of RNA polymerase, respectively. Lane 4: 4φOP-Cu+ digestion of the lac UV-5 promoter. Lanes 510: 5φOP-Cu+ footprint of the open complex in the presence of increasing concentrations of (5φNC)2Cu+: 1, 3, 10, 30, 100 µM. (φ = phenyl).
< previous page
page_507
next page >
< previous page
page_508
next page > Page 508
Fig. 11. Specificity of the open complex for tetrahedral complexes of 1,10-phenanthroline. The tetrahedral complex of cuprous ion, but not the octahedral complex of ferrous ion, inhibits runoff transcription of the 66 nt from the lac UV-5 operon.
< previous page
page_508
next page >
< previous page
page_509
next page > Page 509
This enzyme is essential in viral infections for the integration of duplex DNA copies into the host cell. Analysis of the activities of integrase suggests that the neocuproine complex binds either to the enzymatic catalytic site or to the interface between the enzyme and a noncanonical DNA structure generated during catalysis. These suggest that cell death could be due to the inhibition of DNA remodeling enzymes related in structure to integrase. 5 Conclusion In summary, the nuclease activity of 1,10-phenanthroline copper is of interest from both a chemical and a biological perspective. Although the mechanism of peroxide activation is not yet known, the investigation of the origin of the oxygen atom incorporated into 5-MF should shed useful light on the detailed chemical mechanism. The tetrahedral cuprous complex of 1,10-phenanthroline has a remarkable range of binding specificities. These include (1) the minor groove of DNA, (2) single-stranded RNA, (3) Holiday junctions, and (4) single-stranded DNA formed at the active site of RNA polymerases in actively transcribing complexes. These studies have led to the discovery that cuprous complexes of neocuproine are general inhibitors of transcription. Abbreviations
ADP
adenosine 5'-diphosphate
ATP
adenosine 5'-triphosphate
CTP
cytidine 5'-triphosphate
EDTA
ethylenediamine-N,N,N',N',-tetraacetic acid
GTP
guanosine 5'-triphosphate
HIV
human immunodeficiency virus
IC50
concentration at which inhibition is 50%
lac UV-5 promoter
high-affinity lac UV-5 promoter
5-MF
5-methylene-2-furanone
NADH
reduced nicotinamide adenine dinucleotide
< previous page
page_509
next page >
< previous page
page_510
next page > Page 510
(NC)2Cu+
2:1 2,9-dimethyl-1,10-phenanthroline-cuprous complex
OP
1,10-phenanthroline
OP-Cu+
1,10-phenanthroline copper
(OP)2Cu+
2:1 1,10-phenanthroline-cuprous complex
Pol I
polymerase I
TAR
transactivator domain of HIV
trp
tryptophan EDCBA operon
UTP
uridine 5'-triphosphate
References 1. D. S. Sigman, T. W. Bruice, A. Mazumder, and C. L. Sutton, Acc. Chem. Res., 26, 98 (1993). 2. D. S. Sigman, Biochemistry, 29, 9097 (1990). 3. C. Q. Pan, R. Landgraf, and D. S. Sigman, Mol. Microbiol., 12, 335 (1994). 4. L. E. Marshall, D. R. Graham, K. A. Reich, and D. S. Sigman, Biochemistry, 20, 244 (1981). 5. A. Mazumder, D. M. Perrin, K. H. Watson, and D. S. Sigman, Proc. Natl. Acad. Sci. USA, 90, 8140 (1993). 6. A. Mazumder, D. M. Perrin, D. McMillen, and D. S. Sigman, Biochemistry, 33, 2262 (1994). 7. D. M. Perrin, L. Pearson, A. Mazumder, and D. S. Sigman, Gene, 149, 173 (1994). 8. L. Chang, and F. Bollum, Proc. Nat. Acad. Sci. USA, 65, 1041 (1970). 9. B. J. Poiesz, G. Weal, and L. A. Loeb, Proc. Natl. Acad. Sci. USA, 71, 4892 (1974). 10. C. F. Springgate, A. S. Mildvan, S. Abramson, J. L. Engle, and L. A. Loeb, J. Biol. Chem., 248, 5987 (1973). 11. J. G. Stavrianpoloulos, J. D. Karkas, and E. Chargaff, Proc. Natl. Acad. Sci. USA, 69, 1781 (1972). 12. J. P. Slater, A. S. Mildvan, and L. A. Loeb, Biochem. Biophys. Res. Commun., 44, 37 (1971).
< previous page
page_510
next page >
< previous page
page_511
next page > Page 511
13. C. F. Springgate, A. S. Mildvan, R. Abramson, J. L. Engle, and L. A. Loeb, J. Biol. Chem., 248, 5987 (1973). 14. A. Mildvan in The Enzymes, Vol. 2 (P. D. Boyer, ed.), Academic Press, New York, 1970, p. 445 ff. 15. D. S. Sigman, G. M. Wahl, and D. J. Creighton, Biochemistry, 11, 2236 (1972). 16. L. S. Beese and T. A. Steitz, EMBO J., 10, 25 (1991). 17. L. S. Beese, V. Derbyshire, and T. A. Steitz, Science, 260, 352 (1992). 18. V. D'Aurora, A. M. Stern, and D. S. Sigman, Biochem. Biophys. Res. Commun., 78, 170 (1977). 19. V. D'Aurora, A. M. Stern, and D. S. Sigman, Biochem. Biophys. Res. Commun., 80, 1025 (1978). 20. D. R. Graham and D. S. Sigman, Inorg. Chem., 23, 4188 (1984). 21. K. E. Walton, P. C. Fitzgerald, M. S. Hermann, and W. D. Behnke, Biochem. Biophys. Res. Commun., 108, 1353 (1982). 22. L. J. Ferrin, A. S. Mildvan, and L. A. Loeb, Biochem. Biophys. Res. Commun., 112, 723 (1983). 23. L. M. Pope, K. A. Reich, D. R. Graham, and D. S. Sigman, J. Biol. Chem., 257, 12121 (1982). 24. J. M. Veal and R. L. Rill, Biochemistry, 27, 1822 (1988). 25. J. M. Veal and R. L. Rill, Biochemistry, 28, 3243 (1989). 26. J. M. Veal, K. Merchant, and R. L. Rill, Nucl. Acids Res., 19, 3383 (1991). 27. J. M. Veal and R. L. Rill, Biochemistry, 30, 1132 (1991). 28. D. R. Graham, L. E. Marshall, K. A. Reich, and D. S. Sigman, J. Am. Chem. Soc., 102, 5419 (1980). 29. K. A. Reich, L. E. Marshall, D. R. Graham, and D. S. Sigman, J. Am. Chem. Soc., 103, 3582 (1981). 30. D. S. Sigman, D. R. Graham, V. D'Aurora, and A. M. Stern, J. Biol. Chem., 254, 12269 (1979). 31. A. Mazumder, C. L. Sutton, and D. S. Sigman, Inorg. Chem., 32, 3516 (1993). 32. M. Kuwabara, C. Yoon, T. E. Goyne, T. Thederahn, and D. S. Sigman, Biochemistry, 25, 7401 (1986). 33. T. E. Goyne and D. S. Sigman, J. Am. Chem. Soc., 109, 2846 (1987).
< previous page
page_511
next page >
< previous page
page_512
next page > Page 512
34. M. P. Deutscher and A. Kornberg, J. Biol. Chem., 244, 3019 (1969). 35. P. T. Englund, R. B. Kelly, and A. Kornberg, J. Biol. Chem., 244, 3045 (1969). 36. G. R. A. Johnson and N. B. Nazhat, J. Am. Chem. Soc., 109, 1990 (1987). 37. G. R. A. Johnson, N. B. Nazhat, and R. A. Saadalla-Nazhat, J. Am. Chem. Soc., 84, 501 (1988). 38. D. S. Sigman, A. Spassky, S. Rimsky, and H. Buc, Biopolymers, 24, 183 (1985). 39. T. B. Thederahn, M. D. Kuwabara, T. A. Larsen, and D. S. Sigman, J. Am. Chem. Soc., 111, 4941 (1989). 40. H. R. Drew, J. Mol. Biol., 176, 535 (1984). 41. L. E. Pope and D. S. Sigman, Proc. Natl. Acad. Sci. USA, 81, 3 (1984). 42. S. Goldstein and G. Czapski, Inorg. Chem., 24, 1087 (1985). 43. S. Goldstein and G. Czapski, J. Am. Chem. Soc., 108, 2244 (1986). 44. R. B. James and R. J. P. Williams, J. Chem. Soc., 2007 (1961). 45. G. J. Murakawa, C.-h. B. Chen, M. D. Kuwabara, D. Nierlich, and D. S. Sigman, Nucl. Acids Res., 17, 5361 (1989). 46. A. Mazumder, C.-h. B. Chen, R. B. Gaynor, and D. S. Sigman, Biochem. Biophys. Res. Commun., 187, 1503 (1992). 47. L. Pearson, C.-h. B. Chen, R. P. Gaynor, and D. S. Sigman, Nucl. Acids Res., 22, 2255 (1994). 48. Q. Guo, M. Lu, N. C. Seeman, and N. R. Kallenbach, Biochemistry, 29, 570 (1990). 49. A. Spassky and D. S. Sigman, Biochemistry, 24, 8050 (1985). 50. K. Kirkegaard, A. Spassky, H. Buc, and J. Wang, Proc. Natl Acad. Sci USA, 80, 2544 (1983). 51. A. Spassky, Biochemistry, 31, 10502 (1992). 52. B. Frantz and T. V. O'Halloran, Biochemistry, 29, 4747 (1990). 53. S. Buratowski, M. Sopta, J. Greenblatt, and P. A. Sharp, Proc. Natl. Acad. Sci. USA, 88, 7509 (1991). 54. M. D. Kuwabara and D. S. Sigman, Biochemistry, 26, 7234 (1987). 55. T. Thederahn, A. Spassky, M. D. Kuwabara, and D. S. Sigman, Biochem. Biophys. Res. Commun., 168, 756 (1990).
< previous page
page_512
next page >
< previous page
page_513
next page > Page 513
56. M. J. Absalon, C. R. Krishnamoorthy, G. McGall, J. W. Kozarich, and J. Stubbe, Nucl. Acids Res., 20, 4179 (1992). 57. J. W. Kozarich, J. L. Worth, B. L. Frank, D. F. Christner, D. E. Vanderwall, and J. Stubbe, Science, 245, 1396 (1989). 58. A. Mohindru, J. M. Fisher, and M. Rabinovitz, Biochem. Pharmacol., 32, 3627 (1983). 59. A. Mohindru, J. M. Fisher, and M. Rabinovitz, Nature, 303, 64 (1983). 60. A. Mazumder, M. Gupta, D. M. Perrin, D. S. Sigman, M. Rabinovitz, and Y. Pommier, AIDS Res. Hum. Retrovirus, 11, 115 (1994).
< previous page
page_513
next page >
< previous page
page_515
next page > Page 515
17 Specific DNA Cleavage by Manganese(III) Complexes Dennis J. Gravert and John H. Griffin Department of Chemistry, Stanford University, Stanford, CA 94305, USA
516
1. Introduction
516
1.1. Metal Complex-DNA Interactions
517
1.1.1. Ligation
517
1.1.2. Intercalation
518
1.1.3. Groove Binding
518
1.2. Oxidase Properties of [SalenMn(III)]+
518
1.2.1. Catalyst for Epoxidation and Carbon-Hydrogen Bond Activation Reactions
519
1.2.2. Synthesis and Characterization of [SalenMn(III)]+ Derivatives
519
2. DNA Cleavage by [SalenMn(III)]+ and Derivatives
519
2.1. Single-Strand Nicking of Supercoiled Plasmid DNA
520
2.2. DNA Affinity Cleaving of End-Labeled DNA Substrates
2.2.1. Global Specificity and Cleavage Efficiency of [SalenMn(III)]+ Complexes: Effects of Ring and (Asymmetric) Bridge Substituents
< previous page
page_515
522
next page >
< previous page
page_516
next page > Page 516
2.2.2. Cleavage Specificity of [SalenMn(III)]+ and Derivatives at Nucleotide Resolution
2.2.3. Basis for Sequence-Specific Cleavage by [SalenMn(III)]+
2.3. Mechanism of DNA Cleavage by [SalenMn(III)]+
2.3.1. Reactive Orientation
2.3.2. Reaction Conditions and Product Analysis
2.3.3. Possible Mechanistic Pathways for Oxidative DNA Cleavage
3. Comparative Cleavage Specificity of Minor Groove Agents
3.1. Representative Minor Groove Agents
3.2. DNA Affinity Cleaving Analysis
4. Conclusions
Abbreviations
References
524
527
527
527
528
529
532
532
532
534
534
535
1 Introduction The ability of proteins to bind specific sequences of DNA is critical to the expression of genes and the regulation of cellular functions. Developing an understanding of the determinants for these sequence-specific interactions leads to new insights of life at the cellular level and accelerates the development of novel pharmaceutical agents. Included among the families of sequence-specific DNA-binding proteins are many examples which utilize metal ions as an integral part of their DNA binding or binding/cleaving functions. In addition to proteins, numerous naturally occurring and synthetic small molecules also exhibit metal ion-dependent specific DNA binding and binding/cleaving properties, which may in turn give rise to antibacterial, antiviral, or antitumor activity. 1.1 Metal Complex-DNA Interactions Examining the structure of DNA is the first step in considering how a metal complex might engage in sequence-specific interactions with this
< previous page
page_516
next page >
< previous page
page_517
next page > Page 517
biopolymer. In the most common conformation of DNA, the B form, the sugar phosphate backbones of two antiparallel strands wind in a right-handed helix, held together by complementary hydrogen-bonded base pairs that are stacked along the helix axis. This arrangement generates major and minor grooves in which the walls are formed by the polyanionic backbones and the floors are formed by groups which protrude from the edges of aromatic base pairs. Sequence-dependent variations in DNA shape [1], as well as the chemical functionalities of the phosphate, deoxyribose, and purine/pyrimidine subunits, may be recognized and distinguished by metal complexes to achieve sequence specificity in binding. Metal complexes interact with DNA principally through three binding modes (ligation, intercalation, and groove binding) at DNA sequences that accommodate their steric requirements and/or provide for favorable covalent, van der Waals, or electrostatic interactions, including hydrogen bonding. 1.1.1 Ligation Ligation of a metal complex to donor groups present on DNA is the principal binding mode of the antitumor drug cisplatin, which covalently binds to N7 atoms of guanine in the major groove of DNA [2]. Intrastrand crosslinking of neighboring guanines by cisplatin can disrupt cellular processes and lead to cell death. The N7 atoms of guanine are also covalently coordinated by bis(phenanthroline)ruthenium dichloride, and enantioselectivity in this process has been demonstrated [3]. Site-and conformation-specific oxidation of guanine bases by macrocyclic nickel complexes ostensibly occurs through a ligand-metal complex [4], while palladium ligation appears to promote specific depurination at adenine [5]. 1.1.2 Intercalation Intercalation is the process by which planar, typically aromatic moieties are inserted between adjacent base pairs of the double helix. Rigid, substitutionally inert, octahedral tris(aromatic chelate) complexes exhibit various forms of selectivity in intercalative binding and have received intense scrutiny [6]. As DNA possesses handedness, enantiomeric forms of these chiral metal complexes may demonstrate enantiospecificity in binding. For example, the ∆ isomer of Rh(phen)2(phi)3+
< previous page
page_517
next page >
< previous page
page_518
next page > Page 518
has a right-handed propeller twist which, when intercalated via the major groove, complements the right-handed helix [7]. Binding of the left-handed enantiomer to the same site on DNA is hindered by steric repulsions between the sugar phosphate backbone and nonintercalated ligands. The intrinsic site selectivity of these complexes may be enhanced by addition of substituents to the nonintercalated ligands which participate in specific hydrogen bonding and van der Waals interactions with groups present in the DNA major groove [8]. 1.1.3 Groove Binding Generally, proteins bind in the major groove of DNA, while smaller molecules often prefer to bind in the narrower minor groove. Included among the minor groove binding and binding/cleaving agents are the metallobleomycins [9,10] and metalloporphyrins [1113]. On the other hand, it was recently definitively shown that dimeric complexes of chromomycin with Mg2+ bind specifically in the major groove of DNA [14]. 1.2 Oxidase Properties of [SalenMn(III)]+ 1.2.1 Catalyst for Epoxidation and Carbon-Hydrogen Bond Activation Reactions Kochi et al. reported in 1986 that [SalenMn(III)]+ (Salen = N,N'-ethylenebis(salicylideneaminato) dianion, Fig. 1) and substituted derivatives of this complex catalyze olefin epoxidation and carbon-hydrogen bond activation in the presence of the terminal oxidant iodosyl benzene [15]. Substituents were found to modulate the catalytic properties of [SalenMn(III)]+, and it was proposed that the activated intermediate responsible for oxidative chemistry is the Mn(V)-oxo species [SalenMn(V)O]+. More recently, Jacobsen et al. developed chiral derivatives of
Fig. 1. [SalenMn(III)]+.
< previous page
page_518
next page >
< previous page
page_519
next page > Page 519
[SalenMn(III)]+ that are useful catalysts for asymmetric olefin epoxidation [16]. This group also demonstrated ''electronic tuning" by substituents in the 5,5' positions in catalytic asymmetric epoxidation reactions [17]. 1.2.2 Synthesis and Characterization of [SalenMn(III)]+ Derivatives [SalenMn(III)]+ is a particularly attractive candidate for development as a catalyst for oxidation reactions because of its ease of synthesis and modification. The parent ligand SalenH2 and substituted derivatives are prepared by condensation of salicylaldehyde derivatives with substituted diamines. Stirring the Schiff base ligands with Mn(OAc)2 under aerobic conditions affords the requisite Mn(III) complexes. [SalenMn(III)]+ complexes are generally crystalline and amenable to characterization by infrared and UV-visible spectroscopy, mass spectrometry, and magnetic susceptibility measurements. 2 DNA Cleavage by [SalenMn(III)]+ and Derivatives We investigated the DNA binding and cleaving properties of [SalenMn(III)]+ and substituted derivatives. This work was prompted by recognition of the structural features and reactive chemistry of [SalenMn(III)]+ that are similar to known DNA binding/cleaving molecules: a planar, crescent shape, positive charge, and ability to activate carbonhydrogen bonds in the presence of terminal oxidants. We reasoned that oxidative modification of deoxyribose residues or nucleobases by activated [SalenMn(III)]+ would lead to DNA strand scission and that the ease with which substituted derivatives could be prepared would provide an extremely rich system in which to study steric and electronic factors involved in DNA recognition and cleavage. 2.1 Single-Strand Nicking of Supercoiled Plasmid DNA The ability of [SalenMn(III)]+ to promote DNA cleavage was demonstrated by single-strand nicking of supercoiled plasmid DNA in the presence of a terminal oxidant (Fig. 2) [18]. At higher concentrations,
< previous page
page_519
next page >
< previous page
page_520
next page > Page 520
Fig. 2. Cleavage of supercoiled plasmid pBR322. Plot of the average number of single-strand breaks per plasmid molecule (S) measured after incubated with [SalenMn(III)]+ in the presence or absence of oxidant (1.0 mM magnesium monoperoxyphthalate). [SalenMn(III)]+ produced linear DNA products which formed through independent yet proximal nicks on opposite DNA strands. While [SalenMn(III)]+ exhibits catalytic behavior in olefin epoxidation, production of single-strand breaks in DNA was not found to be catalytic under the conditions examined. DNA cleavage required the presence of terminal oxidants such as magnesium monoperoxyphthalate, potassium peroxymonosulfate, hydrogen peroxide, potassium superoxide, or iodosyl benzene. No cleavage was observed in the presence of sodium hypochlorite (bleach) or tert-butyl hydroperoxide. 2.2 DNA Affinity Cleaving of End-Labeled DNA Substrates The specificity and mechanistic aspects of DNA cleavage by [SalenMn(III)]+ and a series of ring- and bridgesubstituted derivatives (Fig. 3)
< previous page
page_520
next page >
< previous page
page_521
next page > Page 521
Fig. 3. Substituted derivatives of [SalenMn(III)]+.
< previous page
page_521
next page >
< previous page
page_522
next page > Page 522
were probed via a two-stage DNA affinity cleaving analysis (Scheme 1) [1921]. Substrates for these studies were derived from plasmid pBR322 (4363 bp). DNA double-strand affinity cleaving was used to screen for DNA binding/cleaving activity, to determine relative cleavage efficiencies, and to define cleavage specificity at a resolution of 2550 bp. After identifying loci of intense double-strand cleavage, restriction fragments containing these sequences served as substrates for high-resolution analysis of cleavage patterns and products. 2.2.1 Global Specificity and Cleavage Efficiency of [SalenMn(III)]+ Complexes: Effects of Ring and (Asymmetric) Bridge Substituents In the double-strand affinity cleaving assay, [SalenMn(III)]+ was found to produce a specific, nonuniform pattern of cleavage. The observed cleavage loci map to regions of pBR322 containing sites of multiple, contiguous A:T base pairs. The loci of greatest cleavage intensity correspond to the most A:T-rich regions of the plasmid (Table 1). The complex bearing methoxy substituents at the 3 and 3' positions of the aromatic ring [3,3'-(OCH3)2] as well as [NaphenMn(III)]+ and the 4,4'-(NEt2)2, 4,4'-(OCH3)2-7,7'-(CH3)2, 5,5'-(OCH3)2, 7,7'-(CH3)2, and 7,7'-Ph2 complexes were found to be incapable of effecting DNA double-strand cleavage in the presence of terminal oxidants. Among the active derivatives, the identity and position of attachment of substituents had no observTABLE 1 Loci of Greatest Double-Strand Cleavage Produced on Plasmid pBR322 by [SalenMn (III)]+ and Derivatives Positiona
Sequence
3231
TTTTAAATTAAAAATGAAGTTTTAAATCAATCTAAAGTATATAT
4164
TTTTTCAATATTATTGAA
4228
TATTTAGAAAAATAAACAAATA
4315
ATTAACCTATAAAAATA
aLowest numbered base pair of the underlined sequences, pBR322 numbering.
< previous page
page_522
next page >
< previous page
page_523
next page > Page 523
Scheme 1. Two-stage DNA affinity cleaving analysis.
< previous page
page_523
next page >
< previous page
page_524
next page > Page 524
able effect on DNA double-strand cleavage specificity; these derivatives produced patterns which were indistinguishable from that of the parent complex at this level of resolution. DNA double-strand cleavage efficiency was found to vary significantly with both the identity and the position of attachment of substituents. Most noteworthy, the 5,5'-Br2, 5,5'-Cl2, 4,4'-Cl2, and 5,5'-(NO2)2 derivatives cleaved DNA with efficiencies equal to or greater than that of [SalenMn(III)]+. Overall, DNA double-strand binding/cleaving efficiency for ring- and imine-substituted derivatives (Fig. 3) was found to decrease in the following order: 5,5'-Br2 = 5,5'-Cl2 > 4,4'-Cl2 > 5,5'-(NO2)2 = [SalenMn(III)]+ > 5,5'-F2 = 3,3'-F2 = 6,6'-Cl2 > 3,3'-Cl2 = 3,3',5,5'-Cl4 = 4,4'(OCH3)2 >> [NaphenMn(III)]+ = 3,3'-(OCH3)2 = 4,4'-(NEt2)2 = 4,4'-(OCH3)2-7,7'-(CH3)2 = 5,5'-(OCH3)2 = 7,7'(CH3)2 = 7,7'-Ph2. These results for DNA cleavage efficiency are in general accord with the results of Kochi et al., who found that 5,5'-Cl2 and 5,5'-(NO2)2 exhibited increased efficiency as catalysts for olefin epoxidation and carbonhydrogen activation relative to [SalenMn(III)]+ and 5,5'-(OCH3)2 [15]. Substituents at the 5,5' positions strongly affect the redox properties of [SalenMn(III)]+ derivatives [15], and it is likely that the great differences in DNA binding/cleaving efficiencies between 5,5'-Br2, 5,5'-Cl2, and 5,5'-(NO2)2, on the one hand, and 5,5'-(OCH3)2, on the other, derive primarily from electronic effects associated with the presence of electron withdrawing bromo, chloro, or nitro groups vs. electron-donating methoxy groups at these positions. Complexes bearing bridge modifications exhibited reduced DNA cleaving activity relative to the parent complex. DNA cleaving efficiency was found to vary not only with the structure of the bridge but with its stereochemistry. Thus, enantioselective recognition of the right-handed double helix was demonstrated by the finding that the cleavage efficiency for (R,R)-1 was five times greater than that for (S,S)-1. DNA cleavage activity for the bridge derivatives (Fig. 3) decreased in the order [SalenMn(III)]+ > (R,R)-1 > [SalophenMn(III)]+ > meso-1 > (S,S)-1 > [SalophenMn(III)]+. 2.2.2 Cleavage Specificity of [SalenMn(III)]+ and Derivatives at Nucleotide Resolution The specificity of DNA cleavage by [SalenMn(III)]+ and derivatives was probed at nucleotide resolution using 517bp EcoRI/Rsa I restric-
< previous page
page_524
next page >
< previous page
page_525
next page > Page 525
tion fragments from pBR322 which contained the series of intense cleavage loci located within positions 41644315 listed in Table 1. Intense cleavage was observed within and adjacent to sites of multiple contiguous A:T base pairs. Cleavage patterns were shifted to the 3' side on one DNA strand relative to the other, indicating that cleavage occurs from the minor groove of right-handed double-helical DNA [20,21]. Finally, single sites of cleavage, as well as symmetrical and unsymmetrical arrays of cleavage extending over several adjacent nucleotides, were observed at different locations. Such nonuniform patterns in cleavage are consistent with a cleavage process involving a bound rather than freely diffusible reactive species (such as hydroxyl radical) [2022]. For the series of aryl derivatives, substituents at the 4,4', 5,5', or 6,6' positions altered the precise patterns of cleavage produced by [SalenMn(III)]+ derivatives in a position-specific but substituent-independent fashion. The 3,3'-F2 and 3,3'-Cl2 complexes produced patterns of cleavage which matched that produced by [SalenMn(III)]+. Illustrative histograms are presented in Fig. 4. Derivatives bearing 5,5'-halogen or nitro substituents produced the same pattern of cleavage, while the regioisomeric complexes 3,3'-Cl2, 4,4'-Cl2, 5,5'-Cl2, and 6,6'-Cl2 produced cleavage patterns that differed from one another. That the 5,5'-disubstituted complexes produce indistinguishable patterns of cleavage suggests that fluoro, chloro, bromo, and nitro groups at these positions exert similar steric and/or electronic effects. Steric differences between reactive intermediates derived from 5,5'-F2 and [SalenMn(III)]+ should be small considering the similarity of the van der Waals radii of fluorine (1.47 Å) and hydrogen atoms (1.20 Å), which suggests that it is the electron-withdrawing character of fluorine and the other 5,5' substituents studied which defines their common DNA cleavage specificity. Derivatives substituted on the bridge also produced 3'-shifted DNA cleavage patterns at sites containing multiple contiguous A:T base pairs. The precise patterns of cleavage produced by these complexes were unique. Furthermore, cleavage patterns were different for (R,R)-1 and (S,S)-1 demonstrating enantiospecific recognition and cleavage of right-handed double-helical DNA. This result also provides strong evidence that DNA cleavage is initiated by specifically bound complexes rather than by a diffusible oxidant produced by the metal complex at a site remote from the DNA target.
< previous page
page_525
next page >
< previous page
page_526
next page > Page 526
Fig. 4. Representative histograms of single-strand DNA cleavage produced by [SalenMn(III)]+ derivatives in the presence of 1.0 mM magnesium monoperoxyphthlate on the 517-bp EcoRI/Rsa I fragment from plasmid pBR322. Lengths of arrows correspond to the relative amounts of cleavage as determined by optical densitometry.
< previous page
page_526
next page >
< previous page
page_527
next page > Page 527
2.2.3 Basis for Sequence-Specific Cleavage by [SalenMn(III)]+ The level of A:T binding/cleaving specificity exhibited by [SalenMn(III)]+ is remarkable considering the small size of this complex. As [SalenMn(III)]+ lacks hydrogen bond acceptors, such as the amide groups found in the A:T specific minor groove ligands netropsin and distamycin [20,23], we believe that the observed specificity is based on minor groove shape recognition. Hydrophobic cationic molecules such as [SalenMn(III)]+ can be sandwiched within the narrow minor groove found in A:T-rich regions of DNA where they engage in favorable van der Waals contacts and liberate ordered solvent molecules. The wider minor groove and protruding guanine N2 amino groups found in G:C-containing sequences do not provide as favorable a receptor site. 2.3 Mechanism of DNA Cleavage by [SalenMn(III)]+ 2.3.1 Reactive Orientation Similarities and differences among the cleavage patterns produced by [SalenMn(III)]+ and substituted derivatives provide a basis for assigning the reactive orientation relative to DNA of the oxidatively activated intermediate derived from [SalenMn(III)]+. The patterns of cleavage produced by derivatives bearing fluoro or chloro substituents at the 3,3' positions match those produced by the parent complex and differ from those produced by any other derivative. This result suggests that the concave edge of the activated crescent-shaped complex, from which extend the 3,3' substituents, is oriented away from the DNA. This is depicted using [SalenMn(V)O]+ as the proposed reactive intermediate in Fig. 5. If this represents the preferred orientation for initiation of DNA cleavage by [SalenMn(III)]+ derivatives, then it would be expected that substituents which extend from the convex edge of the complex could diminish cleavage efficiency and alter cleavage specificity. Consistent with these expectations are the findings that the complexes 7,7'-Ph2 (which is known to catalyze olefin epoxidation [15]) and 7,7'-(CH3)2are devoid of DNA binding/cleaving activity, while bridge-substituted complexes 1 (see Fig. 3) as well as [SalenMn(III)]+ exhibit diminished DNA binding/cleaving efficiency and altered patterns of cleavage relative to [SalenMn(III)]+.
< previous page
page_527
next page >
< previous page
page_528
next page > Page 528
Fig. 5. Proposed model of the reactive orientation for DNA cleavage by [SalenMn(V)O]+. Circles with two dots represent lone pairs of electrons on adenine N3 or thymine O2 atoms at the edges of the base pairs on the floor of the minor groove of B-form DNA. 2.3.2 Reaction Conditions and Product Analysis The effects of variations in some reaction conditions on DNA cleavage by [SalenMn(III)]+ have been analyzed [24,25]. DNA cleavage efficiency was greatest at pH 6.57.0 and fell off rapidly above pH 8. Cleavage efficiency was also observed to decrease with increasing ionic strength, which would be expected to weaken the coulombic driving force for association of the [SalenMn(III)]+ cation with the DNA polyanion. Heating DNA cleavage products generated by [SalenMn(III)]+ in 0.1 N NaOH altered and enhanced the intensity of DNA cleavage patterns, demonstrating that the complex mediates the formation of both direct DNA strand breaks and base-labile lesions. Finally, DNA cleavage efficiency, specificity, and the ratio of direct to base labile lesions were unaffected when reactions were carried out in the presence of terminal oxidant but in the absence of dioxygen. This limits consideration of mechanisms for [SalenMn(III)]+ -mediated DNA degradation to O2-independent processes.
< previous page
page_528
next page >
< previous page
page_529
next page > Page 529
Cleavage of calf thymus DNA by a combination of [SalenMn(III)]+ and hydrogen peroxide resulted in release of the free nucleotide bases cytosine, thymine, adenine, and guanine, which is consistent with cleavage mechanisms involving oxidative degradation of DNA deoxyribose residues. The natures of DNA end groups produced upon cleavage by [SalenMn(III)]+ have been probed by determining the electrophoretic mobilities of end-labeled cleavage fragments in sequencing gels relative to the mobilities of fragments with known termini and the effects of phosphatase treatment on cleavage fragment mobilities. Using 5'endlabeled substrates, it was determined that [SalenMn(III)]+ mediates the direct formation of DNA cleavage products bearing 3'-phosphate groups as well as latent lesions that afford strand scission products terminating in 3'phosphate groups upon treatment with base. Using 3'-end-labeled substrates, it was found that [SalenMn(III)]+ mediates the direct formation of DNA cleavage products bearing 5'-phosphate and 5'-nonphosphate, nonhydroxyl groups. Base treatment converts the 5'-nonphosphate termini to phosphate termini and develops latent lesions into strand scission products bearing 5'-phosphate end groups. 2.3.3 Possible Mechanistic Pathways for Oxidative DNA Cleavage The possible mechanistic pathways for [SalenMn(III)]+ -mediated DNA cleavage are limited by the nature of the observed end groups in the products and the findings that (1) the complex effects cleavage from the DNA minor groove, (2) cleavage is oxidative in nature but is neither dependent on nor affected by the presence of dioxygen, and (3) both direct and base-labile lesions are produced. Figure 6 depicts plausible cleavage mechanisms involving initial hydroxylation of deoxyribose residues at positions accessible to a reactive species located within the minor groove: CH1', C-H2'α, C-H4', and C-H5'. Following the work of Kochi et al. [15], C-H activation is proposed to involve initial hydrogen atom abstraction by [SalenMn(V)O]+ to form an alkyl radical and [SalenMn(IV)OH]+ followed by radical oxidation to a carbenium ion (perhaps by electron transfer to the proximal [SalenMn(IV)OH]+), and trapping of the carbenium ion by water or the coordinated hydroxyl of
< previous page
page_529
next page >
< previous page
page_530
next page > Page 530
Fig. 6. Plausible mechanisms for oxidative DNA cleavage mediated by [SalenMn(III)]+. (top) Formation of [SalenMn(V)O]+ from [SalenMn(III)]+ and terminal oxidant. (upper middle) Hydroxylation of DNA deoxyribose residues at positions C-H1' (left), C-H2'αa (center left), C-H4' (center right), and C-H5' (right). (lower middle and bottom) Intermediate and final products obtained from decomposition of hydroxylated deoxyribose residues. B, nucleobases; P, DNA phosphate.
< previous page
page_530
next page >
< previous page
page_531
next page > Page 531
[SalenMn(IV)OH] produced on electron transfer from the alkyl radical to [SalenMn(IV)OH]+. C-H1' activation would produce a hemiortho ester which decomposes to a 3'-dehydrolactone derivative with liberation of free base and a DNA fragment with a 5'-phosphate group. The dehydrolactone would undergo further decomposition to liberate 5-methylene-2-furanone and a DNA fragment with a 3'-phosphate group [26]. C-H2'α activation would afford a ribonucleotide residue within a DNA molecule. Such a species should hydrolyze under basic conditions to liberate a DNA fragment having a 5'-hydroxyl group and a 3'- or 2'-phosphate group. C-H4', activation would afford an unstable hemiacetal derivative which would decompose to liberate free base and produce a bis(hemiacetal). This species represents the base-labile lesion produced by bleomycin:Fe under lowoxygen conditions [27]. In the presence of base this species decomposes to a DNA fragment having a 5'-phosphate group and a DNA fragment having a 3'-nonphosphate group. C-H5', activation would afford a hemiacetal phosphate ester which should decompose under ambient conditions to liberate a DNA fragment with a 3'-phosphate group and a DNA fragment with a nucleoside 5'-aldehyde group. DNA fragments terminating in nucleoside 5'-aldehyde groups can be converted to corresponding fragments that terminate in 5'-phosphate groups by base workup, which would also liberate the DNA base and furfural [28]. The experimental results are inconsistent with a mechanism which proceeds via hydroxylation at C-H4', as this pathway leads to the formation of 3'-nonphosphate ends both prior and subsequent to treatment with base. Hydroxylation at C-H2'α does not appear to be a major pathway, as direct strand scission is facile and the expected 5'hydroxyl products were not observed in appreciable quantities. Overall, the results could be explicable by hydroxylation at C-H5', but only if decomposition of the nucleoside 5'-aldehyde proceeds appreciably without base workup. However, further characterization of the reaction products (especially those derived from deoxyribose oxidation) and analysis of kinetic isotope effects on cleavage of specifically labeled substrates [29] will be necessary to fully delineate the mechanism(s) of [SalenMn(III)]+-mediated DNA cleavage.
< previous page
page_531
next page >
< previous page
page_532
next page > Page 532
3 Comparative Cleavage Specificity of Minor Groove Agents 3.1 Representative Minor Groove Agents We have compared the DNA binding/cleaving specificity of [SalenMn(III)]+ with other small molecules which effect DNA cleavage from the minor groove [25,30]: bis(netropsin)succinamide-EDTA:Fe (BNSE:Fe) [18], neocarzinostatin (NCZS) [31], methidiumpropyl-EDTA:Fe (MPE:Fe) [32], bleomycin:Fe (Blm:Fe) [9,10], and [mesotetrakis(4-N-methylpyridiniumyl)porphyrinato]manganese(III) (MnTMPP) [1113]. Each of these agents requires a cofactor for activation. The iron-containing agents utilize dioxygen and a reducing agent, NCZS requires a thiol, and, like [SalenMn(III)]+, MnTMPP requires a terminal oxidant to cleave DNA [1113]. 3.2 DNA Affinity Cleaving Analysis The ''global" specificities of DNA cleavage produced by [SalenMn(III)]+ and the other agents were first examined at low resolution. Overlaid densitometric scans of cleavage patterns produced by BNSE:Fe, [SalenMn(III)]+ (MnS), NCZS, MPE:Fe, and Blm:Fe on linearized, end-labeled pBR322 are presented in Fig. 7. Not shown is the densitometric trace for MnTMPP, which is virtually indistinguishable from that for [SalenMn(III)]+. Other striking positive and negative correlations were observed among the DNA double-strand cleavage patterns produced by these compounds. The most intense cleavage loci observed with BNSE:Fe, [SalenMn(III)]+ MnTMPP, and NCZS correlate with regions of diminished cleavage intensity observed with MPE:Fe and Blm:Fe. Thus the tallest densitometric peaks observed with one subset of compounds align with deepest valleys observed with the other subset. The regions of alignment map to the most A:T-rich segments of pBR322. DNA cleavage by these small molecules was then examined at nucleotide resolution using end-labeled restriction fragments spanning the regions of correlated intense/weak cleavage. Here an inexact but notable correlation was observed between the intense cleavage produced by BNSE:Fe, [SalenMn(III)]+ MnTMPP, and NCZS within
< previous page
page_532
next page >
< previous page
page_533
next page > Page 533
Fig. 7. Optical densitometry of DNA double-strand cleavage produced by small molecules (Blm:Fe, MPE:Fe, NCZS, MnS, and BNSE:Fe, from top) on Sty I-linearized pBR322 plasmid DNA labeled at one 3' end with 32P-TTP. At bottom is the densitometry trace of the molecular weight marker lane. Lengths of the double-strand DNA fragments are indicated, as are the corresponding positions on pBR322 (in parentheses). (Reproduced with permission from Ref. 30.)
< previous page
page_533
next page >
< previous page
page_534
next page > Page 534
A:T-rich tracts, where cleavage by MPE:Fe and Blm:Fe was noticeably diminished. These correlations are consistent with the known DNA cleavage specificities of these reagents. The cleavage patterns produced by [SalenMn(III)]+ and MnTMPP were the most strongly correlated, and it is clear that the factors governing DNA minor groove recognition and cleavage are similar for these two manganese complexes. 4 Conclusions We found that the combination of [SalenMn(III)]+ and a terminal oxidant leads to efficient and specific oxidative cleavage of right-handed double-helical DNA from the minor groove at A:T-rich sites. The ease with which derivatives of the parent complex may be synthesized allowed us to probe the effects of changes in electronic and steric properties on DNA binding/cleaving activity in a systematic fashion. We find that (1) DNA cleavage efficiency varies with the identity and position of attachment of substituents, (2) the precise patterns of cleavage at A:T-rich sites vary with the position of attachment but not the identity of substituents, and in some cases indicate an electronic basis for the observed changes, (3) 3,3'-disubstituted derivatives display patterns of cleavage that are indistinguishable from that produced by the parent complex, indicating that the reactive orientation of the activated form of [SalenMn(III)]+ is one in which the concave edge of the crescent-shaped complex is oriented away from the DNA double helix, and (4) enantiomeric bridge-substituted derivatives exhibit different binding/cleaving efficiencies and produce different patterns of cleavage, demonstrating enantiospecific recognition and cleavage of DNA. Finally, in comparing the cleavage specificity of [SalenMn(III)]+ with that of other agents, we found that small molecules generate surprisingly distinct patterns of DNA double-strand cleavage and that DNA double-strand affinity cleaving provides a powerful approach to defining and relating the ''global" DNA cleavage specificities of small molecules. Abbreviations
Blm:Fe
bleomycin:Fe(II)
BNSE:Fe
bis(netropsin)succinamide-EDTA:Fe(II)
< previous page
page_534
next page >
< previous page
page_535
next page > Page 535
EDTA
ethylenediaminetetraacetic acid
Et
ethyl
MnS
[N,N' -ethylenebis(salicylideneaminato)]manganese(III)
MnTMPP
[meso-tetrakis(4-N-methylpyridiniumyl)porphyrinato]-manganese(III)
MPE:Fe
methidiumpropyl-EDTA:Fe(II)
NCZS
neocarzinostatin
OAc
acetate
Ph
phenyl
phen
1,10-phenanthroline
phi
9,10-phenanthrenequinone diimine
Salen
N,N' -ethylenebis(salicylideneaminato) dianion
TTP
thymidine triphosphate
References 1. R. E. Dickerson in Methods in Enzymology, Vol. 211 (D. M. J. Lilley and J. E. Dahlberg, eds.), Academic Press, San Diego, 1992, p. 67111. 2. J. K. Barton and S. J. Lippard, Metal Ions in Biology, 1, 31113 (1980). 3. J. K. Barton and E. Lolis, J. Am. Chem. Soc., 107, 708709 (1986). 4. J. G. Muller, X. Chen, A. C. Dadiz, S. E. Rokita, and C. J. Burrows, J. Am. Chem. Soc., 114, 64076411 (1992). 5. B. L. Iverson and P. B. Dervan in Methods in Enzymology, Vol. 218 (R. Wu, ed.), Academic Press, San Diego, 1993, p. 222227. 6. A. H. Krotz, L. Y. Kuo, T. P. Shields, and J. K. Barton, J. Am. Chem. Soc., 115, 38773882 (1993). 7. D. Campisi, T. Morii, and J. K. Barton, Biochemistry, 33, 41304139 (1994). 8. A. H. Krotz, B. P. Hudson, and J. K. Barton, J. Am. Chem. Soc., 115, 1257712578 (1993). 9. S. M. Hecht, Acc. Chem. Res., 19, 383391 (1986). 10. J. Stubbe and J. W. Kozarich, Chem. Rev., 87, 11071136 (1987).
11. R. J. Fiel, T. A. Beerman, E. H. Mark, and N. Datta-Gupta, Biochem. Biophys. Res. Commun., 107, 10671074 (1982). 12. G. Pratviel, M. Pitie, J. Bernadou, and B. Meunier, Nucl. Acids Res., 19, 62836288 (1991).
< previous page
page_535
next page >
< previous page
page_536
next page > Page 536
13. U. Sehlstedt, S. K. Kim, P. Carter, J. Goodisman, J. F. Vollano, B. Norten, and J. C. Dabrowiak, Biochemistry, 33, 417426 (1994). 14. D. J. Silva and D. E. Kahne, J. Am. Chem. Soc., 115, 79627970 (1993). 15. K. Srinivasan, P. Michaud, and J. K. Kochi, J. Am. Chem. Soc., 108, 23092320 (1986). 16. W. Zhang, J. L. Loebach, S. R. Wilson, and E. N. Jacobsen, J. Am. Chem. Soc., 112, 28012803 (1990). 17. E. N. Jacobsen, W. Zhang, and M. L. Güler, J. Am. Chem. Soc., 113, 67036704 (1991). 18. D. J. Gravert and J. H. Griffin, J. Org. Chem., 58, 820822 (1993). 19. D. J. Gravert and J. H. Griffin, submitted. 20. P. B. Dervan, Science, 232, 464470 (1986). 21. J. S. Taylor, P. G. Schultz, and P. B. Dervan, Tetrahedron, 40, 457465 (1984). 22. T. D. Tullius, B. A. Dombroski, M. E. A. Churchill, and L. Kam in Methods in Enzymology, Vol. 155 (D. M. J. Lilley and J. E. Dahlberg, eds.), Academic Press, San Diego, 1987, p. 537558. 23. M. L. Kopka, C. Yoon, D. Goodsell, P. Pjura, and R. E. Dickerson, Proc. Natl. Acad. Sci. USA, 82, 13761380 (1985). 24. J. H. Griffin, PhD thesis, California Institute of Technology, Pasadena. 25. D. J. Gravert and J. H. Griffin, unpublished results. 26. T. E. Goyne and D. S. Sigman, J. Am. Chem. Soc., 109, 28462848 (1987). 27. H. Sugiyama, C. Xu, N. Murugesan, and S. M. Hecht, J. Am. Chem. Soc., 107, 41044105 (1985). 28. M. Pitie, G. Pratviel, J. Bernadou, and B. Meunier, Proc. Natl. Acad. Sci. USA, 89, 39673971 (1992). 29. J. W. Kozarich, L. Worth, Jr., B. L. Frank, D. F. Christner, D. E. Vanderwall, and J. Stubbe, Science, 245, 13961399 (1989). 30. J. H. Griffin, Bioorg. Med. Chem. Lett., 5, 7376 (1995). 31. L. F. Povirk, C. W. Houlgrave, and Y.-H. Han., J. Biol. Chem., 263, 1926319266 (1988). 32. R. P. Hertzberg and P. B. Dervan, J. Am. Chem. Soc., 104, 313315 (1982).
< previous page
page_536
next page >
< previous page
page_537
next page > Page 537
18 Nickel Complexes As Probes of Guanine Sites in Nucleic Acid Folding Cynthia J. Burrows1 and Steven E. Rokita2 1Department of Chemistry, University of Utah, Salt Lake City, UT 84112, USA 2Department of Chemistry and Biochemistry, University of Maryland, College Park, MD 20742, USA
538
1. Introduction
1.1. Using Chemical Probes of Nucleic Acid Structure
1.2. Overview of Nickel Chemistry with Nucleic Acids
539
540
2. DNA Secondary Structure
2.1. Reaction of Single- vs. Double-Stranded DNA with Nickel Complexes
2.2. Studies of Mismatched, Bulged, and Hairpin Gs
2.3. Sequence-Specific Chemistry of Nickel Peptide-Protein Conjugates
540
542
543
546
3. RNA Tertiary Structure
3.1. Comparison of Nickel Reactivity with a Known Structure: tRNAPhe
3.2. Applications to Ribozymes: Group I Intron and Hairpin Ribozyme
3.3. Conformational Changes in an RNA Pseudoknot
< previous page
538
page_537
546
547
550
next page >
< previous page
page_538
next page > Page 538
3.4. Studies of an Antisense RNA-RNA Complex: micF RNA and ompF mRNA
551
553
4. Experimental Methods
553
4.1. Nucleobase Oxidation vs. Alkylation
4.2. Base Lability vs. Primer Extension Analysis
554
555
5. Mechanistic Considerations
555
5.1. Base vs. Sugar Modification
555
5.2. Ligand Effects
556
5.3. Role of Nickel(III)
557
6. Conclusions
557
Abbreviations
558
References
1 Introduction The biological activity of a nucleic acid is intimately married to its three-dimensional structure. This is particularly true for nontranslated RNAs in which a wealth of folding patterns gives rise to the various functions of RNAs including protein binding and ribozyme activity. Determination of tertiary structure through crystallographic and nuclear magnetic resonance (NMR) methods is limited by the difficulty of obtaining crystals and the complexity of NMR analysis for large biopolymers. As a consequence, most structural investigations begin with extensive characterization using chemical and enzymatic probes. 1.1 Using Chemical Probes of Nucleic Acid Structure Various chemical and enzyme reagents selectively modify nucleic acids by recognizing either a certain base or the solvent exposure of a particular site, or both [13]. The sites of reaction can be easily visualized by gel electrophoresis and autoradiography of radiolabeled nucleic acids if the modification leads to either strand scission (most commonly) or to sufficient base damage that polymerase-dependent chain extension is inhibited (see Sec. 4.2). In terms of guanine reactivity, dimethyl sulfate
< previous page
page_538
next page >
< previous page
page_539
next page > Page 539
(DMS) is a common reagent for alkylation of the N7 position, and most guanines in folded DNA and RNA structures are reactive due to the small size of the reagent and the exposure of N7 in the major groove of
DNA. T1 nuclease is an example of an enzymatic probe for guanine; it is most reactive with highly exposed guanines present in single-stranded regions of RNA and consequently reacts with a small number of sites. A concern in the use of either chemical or enzymatic probes is the question of whether interaction with the structural probe distorts the native structure of the nucleic acid leading to spurious results. It is therefore of interest to generate a family of structural probes that sample the folded surface of the nucleic acid in specific ways without altering the conformation of the biopolymer in the process. These probes will be most useful if they are intermediate in size between the small and nonselective reagents such as DMS and the large, highly selective enzymes such as T1 nuclease. 1.2 Overview of Nickel Chemistry with Nucleic Acids The properties of nickel(II) coordination compounds lend themselves well to the study of nucleic acid chemistry. The Ni2+ ion is known to bind to guanine's N7 [4], and it is in fact this property that is thought to be responsible for nickel's ability to induce B to Z transitions in duplex DNA [5]. In addition, compounds derived from nickel ores have been found to be carcinogenic, and this activity has been linked to oxidative chemistry mediated by nickel leading to DNA strand breaks, DNA-DNA crosslinks, and DNA-protein crosslinks [6,7]. The coordination number, geometry and redox chemistry of nickel(II) are highly dependent on the ligands surrounding the metal ion. These properties together suggest that nickel complexes would be excellent candidates for inducing site-
< previous page
page_539
next page >
< previous page
page_540
next page > Page 540
Fig. 1. Nickel complexes used as structural probes of nucleic acids. specific redox chemistry of nucleic acids. Indeed, a wide range of square planar, tetradentate complexes of nickel(II) promote guanine-specific oxidation that is dependent on the exposure of the nucleobase in folded nucleic acid structures. Compounds 13 of Fig. 1 have been studied most extensively in this regard [8]. 2 DNA Secondary Structure 2.1 Reaction of Single- vs. Double-Stranded DNA with Nickel Complexes Both NiCR, 1 [9], and Ni(cyclam), 2, are square planar, cationic complexes exhibiting high water solubility and stability toward hydrolysis at pH 7. Oligodeoxynucleotides (ODNs) display only weak binding to nickel(II) in the absence of oxidants; however, in the presence of water-soluble peracids such as potassium monopersulfate (KHSO5) or magnesium monoperoxyphthalate (MMPP), a rapid reaction occurs leading to oxidative modification of guanine bases [10]. No evidence is seen for direct strand scission that could have resulted from hydrogen atom abstraction from C1' or C4' of the ribose moiety, and no 3'-glycolates have been detected [11]. Instead modification of the nucleobase was indicated by the alkaline lability of the ODN. Only guanines were reactive. Interestingly, all Gs present in a single-stranded ODN were reactive while reaction of the same ODN annealed to its complementary
< previous page
page_540
next page >
< previous page
page_541
next page > Page 541
Fig. 2. Synthetic oligodeoxynucleotides used in reaction with NiCR/KHSO5. Arrows indicate sites of modification observed for oligos labeled (*) at the 5' end with 32P. strand revealed that Gs were unreactive unless residing at a helix terminus (Fig 2a vs. 2b). Apparently, Gs participating in a Watson-Crick base pair within a helical region are inaccessible to reaction with terminus, are substantially more reactive.
, but those displaying high solvent accessibility, such as in a single-stranded region or at a helix
The unreactivity of G-C pairs within a B helix is likely due to the inability of G N7 to coordinate to octahedral metal ions. During reaction with nickel complexes under oxidative conditions, a transient nickel(III) complex is proposed to be the active agent recognizing Gs and mediating base oxidation [12]. Nickel(II) tetraazamacrocycles become six coordinate upon oxidation to Ni(III) by coordination of anions or solvent. The lone pair on N7 of G is the most common transition metal binding site in nucleic acids and is oriented in the major groove of B-form DNA. While it is the site of binding of square planar metal complexes such as Pt(II) compounds, there are no known examples of an octahedral metal ion directly coordinating to N7 in B-DNA via a metal-N7 bond. This is because N7 is located deep in the major groove and axial ligation to a complex like NiCR would cause severe steric interactions with the nucleotide toward the 5' direction [13]. On the other hand, G N7 is quite accessible on the surface of duplex DNA when it is folded into a left-
< previous page
page_541
next page >
< previous page
page_542
next page > Page 542
and bound to Z-DNA have handed helix in the Z form, and crystal structures of both been determined [14,15]. Therefore, it is theoretically possible that nickel(II) complexes will mediate oxidative reactions at all Gs of G-C pairs in Z-DNA, while for B-DNA these Gs are inaccessible and unreactive unless located at the end of a helix. 2.2 Studies of Mismatched, Bulged, and Hairpin Gs The high selectivity of nickel(II) complexes for only accessible Gs in duplex DNA suggested their application as a structural probe for nucleic acid folding. A series of duplex ODNs were studied in which Gs were located at unusual sites including mismatches, bulges, and hairpin loops [16]. As predicted, the reaction with NiCR/KHSO5 recognized only those Gs whose N7 was highly exposed in comparison to other normal G-C pairs in the same molecule (see Fig. 2cf). While the observed recognition of a G-G mismatch was very exciting, in this particular case it may be an artifact; the ODN shown in Fig. 2c is self-complementary and can also fold into a hairpin loop, which may account for the reactivity of the central G. Indeed, subsequent studies of G-X mismatches have shown the G to be relatively unreactive [17]. These base pairs, while somewhat less stable than G-C in non-self-complementary sequences, may still be largely hydrogen-bonded and stacked within the helix. Studies of a double-stranded ODN containing a dynamic bulged nucleotide (Fig. 2d) showed that each of the two Gs participating in the bulge was reactive with NiCR in the presence of KHSO5 [16]. The extent of reaction was not equivalent for the two Gs but rather was distributed 60:40 between G7 and G8. The higher reactivity of G7 may indicate a higher population of a conformer with G7 in an extrahelical position. Studies of other migrating G bulges by either NMR [18] or NiCR reaction [19] are consistent with this preference for the bulged base to be located between two other purines. Furthermore, this unequal reactivity suggests that nickel complexes probe the preexisting equilibrium of nucleic acid conformations rather than generate unique nickel-dependent structures. Guanines located in a single-stranded hairpin loop are reactive
< previous page
page_542
next page >
< previous page
page_543
next page > Page 543
with NiCR/KHSO5 (Fig. 2e), and again their degree of reactivity is thought to be a measure of the solvent (and therefore, nickel) accessibility of N7 [16]. However, lack of reactivity may or may not be an indication of base pairing. For example, the hairpin ODN shown in Fig. 2f shows high reactivity at G7 and G8 (large arrows), but modest reactivity at G9, thought to be part of the hairpin, and similar reactivity at G10 that is predicted to be basepaired with C6. The reduced reactivity of G9 could be a result of stacking of the purine base within a hydrophobic pocket created by the loop. On the other hand, reduced reactivity would be predicted if an alternative pairing scheme of C6G9 were adopted leaving G10 as a bulged base. This latter scheme would account for the increased reactivity of G10, although one cannot exclude the possibility that the stem simply opens up at C6G10 making all four of G7G10 available for reaction to varying degrees. Thus, observation of reactivity of a particular nucleotide with a chemical probe is only an indication of solvent accessibility and does not itself define structure. Complementary techniques such as NMR provide supporting evidence for base-pairing and stacking interactions. Hairpin loops are also present in telomeric structures formed at the G-rich ends of chromosomes. Interestingly, nickel bleomycin was found to mediate oxidative modification of guanines in TTG loops bridging a G-quartet ODN structure [20] (see Fig. 3). The oxidation reaction could be effected by either reaction with KHSO5 or formation of nickel(III) using Ir(IV) as oxidant. Although all guanines were modified to a certain extent, G9 and G15 residing in predicted loops were substantially more reactive. 2.3 Sequence-Specific Chemistry of Nickel Peptide-Protein Conjugates While the intrinsic selectivity of nickel(II) complexes favors reaction at guanine residues under oxidative conditions, other sites may be targeted by conjugation to sequence-specific DNA-binding agents. Excellent examples are found in two independent studies in which a nickel-peptide complex was appended to a segment of a DNA-binding protein. These studies build on early work with metal-binding peptides such as glycylglycylhistidine (GGH) in which a square planar complex is formed
< previous page
page_543
next page >
< previous page
page_544
next page > Page 544
Fig. 3. Preferential base sites of G-quartet d(T2G4)4 attacked by nickel(III) bleomycin species. (Adapted with permission from Ref. 20.) through ligation to the terminal amino group, two deprotonated amide nitrogens, and histidine's imidazole group [21]. Mack and Dervan found that a single thymine residue near the hixL recognition site was the principle site of reaction with MMPP when NiGGH was appended to the amino terminus of the three-helix bundle (amino acids 139190) of Hin recombinase (see Fig. 4) [22,23]. As with NiCR/KHSO5, strand cleavage was only observed after an alkaline workup, but labeling studies pointed to involvement of the C4' hydrogen. Thus, the chemical mechanism of thymidine oxidation using a nickel-peptide in conjunction with a peracid may or may not be analogous to guanine chemistry with nickel(II) macrocycles. More recently, Nagaoka and coworkers showed that conjugation of NiGGH to the N terminus of a three-zinc-finger protein, amino acids 529696 of the transcription factor Spl, led to direct strand scission of a restriction fragment at two specific cytosine residues using MMPP as
< previous page
page_544
next page >
< previous page
page_545
next page > Page 545
Fig. 4. Proposed binding interaction between NiGGH-Hin recombinase (139190) and DNA. Arrows indicates sites and extent of reaction. (Adapted with permission from Ref. 22.)
< previous page
page_545
next page >
< previous page
page_546
next page > Page 546
oxidant [24]. Curiously, no alkaline workup was required for strand scission suggesting yet another mechanism of oxidation of cytosine residues. Although cleavage efficiency was relatively low, the experiment demonstrates nicely the potential design of a family of artificial restriction endonucleases based on various zinc fingers as recognition units and a nickel(II) complex as the catalytic cleavage agent. 3 RNA Tertiary Structure 3.1 Comparison of Nickel Reactivity with a Known Structure: tRNAPhe While DNA adopts a predominant B-helical form in solution, many types of RNA fold into complex threedimensional structures. Hence, much of the current application of chemical probes to the determination of nucleic acid structure is directed at RNA. In order to determine the utility of nickel probes such as NiCR, it was necessary to first examine reaction specificity with a molecule whose structure was well established. tRNAPhe is perhaps the best studied RNA structure by physical, chemical, and computational methods, and its crystal structure has been determined [25]. It is therefore a logical first study of any new structural probe. In the presence of NiCR and either KHSO5 or MMPP, the native form of tRNAPhe obtained in the presence of Mg2+ displayed four reactive guanine sites [26]. Three of these (G18, G19, and G20) are located in the D loop, and one (G34) is near the bottom of the anticodon loop (see Fig. 5). In the absence of Mg2+, the molecule is partially denatured, and eight more Gs become reactive upon unfolding. Analysis of the four principle sites of reaction in native tRNAPhe indicates that an excellent correlation exists between the exposure of G N7 as determined crystallographically and the solution reactivity of Gs with NiCR. Computations based on the crystal structure that include consideration of both the accessible surface area of N7 and the local electrostatics predict an order of reactivity of G19 > G20 > G18 > all others [27]. The experimentally determined nature of G N7 using NiCR as a probe reflected this trend with high accuracy as shown in Fig. 6. The ability of complexes such as NiCR to reliably probe the surface of folded tRNAPhe, its use in native as well as partially denaturing
< previous page
page_546
next page >
< previous page
page_547
next page > Page 547
Fig. 5. Secondary structure of tRNAPhe. Solid (and open) arrows indicate sites of modification by NiCR/KHSO5 in the presence (and absence) of Mg2+. (Adapted with permission from Ref. 8.) conditions, and the apparent success in recognition without structural perturbation render them useful probes in applications to RNA molecules of less well-defined structure. In addition, the correlation with the local environment of G N7 confirms the site of recognition and reaction with nickel(II) complexes under oxidative conditions. 3.2 Applications to Ribozymes: Group I Intron and Hairpin Ribozyme The catalytic activity of the Tetrahymena group I intron RNA requires the molecule to adopt a specific folded structure. A variety of methodschemical, enzymatic, spectroscopic, and computationalhave been applied to the determination of the L-21 Sca I derivative of this intron [28].
< previous page
page_547
next page >
< previous page
page_548
next page > Page 548
Fig. 6. Calculated (open circles) vs. observed (filled circles) reactivity of G N7 in tRNAPhe. Calculations are based on the accessible surface-integrated field index [27]. Reaction with NiCR/KHSO5 confirmed that a number of guanines are exposed in predicted hairpin loops (Fig. 7) [26]. Two sites, G257 and G331, apparently participate in base pairs to cytosines but are nevertheless reactive because they terminate helical regions. Interestingly, G264, which is buried in the P7 helix as a base pair with C311, was found to be highly reactive. On inspection of the predicted structure [29], this result is less surprising since N7 of G264 appears easily accessible in a large cavity that forms the ribozyme's active site. During the self-splicing reaction, G264 participates in hydrogen bonding, via N7, to a guanosine cofactor necessary for activity. NiCR probes this binding surface by occupying the site of the guanosine cofactor. The 50-nucleotide hairpin ribozyme derived from satellite RNA associated with tobacco ringspot virus was subjected to various base-specific chemical probes in studies by Butcher and Burke [30]. Diethylpyrocarbonate (DEPC), which modifies N7 of adenine, and kethoxal, which reacts with N1 and N3 of guanines, provided complementary information to that provided by NiCR. In the absence of Mg2+, the ribozyme is semidenatured. Major conformational changes occur upon addition of Mg2+ to generate the native, folded structure. Interestingly, the conformation was not significantly altered upon substrate binding indicating that the folded ribozyme was preorganized for recognition of the target oligonucleotide.
< previous page
page_548
next page >
< previous page
page_549
next page > Page 549
Fig. 7. Secondary structure of L-21 Sca I derivative of Tetrahymena group I intron. Arrows point to sites of modification with NiCR/KHSO5. (Adapted with permission from Ref. 8.)
< previous page
page_549
next page >
< previous page
page_550
next page > Page 550
3.3 Conformational Changes in an RNA Pseudoknot Ribosomal frameshifting is a mechanism by which organisms may regulate the expression of genes. Many retroviral RNAs contain sequences proposed to adopt a pseudoknot (VPK) structure in a region down-stream of a frameshift site. In a study of mouse mammary tumor virus mRNA by Chen et al., the structures of various mutant sequences were probed with chemical and enzymatic reagents in order to characterize the pseudoknot structures [31]. VPK is a 34-base oligonucleotide with the same frameshifting efficiency as the wild-type pseudoknot (Fig. 8).
Fig. 8. Reactivity of the VPK pseudoknot and the ∆A14 mutant with endonucleases V1, S1, and T1 and with NiCR/KHSO5. (Adapted with permission from Ref. 31.)
< previous page
page_550
next page >
< previous page
page_551
next page > Page 551
The reactivity of G7 with both RNase T1 and NiCR confirmed that this base is part of loop 1 of the pseudoknot. The mild reactivity of G28 is consistent with the G28-U13 base pair being located at the junction between stem 2 and loop 2, although it is not stably paired. The crucial role for A14 in formation of the pseudoknot structure was seen by examining NiCR reactivity of a mutant, ∆A14, lacking only that nucleotide (Fig. 8). Deletion of A14 resulted in a completely different folding pattern, now a hairpin structure rather than a pseudoknot, and all frameshifting activity was lost. This experiment demonstrates the strong connection between sequence, structure, and function and illustrates the utility of chemical probes such as NiCR in analyzing such relationships. 3.4 Studies of an Antisense RNA-RNA Complex: micF RNA and ompF mRNA NiCR has also been applied to the study of a naturally occurring antisense RNA-RNA complex. micF RNA regulates the stability of ompF mRNA in response to changes in temperature and other environmental factors. The secondary structure of E. coli ompF mRNA as well as its complex with micF RNA have been studied by Schmidt and coworkers [32]. A number of the structural conclusions are reminiscent of the early study of model ODNs using NiCR [16]. For example, it is interesting to note that differing reactivities with NiCR are observed for G152 compared to G156 in loop II (Fig. 9). This result suggests the possibility of a G156-C150 base pair that might be only weakly formed across the base of the hairpin loop leaving U157 bulged out. At another site, G131 and G132 are strongly and equally reactive. This observation implies that they participate in a migrating bulge with C178 in which the two bulged conformers are present to about equal extent. NiCR probing of the duplex region formed upon binding of micF RNA to ompF mRNA shows a third interesting pointnone of the bases G99, G100, nor G101 of ompF mRNA were reactive with NiCR suggesting that they are entirely base paired with nucleotides G32, G33, and C34 of the micF RNA strand. This must involve formation of a G-G mismatch in which N7 of G99 is stacked within the helix and protected from reaction rather than bulged out.
< previous page
page_551
next page >
< previous page
page_552
next page > Page 552
Fig. 9. A portion of the proposed secondary structure of the ompF-213 mRNA/micF RNA duplex. Arrows indicate sites of reaction with NiCR/KHSO5 (Ni), RNases T1 (T1) and T2 (T2) and basal cleavage (B). (Adapted with permission from Ref. 32.)
< previous page
page_552
next page >
< previous page
page_553
next page > Page 553
4 Experimental Methods 4.1 Nucleobase Oxidation vs. Alkylation Chemical probes of nucleic acid structure must mediate a detectable modification in the nucleic acid at a specific nucleotide site. The modification observed might be in the form of direct strand scission or a chemical change in the nucleobase structure that is detected by either alkaline lability or primer extension analysis (see Sec. 4.2). For the guanine-specific chemistry described above, the actual products of nickel-catalyzed oxidation with KHSO5 or MMPP have not been determined. However, it is well established that a variety of guanine oxidation products are sensitive to depurination, particularly at high pH or in the presence of certain amines such as piperidine or butylamine [33]. Similarly, alkylation of G at N7 leads to depurination and strand scission after alkaline workup, and this is the mechanism in place during the Maxam-Gilbert sequencing reaction for G which employs DMS as a methylating agent [34]. Although oxidation of G is the common result of nickel(II)-mediated reactions of nucleic acids, it was serendipitously found that certain nickel(II) complexes result in guanine alkylation (or arylation) [35]. Complex 3, a water-soluble derivative of nickel(salen), formed covalent adducts with ODNs under oxidative conditions. Extensive studies of substituent effects demonstrated the involvement of the phenol moiety of the salen ligand in the coupling reaction with DNA. Based on available evidence, it is proposed that KHSO5 oxidation of 3 leads initially to an unstable nickel(III) complex which decomposes by ligand-to-metal electron transfer generating a phenoxy radical. Phenoxy and alkyl radicals are known to react with nucleobases [36], and such a reaction would form a major covalent modification to the base that could be easily detected [Eq. (1)]. That guanine is the only reactive
< previous page
page_553
next page >
< previous page
page_554
next page > Page 554
base with 3 is known from primer extension studies with RNA [37]. However, at least two different coupling reactions occur with guanine and 3+ since it was found that some of the alkylated lesions were alkaline labile and some were not. In the cases studied, the sites of reactivity were consistent with the notion that exposure of the guanine nucleobase was again important for reaction. 4.2 Base Lability vs. Primer Extension Analysis Polyacrylamide gel electrophoresis followed by autoradiography of 32P-labeled nucleic acids is the common method of analysis of chemical or enzymatic modification. However, electrophoretic mobility is only affected by substantial changes in the size or charge of the biopolymer; thus minor chemical modifications to a single nucleotide may go undetected in a sizable oligomer. Modifications to purine bases are therefore analyzed in one of two ways. First, depurination can be detected by treatment under alkaline conditions or with a suitable amine at lower pH. For DNA, an alkaline treatment with 0.2 M piperidine at 90°C for 30 min is standard [16]. For RNA, such alkaline conditions are too harsh and would lead to indiscriminate base-catalyzed hydrolysis. Instead, chemically modified RNA samples are treated with a 1 M solution of aniline acetate (pH 4.5) at 60°C for 20 min [26]. In either case, the nucleic acid strand undergoes an elimination reaction generating two fragments, each with phosphate termini. Only the fragment containing the 32P label is detected by autoradiography, and thus care must be taken to perform chemical modification reactions under ''single-hit" conditions. For the nickel complexes described herein, concentrations of 110 µM are typical, although oxidants may be present in large excess (10100 µM). An alternative to visualization of chemical modification through strand scission is to use a primer extension method. In this technique, a radiolabeled primer strand complementary to a short segment of the strand of interest is synthesized and chain extension is continued until a site of chemical modification is encountered. The polymerase enzyme will then pause or stop at the site of an unrecognizable base, creating oligonucleotides whose length and sequence can be correlated to the original modified nucleic acid. This technique is quite sensitive to changes in the nucleotide structure, unless the chemical modification is minor or converts one base to an analog of another. For RNA, both
< previous page
page_554
next page >
< previous page
page_555
next page > Page 555
oxidative modification by NiCR/KHSO5 and oxidative alkylation by 3/KHSO5 yield polymerase stop points [37]. 5 Mechanistic Considerations 5.1 Base vs. Sugar Modification Information regarding the mechanism of nickel-mediated nucleic acid oxidation is gleaned from both the site of reaction as well as the products formed from oxidation. The high specificity of most of the nickel reactions for accessible guanine residues argues against the intermediacy of a freely diffusible radical such as HO. Reagents based on Fenton chemistry have been extremely useful as probes of nucleic acid structure in which the hydroxyl radical can survey the solvent-accessible surface of a folded nucleic acid with little base specificity [3840]. Hydroxyl radical reacts with sugar residues in nucleic acids via hydrogen atom abstraction leading to a cascade of events culminating in direct strand scission; however, HO is also reactive with nucleobases, although this chemistry is not generally observed unless an alkaline work-up is included. The fact that neither direct strand scission nor phosphoglycolate products have been observed in nickel/KHSO5 strongly supports the notion that the chemical modification is centered on the guanine heterocycle [11]. Whether this specificity derives from binding of the nickel complex to G N7, a higher intrinsic susceptibility of accessible Gs to oxidation, or both is not yet known. A more complete study of the products of G oxidation will aid in this determination. 5.2 Ligand Effects Not all nickel(II) complexes are applicable as probes of nucleic acid structure. In an extensive study of ligand effects, it was found that the ligand field strength, coordination number, reduction potential, and geometry of the complex were key factors in determining the reactivity of the nickel compound [12]. With the exception of nickel bleomycin, which is thought to be five coordinate [20], all of the active complexes were four coordinate and square planar with relatively strong in-plane ligand fields. These conditions are met using ligands that provide strong nitrogen (amine, imine, pyridine) and oxygen (phenolate) donors
< previous page
page_555
next page >
< previous page
page_556
next page > Page 556
in a tetradentate, preferably macrocyclic, arrangement with a ring size of 1314 in the case of macrocycles. Ligands that enforce a cis geometry (such as tren) or generate octahedral complexes with no vacant coordination sites [such as tris(phenanthroline)] render the nickel complex completely inactive. Similarly, a weak field species such as has no activity with KHSO5 and nucleic acids. Steric encumbrance appears to play a smaller role in the reactivity of a complex. Highly substituted ligands such as hexamethylcyclam result in somewhat less reactive complexes depending on the stereochemistry of the substituents. Whether this phenomenon is due to hindrance in formation of an appropriate six-coordinate nickel(III) intermediate or to an inability to bind properly to the nucleic acid is a subject for further study. 5.3 Role of Nickel(III) The strong dependence of nickel-mediated nucleic acid oxidation on ligand field strength and reduction potential suggests a crucial role for nickel(III). These criteria for ligand design are consistent with formation of an intermediate nickel(III) complex whose additional axial ligands (likely H2O or buffer, initially) can be exchanged for G N7 target sites and oxidant. In studies of nickel(cyclam) with a variety of oxidants, it was found that only those oxidants likely to coordinate to nickel(III) at pH 7 were effective, namely peracids, while H2O2 was not [41]. Chemical generation of nickel(III) in the absence of peracids did not lead to guanine modification when cyclam was the ligand [41], although it did when bleomycin was used [11]. One possible role for nickel(III) is to act as a template to bring together guanine and an
< previous page
page_556
next page >
< previous page
page_557
next page > Page 557
oxidant during transient formation of a six-coordinate complex [Eq. (2)]. An alternative role involving one-electron oxidation of guanine might be possible in the case of nickel bleomycin although it does not appear to be operative with NiCR and related compounds. 6 Conclusions Nickel(II) complexes such as NiCR display surprising selectivity for accessible guanine residues in folded nucleic acid structures, and their reactivity is consistent with G N7 being the recognition site. Although binding of a transient nickel(III) intermediate is proposed, this binding does not appear to alter the conformation of the nucleic acid, and studies with conformationally mobile guanine sites may be accurately probed. The experimental method is applicable to both DNA and RNA structures of any size, and sites of reaction may be determined by either alkaline treatment or primer extension techniques. The possibility of altering the site of reaction through appendage of a sequence-specific DNA-binding agent has been demonstrated, and the future possibilities are expansive. Abbreviations
ASIF
accessible surface-integrated field
C
cytosine
CR
2,12-dimethyl-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca- 1(17),2,11,13,15pentaene
cyclam 1,4,8,11-tetraazacyclotetradecane DEPC
diethylpyrocarbonate
DMS
dimethyl sulfate
G
guanine
GGH
glycylglycylhistidine
MMPP magnesium monoperoxyphthalate NMR
nuclear magnetic resonance
ODN
oligodeoxynucleotide
salen
N,N'-ethylenebis(salicylaldimine)
tren
tris(aminoethyl)amine
< previous page
page_557
next page >
< previous page
page_558
next page > Page 558
tRNA
transfer RNA
VPK
viral pseudoknot
References 1. P. E. Nielsen, J. Mol. Recog., 3, 1 (1990). 2. C. Ehresmann, F. Baudin, M. Mougel, P. Romby, J.-P. Ebel, and E. Ehresmann, Nucl. Acids Res., 22, 9109 (1987). 3. J. J. Hayes, Chem. Biol., 2, 127 (1995). 4. S. Adam, P. Bourtayre, J. Liquier, and E. Taillandier, Nucl. Acids Res., 14, 3501 (1986). 5. J. A. Taboury, P. Bourtayre, J. Liquier, and E. Taillandier, Nucl. Acids Res., 12, 4247 (1984). 6. K. S. Kasprzak, Chem. Res. Toxicol., 4, 604 (1991). 7. R. B. Ciccarelli and K. E. Wetterhahn, Cancer Res., 44, 3892 (1984). 8. C. J. Burrows and S. E. Rokita, Acc. Chem. Res., 27, 295 (1994). 9. J. L. Karn and D. H. Busch, Nature, 211, 160 (1966). 10. X. Chen, S. E. Rokita, and C. J. Burrows, J. Am. Chem. Soc., 113, 5884 (1991). 11. S. E. Rokita, P. Zheng, N. Tang, C.-C. Cheng, R.-H. Yeh, J. G. Muller, and C. J. Burrows in Metals and Genetics (B. Sarkar, ed.), Marcel Dekker, New York, 1995. 12. J. G. Muller, X. Chen, A. C. Dadiz, S. E. Rokita, and C. J. Burrows, J. Am. Chem. Soc., 114, 6407 (1992). 13. C. J. Burrows, J. C. Muller, H.-C. Shih, and S. E. Rokita in Supramolecular Stereochemistry (J. Siegel, ed.), Kluwer, Dordrecht, 1995, pp. 5762. 14. T. F. Kagawa, B. H. Geiertanger, A. H.-J. Wang, and P. S. Ho, J. Biol. Chem., 266, 20175 (1991). 15. Y.-G. Gao, M. Sriram, and A. H.-J. Wang, Nucl. Acids Res., 21, 4093 (1993). 16. X. Chen, C. J. Burrows, and S. E. Rokita, J. Am. Chem. Soc., 114, 322 (1992). 17. P. Zheng, C. J. Burrows, and S. E. Rokita, unpublished results.
< previous page
page_558
next page >
< previous page
page_559
next page > Page 559
18. S. A. Woodson and D. M. Crothers, Biochemistry, 27, 436 (1988). 19. R. L. Wurdeman, M. C. Douskey, and B. Gold, Nucl. Acids Res., 21, 4975 (1993). 20. L. L. Guan, J. Kuwahara, and Y. Sugiura, Biochemistry, 32, 6141 (1993). 21. W. Bal, M. I. Djuran, D. W. Margerum, E. T. Gray, Jr., M. A. Mazid, R. T. Tom, E. Nieboer, and P. J. Sadler, J. Chem. Soc., Chem. Commun., 1889 (1994), and references therein. 22. D. P. Mack and P. B. Dervan, J. Am. Chem. Soc., 112, 6369 (1990). 23. D. P. Mack and P. B. Dervan, Biochemistry, 31, 9399 (1992). 24. M. Nagaoka, M. Hagihara, J. Kuwahara, and Y. Sugiura, J. Am. Chem. Soc., 116, 4085 (1994). 25. S. H. Kim, F. L. Suddath, F. J. Quigley, A. McPherson, J. L. Sussman, A. H. J. Wang, N. C. Seeman, and A. Rich, Science, 185, 435 (1974). 26. X. Chen, S. A. Woodson, C. J. Burrows, and S. E. Rokita, Biochemistry, 32, 7610 (1993). 27. R. Lavery and A. Pullman, Biophys. Chem., 19, 171 (1985). 28. T. R. Cech, Annu. Rev. Biochem., 59, 259 (1990). 29. F. Michel and E. Westhof, J. Mol. Biol., 216, 585 (1990). 30. S. E. Butcher and J. M. Burke, J. Mol. Biol., 244, 52 (1994). 31. X. Chen, M. Chamorro, S. E. Lee, L. X. Shen, J. V. Hines, I. Tinoco, Jr., and H. E. Varmus, EMBO J., 14, 842 (1995). 32. M. Schmidt, P. Zheng, and N. Delihas, Biochemistry, 34, 3521 (1995). 33. M. Kouchakdjian, V. Boudepoudi, S. Shibutani, M. Eisenberg, F. Johnson, A. P. Grollman, and D. G. Patel, Biochemistry, 30, 1403 (1991). 34. A. M. Maxam and W. Gilbert, Meth. Enzymol., 65, 499 (1980). 35. J. G. Muller, S. J. Paikoff, S. E. Rokita, and C. J. Burrows, J. Inorg. Biochem, 54, 199 (1994). 36. O. Augusto, E. L. Cavlier, E. G. Rogan, N. S. V. RamaKrishna, and C. Kolar, J. Biol. Chem., 265, 22093 (1990). 37. S. A. Woodson, J. G. Muller, C. J. Burrows, and S. E. Rokita, Nucl. Acids Res., 21, 5524 (1993). 38. P. B. Dervan, Science, 232, 464 (1986).
< previous page
page_559
next page >
< previous page
page_560
next page > Page 560
39. M. A. Price and T. D. Tullius, Meth. Enzymol., 212, 194 (1992). 40. J. A. Latham and T. R. Cech, Science, 245, 276 (1989). 41. J. G. Muller, X. Chen, A. C. Dadiz, S. E. Rokita, and C. J. Burrows, PureAppl. Chem., 65, 545 (1993).
< previous page
page_560
next page >
< previous page
page_561
next page > Page 561
19 Hydrolytic Cleavage of RNA Catalyzed by Metal Ion Complexes Janet R. Morrow Department of Chemistry, Natural Sciences and Mathematics Complex, State University of New York at Buffalo, Buffalo, NY 14260, USA
1. Introduction
1.1. RNA Cleavage by Metal Ions
1.2. Nucleophilic Displacement Reactions of Phosphate Diesters
1.3. Aqua-Metal Ions
2. RNA Cleavage by Metal Ion Complexes
2.1. Catalysis by Electrophilic Activation
2.1.1. RNA and RNA Models
2.1.2. Phosphate Diester Complexation
2.1.3. Dimeric Metal Ion Complexes
2.1.4. Ligand Design
2.2. Ligand-Facilitated Hydrolysis
2.2.1. Hydroxide or Alkoxide Ligands
2.2.2. Bifunctional Catalysts
2.3. Lanthanide(III) Complexes
2.3.1. Macrocyclic Ligands
562
562
563
565
566
566
566
567
569
570
571
571
572
573
573
< previous page
page_561
next page >
< previous page
page_562
next page > Page 562 578
2.3.2. Choice of Lanthanide Ion
580
2.3.3. Catalytic Turnover
580
3. Catalyst Specificity
580
3.1. RNA Sequence and Structure
581
3.2. Sequence-Specific RNA Cleavage
585
4. Summary and Outlook
585
Abbreviations
586
References
1 Introduction 1.1 RNA Cleavage by Metal Ions There is much interest in the study of metal ions and metal ion complexes that hydrolytically cleave RNA. The first examples of metal ionfacilitated cleavage of RNA were reported 40 years ago and involved the cleavage of dinucleotides [1,2] and homopolymers [3,4] of RNA by aquametal ions. Several years later, reports of the hydrolysis of tRNAphe by Pb2+ ions suggested that strong metal ion binding sites in RNA could facilitate highly site-specific hydrolytic cleavage [57]. Following these discoveries ribozymes, RNA molecules that catalyze hydrolysis or phosphoryl transfer reactions at the phosphate diester backbone of RNA, were first reported [8,9]. Ribozymes require metal ions for activity and contain metal ions at the active site as well as in structural roles. The role of metal ions in catalyzing phosphoryl transfer reactions in ribozymes is an active area of research [1014]. More recently, metal ion complexes have been investigated as catalysts for the hydrolytic cleavage of RNA dinucleotides and oligomers [1524]. Mechanisms of metal ion complex-facilitated RNA cleavage may be similar to those for cleavage of RNA by metal ions in ribozymes. However, there are additional considerations that are important for metal ion complex catalysts. Ligands are designed to minimize metal ion dissociation from the complex and to enhance catalytic hydrolysis. Metal ion complex catalysts may be useful for the sequence-specific or structure-specific hydrolytic cleavage of RNA. Here the speci-
< previous page
page_562
next page >
< previous page
page_563
next page > Page 563
ficity of cleavage derives from a nucleic acid recognition agent attached to the metal ion complex. The nucleic acid recognition agent may be an oligodeoxyribonucleotide with a base sequence that is complementary to that of a message RNA (an antisense oligonucleotide) [2528]. Such sequence-specific agents that catalyze the destruction of an m-RNA may be useful as a new class of antisense oligonucleotides [29]. In addition, the use of metal ion complexes has contributed to elucidation of the mechanisms by which metal ions catalyze reactions of phosphate diesters. In this chapter, metal ion complexes as catalysts for the hydrolytic cleavage of RNA are reviewed. The general mechanisms of hydrolytic cleavage by metal ions and the important considerations in ligand design will be covered in the first part of the chapter. This will be followed by a discussion of the effect of RNA structure and sequence on cleavage rates and by new data on sequence-specific cleavage of RNA. Work reported after 1991 will be emphasized as much of the work prior to that date has been covered in a previous review article [30]. 1.2 Nucleophilic Displacement Reactions of Phosphate Diesters The metal ion complex catalysts that will be discussed in this chapter are those that cleave RNA by facilitating phosphate diester transesterification (Fig. 1). In this type of reaction, intramolecular attack of the 2'-hydroxyl of ribose on the phosphate diester produces a cyclic phosphate ester and a 5'-hydroxyl end group with concomitant cleavage of the RNA strand. The reaction in Fig. 1 is a transesterification or transphosphorylation reaction rather than a hydrolysis reaction, which would have water or hydroxide as a nucleophile. Nevertheless, this type of cleavage is commonly referred to as RNA hydrolysis or RNA hydrolytic cleavage. This terminology may arise from the mechanism of cleavage of RNA by ribonucleases such as ribonuclease A where cleavage of the phosphate diester backbone (transesterification) is followed by a true hydrolysis reaction involving attack of water or hydroxide on the cyclic phosphate ester [31]. Metal ion complexes that promote hydrolytic cleavage of RNA are more likely to be selective for RNA than are metal complexes that
< previous page
page_563
next page >
< previous page
page_564
next page > Page 564
Fig. 1. Hydrolytic cleavage of RNA. promote oxidative cleavage of nucleic acids. Since cleavage of the phosphate diester backbone of RNA is facilitated by the intramolecular nature of the nucleophilic displacement reaction, compounds that promote RNA cleavage will not necessarily promote hydrolysis of DNA or other biopolymers. For the design of hydrolysis catalysts it is preferable to choose metal ions that are not redox-active in order to avoid the possible production of oxygen radicals. Such diffusible oxygen radicals or, alternately, reactive metal oxygen species may nondiscriminately destroy RNA and DNA as well as the catalyst itself [32,33]. Metal ions facilitate other types of nucleophilic displacement reactions of phosphate esters including the second step in RNA cleavage, which is a true hydrolysis reaction [17,34,35]. However, most of the work to date is concerned with the hydrolysis of phosphate diesters with good leaving groups such as 4-nitrophenylate and there are relatively few examples of the hydrolysis of inactivated esters [30]. Hydrolysis of both activated and inactivated phosphate diesters by metal ion complexes is thought to occur by attack of a metal-bound hydroxide or water
< previous page
page_564
next page >
< previous page
page_565
next page > Page 565
ligand on the coordinated phosphate ester. There are also a few examples of displacement reactions with metal ionbound amido [36], peroxide [3739], or alkoxide [40] as nucleophiles. Metal ion-bound peroxide ligands are especially strong nucleophiles toward phosphate esters and have been shown to induce DNA cleavage. Several recent reviews cover metal ion-promoted displacement reactions of phosphorus(V) compounds [30,41,42]. As this is a rapidly expanding area, the scope of this chapter will be limited to the transesterification or hydrolytic cleavage of RNA by metal ion complexes. 1.3 Aqua-Metal Ions The first examples of RNA cleavage by metal ions used aqua-metal ions with no ligands other than buffer. A large number of divalent and trivalent aqua-metal ions promoted RNA dinucleotide cleavage including La3+, Bi3+, Zn2+, and Pb2+ [14]. Later work by Eichhorn, Butzow, and coworkers detailed the cleavage of homopolymers of RNA by divalent metal ions [4346]. More recent work involved the study of Ln3+- (Ln = lanthanide), Zn2+-, Cu2+-, and Ni2+ -promoted cleavage of RNA dinucleotides or RNA oligomers [15,16,4749]. In these studies that do not make use of strongly chelating ligands, precipitation of metal ion hydroxide or metal ion nucleotide complexes may occur at neutral or slightly basic pH values to give heterogeneous mixtures. Thus it is difficult to interpret much of the kinetic data of RNA cleavage by aquametal ion complexes. This work is summarized in a review article [30] and will not be discussed further here. Self-cleaving RNAs are mentioned briefly here since the cleavage reaction is phosphate diester transesterification as shown in Fig. 1. The reader is referred to recent review articles for more information about this topic [12,50,51]. Selfcleaving RNAs have an absolute requirement for metal ions, but the metal ion requirements are frequently not very specific. For example, there are reports that self-cleaving RNAs will cleave in the presence of a wide variety of metal cations [52]. However, small ribozymes that cleave only in the presence of Mn2+ have been reported [53] and there are ribozymes that cleave only in the presence of Pb2+ as developed by in vitro selection methods [5456]. The hammerhead ribozyme is among the best characterized of the self-cleaving
< previous page
page_565
next page >
< previous page
page_566
next page > Page 566
RNA from both a kinetic and structural standpoint [57,58]. A small oligoribonucleotide that forms a hammerheadlike structure cleaves in the presence of a limited number of cations including Mg2+, Ca2+, and Mn2+ [58]. 2 RNA Cleavage by Metal Ion Complexes 2.1 Catalysis by Electrophilic Activation 2.1.1 RNA and RNA Models Study of the mechanism of RNA cleavage by metal ions is made challenging by the presence of many different metal ion binding sites on nucleic acids and by the presence of multiple cleavage sites. As a result, workers have frequently relied on the use of model compounds for mechanistic studies of RNA hydrolysis. The most frequently used RNA model compound, hpnp (2-hydroxypropyl-4-nitrophenylphosphate), was first prepared by Usher and Brown [59]. Other model compounds as discussed below contain cleverly placed ligands that position the metal ion to test a specific catalytic mode. Comparison of the hydrolysis of model compounds to RNA hydrolytic cleavage must be made cautiously, however. The availability of other competing binding sites such as RNA nitrogenous bases may affect the rate of RNA cleavage by a metal ion. More important, RNA structure dramatically modifies metal ion cleavage rates. This latter topic will be discussed in Sec. 3.1. In the present section, complexes that promote cleavage of model RNA compounds or short RNA oligomers will be discussed. Hydrolytic cleavage of RNA or RNA models is established by product analysis. This is especially important for metal ions that are redoxactive and may promote oxidative cleavage of RNA as well as hydrolytic cleavage. For RNA model substrates and for dinucleotides of RNA, 31P nuclear magnetic resonance (NMR) spectroscopy or highperformance liquid chromatography (HPLC) methods are useful for product identification [1619,60]. For longer RNAs, products that result from hydrolytic cleavage such as cyclic phosphate esters may be identified by use of HPLC [15]. High-resolution gel electrophoresis is also useful to compare the end groups from metal ion-catalyzed cleavage to the known end groups resulting from enzymatic cleavage [21,22].
< previous page
page_566
next page >
< previous page
page_567
next page > Page 567
2.1.2 Phosphate Diester Complexation Electrophilic activation of the phosphate diester by the metal cation is of primary importance in metal ion-catalyzed hydrolytic cleavage of RNA. Binding of a Lewis acidic metal ion to a terminal oxygen of the phosphate ester polarizes the P-O bond and activates the phosphorus(V) center to nucleophilic attack by the 2'-hydroxyl group. This type of displacement reaction may go by way of a concerted associative process with a pentacovalent phosphorus as a transition state. The metal ion may bind to and stabilize this transition state and may neutralize the negative charge on the pentacovalent phosphorus. Electrophilic catalysis of the hydrolytic cleavage of RNA and RNA model compounds by protons has been studied in detail [6165]. Kinetic studies on RNA models suggest that protonation of one of the phosphoryl oxygens would accelerate phosphate ester transesterification by approximately 105 [6164]. Although metal ions do not polarize ligands as effectively as do protons, other modes of catalysis such as bidentate binding to a pentacovalent phosphorus transition state [60] or bifunctional mechanisms as discussed in the next section may be important. Several studies suggest that the first step in metal ion-promoted phosphate diester transesterification is the formation of a complex between the metal ion and the phosphate diester. Saturation kinetics for the transesterification of the RNA model compound hpnp at 37°C, pH 6.85 by metal ions suggest the preequilibrium formation of a reactive complex [60]. At saturating concentrations of La(III) or Pb(II) ions a greater than 104 rate enhancement for phosphate ester transesterification is observed. Similar kinetic studies in addition to phosphate diester binding studies with lanthanide(III) macrocyclic complexes indicate that a complex between the phosphate diester and the metal ion macrocyclic complex is initially formed [23,65]. A further example has a hydroxyquinoline ligand placed strategically to enhance binding of a zinc(II) ion to the phosphate ester of a dinucleotide (Fig. 2) [66]. With Zn(II) present at saturating concentrations, a rate enhancement of greater than 104 is observed. An interesting question is whether the metal ion is coordinated directly to the phosphate diester in these complexes. The metal ion may also interact indirectly through a water ligand or other type of ligand that may hydrogen-bond to the phosphate diester. This point was dis-
< previous page
page_567
next page >
< previous page
page_568
next page > Page 568
Fig. 2. Hydrolytic cleavage of RNA by a zinc(II) complex as in Ref. 66. cussed lucidly in a recent commentary [67]. It is relevant to mention here that hydrolytic cleavage of RNA and RNA models has been accomplished by the use of organic receptors that bind the phosphate ester through a combination of electrostatic and hydrogen bonding interactions [6870]. However, the rates of hydrolysis are substantially slower than those with metal ion complexes. Since direct metal ion complexation to the phosphate ester will be more effective in polarizing the P-O bond than indirect coordination, it seems likely that metal ion complexes with exchange-labile sites suitable for the direct coordination of a phosphate diester will be more potent catalysts for RNA cleavage. In support of this hypothesis, we have observed that the transesterification of the RNA model hpnp by inert metal ion complexes such as [Co(NH3)5(OH]2+ is very slow in comparison to catalysis by metal ions that may bind directly to the phosphate ester [60]. Metal ion facilitation of leaving group departure may also be an important mode of metal ion-catalyzed RNA hydrolytic cleavage. It is in general difficult to discriminate between catalysis by binding to the phosphate ester terminal oxygen and catalysis by binding to the leaving group. An approach to the study of leaving group effects involves the
< previous page
page_568
next page >
< previous page
page_569
next page > Page 569
strategic placement of a ligand adjacent to the leaving group. Although no RNA model compounds have been studied, there are several studies of phosphate esters or phosphonate esters containing ligands that bind the metal ion in close proximity to the leaving group [71,72]. Rate accelerations of up to 108 have been reported for the hydrolysis of these substrates in the presence of metal ions. 2.1.3 Dimeric Metal Ion Complexes Two metal ions may very effectively promote phosphoryl transfer reactions at phosphate esters. In such processes catalyzed by metalloenzymes, the reaction is facilitated by two metal ions which are 3.9 Æ apart [73]. It is proposed that one metal ion activates the nucleophile and the second metal ion binds to and stabilizes the leaving group. Both metal ions may coordinate to and stabilize a pentacoordinate phosphorus transition state. There are currently a few examples of dimeric metal ion complexes where kinetic data support the proposal that two metal ions may act in tandem to promote phosphate diester substitution reactions. For the RNA model compound hpnp, transesterification is accelerated 50-fold more rapidly by a dimeric copper(II) complex than by the monomeric copper(II) complex [74]. In addition, a structurally similar monomeric copper(II) complex promotes transesterification of hpnp with a secondorder dependence on copper(II) complex, again suggesting that a dimeric complex may be more effective in promoting phosphate diester transesterification than the monomeric complex [75]. Double Lewis acid activation is proposed in these examples since the transesterification of hpnp goes by intramolecular nucleophilic attack and a good leaving group is present (Fig. 3). Other examples where two metal ions accelerate displacement reactions at phosphate esters include early studies by Jones et al. [76] on cobalt(III) dimers and more recent studies by Vance and Czarnik [77] on the use of dimers of cobalt(III) macrocyclic amine complexes to hydrolyze a bis(4nitrophenyl)phosphate ester. The Jones study indicates an additional 50-fold in acceleration in the presence of two metal ions and the Vance and Czarnick example has a 10-fold acceleration for the dimeric complex over the monomeric complex. A third example describes the involvement of two Co(III) amine complexes in the hydrolysis of phosphate monoesters with poor leaving groups [78]. Finally, dimeric metal
< previous page
page_569
next page >
< previous page
page_570
next page > Page 570
Fig. 3. Double Lewis acid activation by two copper(II) centers as in Refs. 74 and 75. ion complexes may be involved [79] in DNA hydrolysis [37], amide hydrolysis [80], and in the hydrolysis of activated phosphate diesters [38,81] or phosphonate esters [82]. 2.1.4 Ligand Design Nucleic acids, being polyanions, bind strongly to metal cations. Some types of RNA have strong metal ion binding pockets, e.g., formation constants of approximately 108 M1 have been reported for binding of Eu(III) ions to tRNA [83]. Metal ion complexes that are to be used in the cleavage of RNA must therefore have ligands that bind strongly to the metal ion if the metal ion complex is to remain intact. Chelating ligands have been shown to increase catalyst longevity presumably by preventing precipitation of the metal ion as oligomeric metal ion hydroxide complexes or as metal ion polynucleotide complexes [17]. In addition, a chelating ligand provides a means of attaching the metal ion to a nucleic acid recognition agent such as an antisense oligonucleotide for sequence-specific RNA cleavage as discussed in the final section. For metal ion complex catalysts, then, it is desirable to have a high formation constant or a low metal ion dissociation rate. Clearly the use of metal ion complexes with sluggish kinetics of dissociation have certain advantages over the use of those that have high formation constants but lack kinetic inertness. Complexes that are kinetically inert to metal ion release remain intact to a larger degree in the presence of the competing ligands found in vivo than do complexes that are not inert to metal ion dissociation [84].
< previous page
page_570
next page >
< previous page
page_571
next page > Page 571
In addition to the formation of a robust metal ion complex, there are several considerations in choosing a chelating ligand. Coordination of the metal ion to a polydentate chelating ligand decreases the number of binding sites for catalysis and may inhibit catalysis completely [65]. In addition, polydentate ligands modify metal ion complex cleavage rates by influencing the Lewis acidity of the metal ion. In a series of Zn(II) macrocyclic amine complexes, the Lewis acidity of the Zn(II) ion, as indicated by the pKa of the metal-bound water, decreases with increasing number of nitrogen donor ligands [17]. Dinucleotide cleavage rates correlate roughly with the Lewis acidity of the complexes. The type of donor ligand also modifies the Lewis acidity of the metal ion complex. Anionic ligands increase the electron density on the metal ion center and decrease the pKa of metal-bound water molecules relative to neutral ligands [85]. Use of anionic ligands may thus have a detrimental effect on metal ion complex catalysis since both the Lewis acidity of the metal ion complex and the overall positive charge on the complex decreases. A positive charge enhances binding of the metal ion complex to RNA and may aid in catalysis by neutralization of the negative charge on the pentacoordinate phosphorus transition state or intermediate. Positively charged La(III), Zn(II), Cu(II), or Ni(II) complexes promote hydrolytic cleavage of oligonucleotides or dinucleotides of RNA much more effectively than do anionic or neutral complexes [15,48]. Finally, as discussed in the next section, a ligand may contain groups that will participate in catalysis. 2.2 Ligand-Facilitated Hydrolysis 2.2.1 Hydroxide or Alkoxide Ligands RNA is susceptible to general base catalysis by organic bases such as imidazole [86]. Metal ion-bound water or hydroxide may also facilitate RNA cleavage by general base catalysis. In an elegant solid state study [6], it was shown that a Pb(II) hydroxide ligand is positioned to deprotonate the 2'-hydroxyl of the ribose to promote cleavage in the solid state. However, for metal ion complexes in solution, electrophilic mechanisms are generally also operative and the importance of general base catalysis is not as clear. If general base catalysis were the only mode of catalysis it would be difficult to reconcile the large rate acceleration of
< previous page
page_571
next page >
< previous page
page_572
next page > Page 572
RNA hydrolytic cleavage in the presence of Pb(II) hydroxide with the relatively modest cleavage rates that are observed in the presence of organic bases with comparable pKa values. Electrophilic activation by coordination of the metal ion to the phosphate ester is clearly of primary importance. Inert metal ion complexes such as [Co(NH3)5(OH)]2+ that bear a hydroxide ligand and no sites for coordination to the phosphate ester show very little activity in promoting phosphate ester transesterification [60]. In contrast, [Co(trien)(OH2)(OH)]2+, which binds phosphate diesters, promotes hydrolysis of RNA dinucleotides [16]. Several metal ion complexes with a hydroxide ligand or an alkoxide ligand have been studied as catalysts for RNA cleavage. The pH-rate profiles for the transesterification of hpnp by the Ln(III) complex of 1,4,7,10-tetrakis(2hydroxyethyl)-1,4,7,10-tetraazacyclododecane (THED) and 1S,4S,7S,10S-tetrakis(2-hydroxypropyl)-1,4,7,10tetraazacyclododecane (s-THP) complexes are sigmoidal; the rate constant becomes independent of pH when the water or alcohol ligand is completely deprotonated [23]. This may be attributed to action of the hydroxide or alkoxide ligand as a general base. Alternately, a kinetically equivalent mechanism has free hydroxide as the base and the protonated form of the macrocyclic complex as the active catalyst. Other examples of hydroxide ligands that may act as general bases include the Zn(II) complex of triazacyclononane [17] and the Cu(II) terpyridyl complex [15,18]. For the Cu(II) terpyridyl complex, which has a bound water with a pKa of 8, a bifunctional mechanism with the complex acting in electrophilic activation of the phosphate ester and as a general base catalyst has been proposed based on the observation of a bell-shaped pH profile. 2.2.2 Bifunctional Catalysts An alternate approach is the construction of ligands that contain a basic group to facilitate deprotonation of the 2'hydroxyl of ribose. These ligands bind metal ions and promote hydrolytic cleavage of RNA with electrophilic catalysis by the metal ion in conjunction with general base catalysis by the ligand. Along these lines, a bifunctional Zn(II) macrocyclic amine complex catalyst has been reported that hydrolyzes the RNA model compound hpnp approximately 20-fold more rapidly than does the unmodified Zn(II) macrocyclic amine complex [87]. A more recent example has an imidazole group tethered to an adenosine base of
< previous page
page_572
next page >
< previous page
page_573
next page > Page 573
Fig. 4. Zinc(II) promoted cleavage of C8-histamino-UpA as in Ref. 88. a dinucleotide of RNA (Fig. 4). In the presence of 1.0 mM ZnCl2, the imidazole modified RNA hydrolyses 10-to 15fold more rapidly than the unmodified RNA [88]. 2.3 Lanthanide(III) Complexes 2.3.1 Macrocyclic Ligands Several lanthanide ion complexes have been prepared for study as catalysts for the hydrolytic cleavage of RNA (Figs. 57) [19,2326]. The complexes that are effective for RNA cleavage have the following important properties in common. First, the complexes are kinetically inert to metal ion dissociation. [Ln(HAM)](OAc)Cl2, [Ln(THED)](CF3SO3)3, and [Ln(s-THP](CF3SO3)3 are prepared in nonprotic solvents, in part because the complexes have low formation constants in water. Nonetheless all complexes are resistant to metal ion release in water in the
< previous page
page_573
next page >
< previous page
page_574
next page > Page 574
Fig. 5. Lanthanide(III) complex catalysts for the hydrolytic cleavage of RNA as described in Refs. 19, 25, and 26. presence of competing ligands. Second, the overall charge on the complexes is positive. Ligands containing neutral, soft nitrogen donor atoms are prevalent. Finally, all complexes have at least one available coordination site for phosphate ester binding. The texaphyrin complexes and the Eu(III) complex of the hexadentate 2,2':6,2''-terpyridine
< previous page
page_574
next page >
< previous page
page_575
next page > Page 575
Fig. 6. Lanthanide(III) complexes of macrocycles with pendant amide groups as described in Refs. 24, 8991. macrocycles (OAMs) have been attached to oligonucleotides to effect sequence-specific cleavage and will be discussed in the final section. Ligand design will be discussed here for the macrocyclic complexes shown in Figs. 6 and 7, which have been studied in my laboratory [23,24,8993]. There are several properties that make lanthanide ions an excellent choice as RNA hydrolysis catalysts. First, trivalent lanthanide ions
< previous page
page_575
next page >
< previous page
page_576
next page > Page 576
Fig. 7. Lanthanide(III) complexes of macrocycles with pendant hydroxyalkyl groups as described in Refs. 23, 92, and 93. are good Lewis acids with very rapid ligand exchange rates. Second, coordination geometries are flexible and hard bases such as oxygen donor ligands are preferred. Since ionic bonding predominates in lanthanide ion complexes, anionic oxygen donor ligands that are hard bases form strong complexes with the lanthanide ions, and complexes of polyaminocarboxylate ligands such as those of 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate (DOTA) or diethylenetriaminepentaacetate (DTPA) have some of the highest known formation constants with the trivalent lanthanide ions [94]. However, negatively charged polyaminocarboxylate complexes of the trivalent lanthanides such as the lanthanum(III) complex of ethylenediamine-N,N,N',N'-tetraacetate (EDTA) do not promote dinucleotide cleavage under conditions where other lanthanide(III) complexes are catalytically active [48]. A good
< previous page
page_576
next page >
< previous page
page_577
next page > Page 577
alternative for the negatively charged carboxylate donor group is an amide group; amide groups have been used in conjunction with carboxylate groups to form strong ligands for the lanthanides [9598]. In my laboratory we have explored the use of macrocyclic amide ligands for the lanthanides (Fig. 6). In addition, we have prepared lanthanide(III) complexes of macrocyclic ligands containing hydroxyethyl or hydroxypropyl groups (Fig. 7). Chelate ring and macrocyclic ring size are important considerations in ligand design. For large metal ions such as the trivalent lanthanides, five-membered chelate rings are favored over six-membered rings. This is dramatically demonstrated by the differences in the properties of the lanthanum(III) complexes of the two tetraamide macrocycles which differ by the number of methylenes in the pendant group. The La(III) complex of 1,4,7,10-tetrakis(2carbamoylethyl)-1,4,7,10-tetraazacyclododecane (TCEC) dissociates rapidly in a couple of hours in water [89] whereas the La(III) complex of 1,4,7,10-tetrakis(carbamoylmethyl)-1,4,7,10-tetraazacyclododecane (TCMC) dissociates much more slowly over a period of several days at 37°C in the presence of Cu2+ as a trapping agent [24]. The necessity of using the 12-membered tetraazamacrocyclic ligand cyclen has also been demonstrated. Lanthanide(III) complexes of derivatives of the 14-membered tetraazamacrocycle cyclam that bear carboxylate or hydroxyethyl pendent groups are substantially less stable than are the analogous cyclen derivatives [92, 99]. The triazacyclododecane macrocyclic ring does not readily accommodate large metal ions; yttrium(III) or lutetium(III) complexes of 1,4,7-tris(carbamoylmethyl)-1,4,7-triazacyclononane (TCMT) dissociate readily in water [90]. Dissociation rates depend on the donor atom type in the pendent groups of the cyclen macrocycle. For example, the macrocyclic ligand containing hydroxypropyl pendant groups (s-THP) forms complexes that are markedly more inert to dissociation [93] than are the hydroxyethyl derivatives (THED) [92]. The half-life for dissociation of the [La(THED)]3+ complex is 21 hr while that for the [La(s-THP)]3+ complex is 73 days. This may be related to the extreme rigidity of the s-THP ligand. Consistent with its high degree of rigidity, the s-THP ligand has less of a strict preference for a particular lanthanide ion. Little difference is observed for the dissociation of the La(III), Eu(III), or Lu(III) complexes of s-THP whereas dissociation rates for the THED or TCMC series are highly dependent on lanthanide(III) ion. The [Eu(TCMC)]3+
< previous page
page_577
next page >
< previous page
page_578
next page > Page 578
ion is the most inert to dissociation of all the complexes. No dissociation is detected over a period of 2 months at 37°C, pH 6.0 in the presence of excess Cu2+ [24]. 2.3.2 Choice of Lanthanide Ion The trivalent lanthanide ion salts (LaCl3, NdCl3, EuCl3, GdCl3, TbCl3, YbCl3, LuCl3) promote transesterification of the model RNA substrate hpnp with pseudo-first-order rate constants that differ by less than a factor of 5 [60]. A greater difference in rates for various lanthanide ions is observed for dinucleotide cleavage in the presence of high concentrations of lanthanide ions; Tm3+ promotes cleavage approximately 100-fold more rapidly than does La3+ [49]. One might have expected large differences in cleavage rates given the large variation in the ionic radii of the trivalent lanthanide ions and the corresponding differences in Lewis acidity of the lanthanide ions in water [100]. In contrast, with the s-THP, THED, and TCMC ligands that are octadentate and nearly encapsulate the lanthanide ion, the choice of a lanthanide ion is crucial to catalyst design. The most dramatic example of this is found in the Ln(III) complexes of TCMC. The La(III) complex of TCMC is 10-coordinate (Fig. 8) and promotes cleavage of RNA oligomers [24]. In contrast, the Eu(III) or the Dy(III) complexes of TCMC do not promote cleavage of RNA oligomers under similar conditions. The Eu(III) complex of TCMC is nine-coordinate in the solid state with one bound water molecule. By use of laser-induced luminescence spectroscopy, we have shown that the Eu(III) ion binds between one and two water molecules in solution [65]. Curiously, phosphate esters such as diethyl phosphate do not bind appreciably to [Eu(TCMC)]3+. Under similar conditions, [La(TCMC)]3+ binds to diethyl phosphate as do other lanthanide(III) complexes that catalyze RNA cleavage. The inability of [Eu(TCMC)]3+ complex to bind to diethyl phosphate may be attributed to the inaccessibility of the Eu(III) ion for binding or, alternatively, the low Lewis acidity of the complex. Making an additional coordination site available creates a catalytically active Eu(III) macrocyclic complex. The Eu(III) complex of the heptadentate ligand 1-(4-nitrobenzyl)-4,7,10tris(carbamoylmethyl)-1,4,7,10-tetraazacyclododecane (NBAC) promotes cleavage of RNA oligomers and transesterification of hpnp [91]. The same trend is observed for the s-THP series of lanthanide(III) complexes. The La(III)
< previous page
page_578
next page >
< previous page
page_579
next page > Page 579
Fig. 8. The structure of the [La(TCMC)(CH3CH2OH)(CF2SO3)]2+ cation is shown. This complex facilitates cleavage of RNA oligomers at 37°C, pH 7.4. (After Ref. 24.) complex of s-THP is more active than the Eu(III) derivative and the Lu(III) complex is completely inactive in promoting RNA cleavage [23]. Kinetic studies of the hydrolytic cleavage of oligomers of adenylic acid by the lanthanide(III) complexes in Figs. 6 and 7 have been reported [19,23]. Pseudo-first-order rate constants for the cleavage of RNA oligomers with 200 µM complex range from 1.5 hr1 for [Eu(HAM)]3+, 0.57 hr1 for [La(TCMC)]3+, 0.68 hr1 for [Eu(THED)]3+, and 0.26 hr1 for [La(s-THP)]3+ at pH 7.4, 37°C. Pseudo-first-order rate constants are dependent on the concentration of lanthanide(III) complex. Preliminary results suggest that the reaction is first order in complex; decreasing
< previous page
page_579
next page >
< previous page
page_580
next page > Page 580
the concentration of [Eu(THED)]3+ by twofold halves the pseudo-first-order rate constant for cleavage of RNA oligomers. For [Eu(THED)]3+, the pH-rate profile for the cleavage of RNA oligomers is sigmoidal and supports participation of a metal-bound alkoxide as a general base catalyst. 2.3.3 Catalytic Turnover Most studies of RNA cleavage by metal ion complexes have an excess of metal ion complex rather than a catalytic amount of complex. An exception is the study of a Schiff base-Eu(III) complex ([Eu(HAM)]3+) that has been shown to catalyze hydrolytic cleavage of a dinucleotide with several catalytic turnovers [19]. More studies of this nature would be of value especially in view of the applications that require a catalytic cleaving agent [29]. It would be useful to determine whether there will be product inhibition for the metal ion complexes that facilitate cleavage of RNA. 3 Catalyst Specificity 3.1 RNA Sequence and Structure There are relatively few reports of the cleavage of more highly structured RNAs by metal ion complexes. In addition, there is little information concerning the effect of base sequence on cleavage. Useful information may be garnered from the more extensive literature on RNA cleavage by aqua-metal ion complexes. Differences in base sequence in short, unstructured RNA dinucleotides have only a minor effect on cleavage rates by metal ions such as Zn2+, Cu2+, La3+, Ni2+, and Co2+ [101]. In contrast, longer RNAs may have secondary and tertiary structural features that will affect cleavage rates by metal ions. Longer RNA molecules may contain metal ion binding sites that bring the metal ion in close proximity to a specific phosphate diester and this may lead to highly specific RNA cleavage. The highly specific hydrolytic cleavage of tRNA by metal ions [57] and the self-cleaving RNAs [5058] are examples where metal ion binding pockets may be of importance in the cleavage event. Interestingly, there are also short RNA fragments that undergo highly site-specific cleavage by metal ions. For example, a RNA strand 30 bases long cleaves specifically in the presence of Pb2+ [56]. An
< previous page
page_580
next page >
< previous page
page_581
next page > Page 581
even smaller self-cleaving RNA consists of a trinucleotide UUU, Mn2+, and the tetranucleotide GAAA, which is cleaved specifically between G and A [102]. Metal ion complexes containing chelating ligands will most likely exhibit specificities for RNA binding that are different from those of aqua-metal ions. Strongly chelating ligands not only will prevent metal ion binding to RNA pockets but the ligands themselves may impart a unique specificity to cleavage. This has been demonstrated for cleavage of tRNAphe by lanthanide(III)-Schiff base complexes. Primary cleavage sites for [Ln(HAM)]3+ are different from the free Ln3+ ion cleavage sites which are attributed to an Ln3+ ion binding pocket in the tRNA [21,22]. Although most [Ln(HAM)]3+ cleavage sites are in the loop regions of the RNA, cleavage is remarkably specific with only a few primary cleavage sites. At present, there is not enough known to predict RNA cleavage sites by these metal ion complexes. More systematic study of the effect of RNA secondary and tertiary structure on metal ion complex cleavage rates will be required if the metal ion complexes are to be used as probes of RNA structure. In addition, it will be necessary to establish whether the structure of the RNA is modified by binding of the metal ion complex. One type of nucleic acid structure that is less susceptible to hydrolytic cleavage by metal ion complexes is doublestranded RNA as well as RNA in RNA-DNA hybrids. In a double-stranded form, the RNA sugar-phosphate backbone is more rigid than in single-stranded RNA. It has been noted that double-stranded RNA may differ in reactivity from single-stranded RNA since the 2'-hydroxyl of the ribose in a double-stranded RNA is not positioned for in-line attack on the phosphate ester [103,104]. Lanthanide(III) ions and complexes do not cleave RNA in RNADNA hybrids under conditions where cleavage of single-stranded RNA is rapid [21]. This effect has now been noted for several metal ion complexes [2527]. Similar effects are observed for double-stranded RNA in tRNAphe as discussed above. 3.2 Sequence-Specific RNA Cleavage There is much interest in facilitating the sequence-specific cleavage of RNA by use of a nucleic acid recognition agent with a tethered metal ion complex hydrolysis catalyst. For single-stranded RNA recognition, a
< previous page
page_581
next page >
< previous page
next page >
page_582
Page 582 short fragment of either DNA or RNA may be used. The oligonucleotide need be only 17 bases long in order to recognize a unique mRNA in a eukaryotic cell [105]. Such short oligodeoxynucleotides with base sequences that are complementary to a base sequence of a mRNA are called antisense oligonucleotides. Antisense oligonucleotides are used to inactivate as well as activate specific genes. The mechanism of antisense oligonucleotide action is actively being investigated and may include binding of the oligonucleotide to the RNA during mRNA processing events, during mRNA transport or mRNA translation. Antisense oligonucleotides are effective tools for the study of gene expression and are of interest for various therapeutic applications. There is much interest in testing the proposal that the addition of a hydrolytic cleaving agent may improve the efficacy of action of the oligonucleotide [29]. There are now several examples of sequence-specific cleavage of RNA by metal ion complexes attached to oligodeoxynucleotides [2528,106]. In Table 1 are summarized the reaction times and percent cleavage of RNA in the presence of several metal ion complex oligonucleotide conjugates. All studies are done with an excess of the conjugate to RNA. Most of the RNA targets are relatively short with the exception of the target RNA for the copper(II) conjugate, which is 159 bases long. The use of short RNA targets circumvents the difficulties that may be associated with the use of longer RNA fragments containing a large degree of RNA tertiary or secondary structure. All of the TABLE 1 Sequence-Specific Cleavage of RNA by Metal Ion Complex Conjugates at 37°C Conjugated complex
Concentration (µM)
Eu(texaphyrin)2+ Eu(OAM)3+ La(IDA)+ Cu(TRPY)2+ Eu(THED)3+
0.025 0.4 10 5 4
RNA
Time (hr) % Cleavage
30 mer
24
30
29 mer
16
5188
39 mer
8
17
159 mer
72
11
25 mer
20
30
From [2528,106], respectively.
< previous page
page_582
next page >
< previous page
page_583
next page > Page 583
lanthanide(III) macrocyclic complexes are inert to metal ion release and the intact metal ion complex is conjugated directly to the oligonucleotide. Both the [Eu(THED)]3+ conjugate and the OAM conjugate are characterized by mass spectrometry. In contrast, the lanthanum(III) iminodiacetate (IDA) complex and the copper(II) complex are prepared in situ by addition of excess metal ion to the conjugate. Interestingly, even with addition of excess metal ion, sequence-specific cleavage appears to predominate, suggesting that the close proximity of the tethered metal ion complex to the RNA target accelerates cleavage. Indeed, the Eu(III) texaphyrin conjugate promoted cleavage approximately 10,000-fold more efficiently than did the free Eu(III) complex. For all conjugates a small number of cleavage sites were observed (24), and all sites are in close proximity to the tethered complex. Several lines of evidence suggest that the sites arise from hydrolytic cleavage. HPLC analysis of the products of cleavage of small RNA oligomers indicate that the end groups initially formed are cyclic phosphate esters [27]. For the copper(II) terpyridyl complex, evidence for hydrolytic cleavage rests on the fact that a different set of cleavage sites is observed upon addition of reducing agents to drive the redox chemistry pathway [28]. For the lanthanide(III) conjugates, cleavage products of both 3'-end-labeled and 5'-end-labeled RNA targets comigrate with products obtained from alkaline hydrolysis or from partial digestion with base-specific ribonucleases [2527,106]. Full kinetic studies have not been reported for any of the conjugates. However, the relatively slow rates that are observed indicate that cleavage is probably the rate-determining step. If the rate constants for the association of RNA and the conjugates are similar to those of unmodified DNAs [107], then association cannot be the rate-determining step. In addition, the length of the oligodeoxyribonucleotides (>14 bases) in the conjugates in Table 1 ensures that the conjugates are essentially fully associated to the RNA under the concentrations given. The rate of the cleavage step is dependent on the metal ion complex catalyst. That the most sluggish reaction is with the copper(II) conjugate correlates to the lower pseudo-first-order rate constants for cleavage of RNA by the free copper(II) complex [15,18] compared to the lanthanide(III) complexes [19,23]. The different linkers used for each complex may also modify cleavage rates. Most of the conjugates have flexible alkylamine linkers at either the 5' or 3' terminus of the oligonucleotide. For the texaphyrin conjugates, only the conjugate with the
< previous page
page_583
next page >
< previous page
page_584
next page > Page 584
5'-terminal amine linker cleaves RNA whereas a conjugate with an internal attachment to a thymine does not cleave RNA. In contrast, the copper(II) terpyridyl complex is linked through an internal thymine, and there are no data for a conjugate linked through a 5' or 3' terminus for comparison. The point of ligand attachment as well as the total number of atoms in the tether are quite different for the copper(II) complex and for the texaphyrin complex, and this may be one factor in the different cleavage rates that are observed. Many of the considerations that have been discussed for gene inactivation by ribozymes are important if the metal ion complex conjugates are to be used for similar purposes [108]. In order to effectively reduce the concentration of a typical eukaryotic mRNA, the ribozymes or conjugates will need to cleave RNA more rapidly than the half-life of the mRNA, which is typically on the order of 3 hr. In addition, the concentration of conjugate or ribozyme required for gene inactivation will be dependent on the number of copies of the target RNA in the cell. Ribozymes have been shown to be more effective than simple antisense oligonucleotides in selective gene inactivation [108,109], presumably due to their ability to cleave RNA. Ribozyme-catalyzed RNA cleavage is rapid under in vitro conditions. For example, with picomolar concentrations of a 19-nucleotide ribozyme and RNA target, the target RNA has a halflife of 4.4 min at 37°C, 10 mM Mg2+ [58]. Notably, most studies where ribozymes have been used to successfully inactivate genes have had a large excess of ribozyme to RNA target, suggesting that catalytic turnover is not important. Further study is required to fully characterize the kinetics of cleavage by the metal ion complex conjugates; however, data in Table 1 would suggest that the cleavage rates may not be sufficiently rapid for gene inactivation by cleavage of mRNA. Since the conjugates in Table 1 are the fruits of the first efforts to design such catalysts, there is much room for improvement. This would include the preparation of more active metal ion complex catalysts. In addition, linkers that would position the metal complex more precisely with respect to the phosphate diester backbone of RNA could be utilized. The atoms in the ligand part of the linker will also be important in the alignment of the catalyst. More careful consideration of the metal ion complex orientation may yield precision cleaving agents [110] that more rapidly and more selectively cleave RNA.
< previous page
page_584
next page >
< previous page
page_585
next page > Page 585
4 Summary and Outlook Much progress has been made during the past 4 years in the design of metal ion complexes for the hydrolytic cleavage of RNA. The use of model complexes has led to a better understanding of factors that are important in the facilitation of hydrolytic cleavage of RNA by metal ion complexes. However, there are still relatively few reports on the cleavage of longer RNA fragments; such studies would be useful to determine the effect of structure on cleavage rates. In addition, the base sequence specificity of cleavage by metal ion complexes is largely unexplored. There are now several examples of sequence-specific cleavage of RNA by metal ion complex-oligonucleotide conjugates. Future work will entail the use of different linkers to position the metal ion complex more precisely as well as the development of new catalysts. Kinetic data on RNA model compounds suggests that dimeric metal ion complex catalysts hold promise for the hydrolytic cleavage of RNA. There are several examples of lanthanide(III) complex catalysts that are inert to metal ion dissociation in solution. Several of these complexes have been tethered to antisense oligonucleotides. These complex conjugates will be especially useful for the study of gene inactivation by the sequence-specific cleavage of RNA as the metal ion will not dissociate readily under physiological conditions. Note Added in Proof Since this chapter was submitted, several reports of RNA cleavage by metal ions [111], dimeric metal ion complexes [112114], and monomeric metal ion complexes [115,116] have appeared. Abbreviations
Ac
acetate
DOTA
1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate
DTPA
diethylenetriaminepentaacetate
< previous page
page_585
next page >
< previous page
page_586
next page > Page 586
EDTA ethylenediamine-N,N,N',N'-tetraacetate HAM
2,7,13,18-tetramethyl-3,6,14,17,23,24-hexaazatricyclo- [17.3.1.1]tetracosa1(23),2,6,8,10,12(24),13,17,19,21-decane
HPLC high-performance liquid chromatography hpnp
2-hydroxypropyl-4-nitrophenylphosphate
IDA
iminodiacetate
NBAC 1-(4-nitrobenzyl)-4,7,10-tris(carbamoylmethyl)-1,4,7,10- tetraazacyclododecane NMR
nuclear magnetic resonance
NTHE 2-(4-nitrobenzyl)-1,4,7,10-tetrakis(2-hydroxyethyl)-1,4,7,10- tetraazacyclododecane OAM
hexadentate 2,2':6',2''-terpyridine macrocycle
tex
texaphyrin
THED 1,4,7,10-tetrakis(2-hydroxyethyl)-1,4,7,10-tetraaza- cyclododecane s-THP 1S,4S,7S,10S-tetrakis(2-hydroxypropyl)-1,4,7,10-tetraaza- cyclododecane TCEC 1,4,7,10-tetrakis(2-carbamoylethyl)-1,4,7,10-tetraazacyclo- dodecane TCMC 1,4,7,10-tetrakis(carbamoylmethyl)-1,4,7,10-tetraazacyclo- dodecane TCMT 1,4,7-tris(carbamoylmethyl)-1,4,7-triazacyclononane trien
triethylenetetramine
TRPY 2,2':6',2"-terpyridine
References 1. K. Dimroth, H. Witzel, W. Hulsen, and H. Mirbach, Justus Liebigs. Ann. Chem., 620, 94 (1959). 2. E. Baumann, H. Trapman, and F. Fischler, Biochem. Z., 328, 89 (1954). 3. G. Eichhorn and J. J. Butzow, Biopolymers, 3, 79 (1965). 4. W. R. Farkas, Biochim. Biophys. Acta, 155, 401 (1967).
5. C. Werner, B. Krebs, G. Keith, and G. Dirheimer, Biochim. Biophys. Acta, 432, 161 (1976).
< previous page
page_586
next page >
< previous page
page_587
next page > Page 587
6. R. S. Brown, B. E. Hingerty, and J. C. Dewan, Nature, 303, 543 (1983). 7. R. S. Brown, J. C. Dewan, and A. Klug, Biochemistry, 24, 4785 (1985). 8. T. R. Cech, A. J. Zaug, and P. J. Grabowski, Cell, 27, 487 (1981). 9. C. Guerrier-Takada, K. Gardiner, T. Marsh, N. Pace, and S. Altman, Cell, 35, 848 (1983). 10. C. A. Grosshan and T. R. Cech, Biochemistry, 28, 6888 (1989). 11. S. C. Dahm and O. C. Uhlenbeck, Biochemistry, 30, 9464 (1991). 12. A. M. Pyle, Science, 261, 709 (1993). 13. J. A. Piccirilli, J. S. Vyle, M. H. Caruthers, and T. R. Cech, Nature, 361, 85 (1993). 14. S. Sawata, M. Komiyama and K. Taira, J. Am. Chem. Soc., 117, 2357 (1995). 15. M. K. Stern, J. K. Bashkin, and E. D. Sall, J. Am. Chem. Soc., 112, 5357 (1990). 16. Y. Matsumoto and M. Komiyama, J. Chem. Soc., Chem. Commun., 1050 (1990). 17. V. M. Shelton and J. R. Morrow, Inorg. Chem., 31, 4295 (1991). 18. A. S. Modak, J. K. Gard, M. C. Merriman, K. A. Winkeler, J. K. Bashkin, and M. K. Stern, J. Am. Chem. Soc., 113, 283 (1991). 19. J. R. Morrow, L. A. Buttrey, V. M. Shelton, and K. A. Berback, J. Am. Chem. Soc., 114, 1903 (1992). 20. F. Chu, J. Smith, V. M. Lynch, and E. V. Anslyn, Inorg. Chem., 34, 5689 (1995). 21. K. Kolasa, J. R. Morrow, and A. Sharma, Inorg. Chem., 32, 3983 (1993). 22. N. Hayashi, N. Takeda, T. Shiiba, M. Yashiro, K. Watanabe, and M. Komiyama, Inorg. Chem., 32, 5899 (1993). 23. K. O. A. Chin and J. R. Morrow, Inorg. Chem., 33, 5036 (1994). 24. S. Amin, J. R. Morrow, C. H. Lake, and M. R. Churchill, Angew. Chem. Int. Ed. Engl., 33, 773 (1994). 25. D. Magda, R. A. Miller, J. L. Sessler, and B. L. Iverson, J. Am. Chem. Soc., 116, 7439 (1994). 26. J. Hall, D. Husken, U. Pieles, H. E. Moser, and R. Haner, Chem. Biol., 1, 185 (1994).
< previous page
page_587
next page >
< previous page
page_588
next page > Page 588
27. K. Matsumura, M. Endo, and M. Komiyama, J. Chem. Soc., Chem. Commun., 2019 (1994). 28. J. K. Bashkin, E. I. Frovlova, and U. Sampath, J. Am. Chem. Soc., 116, 5981 (1994). 29. C. A. Stein and J. S. Cohen, Cancer Res., 48, 2659 (1988). 30. J. R. Morrow in Advances in Inorganic Biochemistry (G. L. Eichhorn and L. G. Marzilli, eds.), Prentice-Hall, Englewood Cliffs, 1994, p. 42. 31. M. R. Eftink and R. L. Biltonen in Hydrolytic Enzymes (A. Neuberger and K. Brocklehurst, eds.), Elsevier, New York, 1987, p. 333. 32. B. C. F. Chu and L. E. Orgel, Proc. Natl. Acad. Sci. USA, 82, 963 (1985). 33. S.-B. Lin, K. R. Blake, P. S. Miller, and P. O. P. Ts'o, Biochemistry, 28, 1054 (1989). 34. R. Breslow and D.-L. Huang, Proc. Natl. Acad. Sci. USA, 88, 4080 (1991). 35. J. K. Bashkin and L. A. Jenkins, J. Chem. Soc., Dalton Trans., 3632 (1993). 36. J. MacB. Harrowfield, D. R. Jones, L. F. Lindoy, and A. M. Sargeson, J. Am. Chem. Soc., 102, 7733 (1980). 37. B. K. Takasaki and J. Chin, J. Am. Chem. Soc., 116, 1121 (1994). 38. B. K. Takasaki and J. Chin, J. Am. Chem. Soc., 115, 9337 (1993). 39. R. Breslow and B. Zhang, J. Am. Chem. Soc., 116, 7893 (1994). 40. J. R. Morrow, K. O. A. Chin, and K. A. Aures in Genetic Response to Metals (B. Sarkar, ed.), Marcel Dekker, New York, 1995, p. 173. 41. J. Chin, Acc. Chem. Res., 24, 145 (1991). 42. P. Hendry and A. M. Sargeson in Progress in Inorganic Chemistry: Bioinorganic Chemistry, Vol. 38 (S. J. Lippard, ed.), John Wiley and Sons, New York, 1987, p. 201. 43. G. L. Eichhorn, J. J. Butzow, P. Clark, and E. Tarien, Biopolymers, 5, 283 (1967). 44. J. J. Butzow and G. L. Eichhorn, Biopolymers, 3, 95 (1965). 45. G. L. Eichhorn, E. Tarien, and J. J. Butzow, Biochemistry, 10, 2014 (1971). 46. J. J. Butzow and G. L. Eichhorn, Biochemistry, 10, 2019 (1971).
< previous page
page_588
next page >
< previous page
page_589
next page > Page 589
47. R. Breslow, D. L. Huang, and E. Anslyn, Proc. Natl. Acad. Sci. USA, 86, 1746 (1989). 48. J. R. Morrow and V. M. Shelton, New J. Chem., 18, 371 (1994). 49. M. Komiyama, K. Matsumura, and Y. Matsumoto, J. Chem. Soc. Chem. Commun., 640 (1992). 50. T. R. Cech, Science, 236, 1532 (1987). 51. R. B. Van Atta and S. M. Hecht, Advances in Inorganic Biochemistry (G. L. Eichhorn and L. G. Marzilli, eds.), Prentice-Hall, Engelwood Cliffs, 1994, p. 1. 52. G. A. Prody, J. T. Bakos, J. T. Buzayan, I. R. Schneider, and G. Bruening, Science, 231, 1577 (1986). 53. V. Dange, R. B. Van Atta, and S. M. Hecht, Science, 248, 585 (1990). 54. T. Pan and O. C. Uhlenbeck, Biochemistry, 31, 3889 (1992). 55. H.-Y. Deng and J. Termini, Biochemistry, 31, 10518 (1992). 56. T. Pan, B. Dichtl and O. C. Uhlenbeck, Biochemistry, 33, 9561 (1994). 57. H. W. Pley, K. M. Flaherty, and D. B. McKay, Nature, 372, 68 (1994). 58. O. C. Uhlenbeck, Nature, 328, 596 (1987). 59. D. M. Brown and D. A. Usher, J. Chem. Soc., 6558 (1965). 60. J. R. Morrow, L. A. Buttrey, and K. A. Berback, Inorg. Chem., 31, 16 (1992). 61. D. A. Usher, D. I. Richardson, and D. G. Oakenfull, J. Am. Chem. Soc., 92, 4699 (1970). 62. K. N. Dalby, A. J. Kirby, and F. Hollfelder, J. Chem. Soc., Perkin Trans., 2, 1269 (1993). 63. A. J. Chandler and A. J. Kirby, J. Chem. Soc., Chem. Commun., 1769 (1992). 64. S. Kuusela and H. Lönnberg, J. Chem. Soc. Perkin Trans., 2, 2109 (1994). 65. S. Amin, D. A. Voss, Jr., W. DeW. Horrocks, Jr., C. H. Lake, M. R. Churchill, and J. R. Morrow, Inorg. Chem., 34, 3294 (1995). 66. R. O. Dempcy and T. C. Bruice, J. Am. Chem. Soc., 116, 4511 (1994).
< previous page
page_589
next page >
< previous page
page_590
next page > Page 590
67. J. K. Bashkin and L. A. Jenkins, Comm. Inorg. Chem., 16, 77 (1994). 68. V. Jubian, R. P. Dixon, and A. D. Hamilton, J. Am. Chem. Soc., 114, 1120 (1992). 69. J. Smith, K. Ariga, and E. V. Anslyn, J. Am. Chem. Soc., 115, 362 (1993). 70. B. Barbier and A. Brack, J. Am. Chem. Soc., 114, 3511 (1992). 71. T. H. Fife and M. P. Pujari, J. Am. Chem. Soc., 110, 7790 (1988). 72. K. A. Browne and T. C. Bruice, J. Am. Chem. Soc., 114, 4951 (1992). 73. T. A. Steitz and J. A. Steitz, Proc. Natl. Acad. Sci. USA, 90, 6498 (1993). 74. M. Wall, R. C. Hynes, and J. Chin, Angew. Chem. Int. Ed. Engl., 32, 1633 (1993). 75. D. Wahnon, R. C. Hynes, and J. Chin, J. Chem. Soc., Chem. Commun., 1441 (1994). 76. D. R. Jones, L. F. Lindoy, and A. M. Sargeson, J. Am. Chem. Soc., 106, 7807 (1984). 77. D. H. Vance and A. W. Czarnick, J. Am. Chem. Soc., 115, 12165 (1993). 78. J. Chin and M. Banaszczyk, J. Am. Chem. Soc., , 4103 (1989). 79. M. W. Gobel, Angew. Chem. Int. Ed. Engl., 33, 1141 (1994). 80. N. N. Murphy, M. Mahroof, and K. D. Karlin, J. Am. Chem. Soc., 115, 10404 (1993). 81. N. Takeda, M. Irisawa, and M. Komiyama, J. Chem. Soc. Chem. Commun., 2773 (1994). 82. A. Tsubouchi and T. C. Bruice, J. Am. Chem. Soc., 116, 11614 (1994). 83. B. F. Rordorf and D. R. Kearns, Biopolymers, 15, 1491 (1976). 84. M. F. Tweedle, J. J. Hagan, K. Kumar, S. Mantha, and C. A. Chang, Mag. Res. Imaging, 9, 409 (1991). 85. I. Bertini, C. Luchinat, M. Rosi, A. Sgamellotti, and F. Tarantelli, Inorg. Chem., 29, 1460 (1990). 86. E. Anslyn and R. Breslow, J. Am. Chem. Soc., 111, 4473 (1989). 87. R. Breslow, D. Berger, and D.-H. Huang, J. Am. Chem. Soc., 112, 3686 (1990).
< previous page
page_590
next page >
< previous page
page_591
next page > Page 591
88. T. P. Prakash and K. N. Ganesh, J. Chem. Soc. Chem. Commun. 1357 (1994). 89. J. R. Morrow, S. Amin, C. H. Lake, and M. R. Churchill, Inorg. Chem., 32, 4566 (1993). 90. S. Amin, C. Marks, L. Toomey, M. R. Churchill, and J. R. Morrow, Inorg. Chim. Acta, in press. 91. S. Amin, Ph.D. thesis, State University of New York at Buffalo, 1995. 92. J. R. Morrow and K. O. A. Chin, Inorg. Chem., 32, 3357 (1993). 93. K. O. A. Chin, J. R. Morrow, C. H. Lake, and M. R. Churchill, Inorg. Chem., 33, 656 (1994). 94. X. Wang, T. Jin, V. Comblin, A. Lopez-Mut, E. Merciny, and J. F. Desreux, Inorg. Chem., 31, 1095 (1992). 95. C. F. Geraldes, A. M. Urbano, M. A. Hoefnagel, and J. A. Peters, Inorg. Chem., 32, 2426 (1993). 96. M. S. Konings, W. C. Dow, D. B. Love, K. N. Raymond, S. C. Quay, and S. M. Rocklage, Inorg. Chem., 29, 1488 (1990). 97. A. D. Sherry, R. D. Brown, III, C. F. G. C. Geraldes, S. H. Koenig, K.-T. Kuan, and M. Spiller, Inorg. Chem., 28, 620 (1989). 98. S. T. Frey, C. A. Chang, J. F. Carvalho, A. Varadarajan, L. M. Schultze, K. L. Pounds, and W. DeW. Horrocks, Jr., Inorg. Chem., 33, 2882 (1994). 99. M. Kodama, T. Koike, A. B. Mahatma, and E. Kimura, Inorg. Chem., 30, 1270 (1991). 100. J. Burgess in Metal Ions in Solution: Basic Principles of Chemical Interaction, Ellis Horwood Ltd., West Sussex, 1988. 101. H. Ikenaga and Y. Inoue, Biochemistry, 13, 577 (1974). 102. S. Kazakov and S. Altman, Proc. Natl. Acad. Sci. USA, 89, 7939 (1992). 103. D. A. Usher, Nature New Biology, 235, 207 (1972). 104. R. Kierzek, Nucleic Acids Res., 20, 5073 (1992). 105. E. Uhlmann and A. Peyman, Chem. Rev., 90, 543 (1990). 106. L. Huang and J. R. Morrow, in preparation. 107. S. M. Freier in Antisense Research and Applications (S. T. Crooke and B. Lebleu, eds.), CRC Press, Ann Arbor, 1993, p. 67.
< previous page
page_591
next page >
< previous page
page_592
next page > Page 592
108. O. C. Uhlenbeck in Antisense Research and Applications (S. T. Crooke and B. Lebleu, eds.), CRC Press, Ann Arbor, 1993, p. 83. 109. N. K. Tanner and M. Vasseur in Antisense Research and Applications (S. T. Crooke and B. Lebleu, eds.), CRC Press, Ann Arbor, 1993, p. 415. 110. D. E. Bergstrom and N. P. Gerry, J. Am. Chem. Soc., 116, 12067 (1994). 111. M. Irisawa, N. Takeda, and M. Komiyama, J. Chem. Soc. Chem. Commun., 1221 (1995). 112. W. H. Chapman, Jr. and R. Breslow, J. Am. Chem. Soc., 117, 5462 (1995). 113. M. J. Young and J. Chin, J. Am. Chem. Soc., 117, 10577 (1995). 114. M. Yashiro, A. Ishikubo, and M. Komiyama, J. Chem. Soc. Chem. Commun., 1793 (1995). 115. J. Hovinen, A. Guzaev, E. Azhayeva, A. Azhayev, and H. Lonnberg, J. Org. Chem., 60, 2205 (1995). 116. B. Linkletter and J. Chin, Angew. Chem. Int. Ed. Engl., 34, 472 (1995).
< previous page
page_592
next page >
< previous page
page_593
next page > Page 593
20 RNA Recognition and Cleavage by Iron(II)-Bleomycin Jean-Marc Battigello, Mei Cui, and Barbara J. Carter Department of Chemistry, University of Toledo, Toledo, OH 43606-3390, USA
594
1. Introduction
594
1.1. Chemotherapeutic Role of Bleomycin
595
1.2. Activation
1.3. Recognition and Cleavage of DNA Substrates
598
2. Cleavage of RNA by Fe(II)-BLM
598
2.1. Historical Perspective
599
2.2. tRNA As Substrate
2.2.1. tRNA Precursor Transcripts (Unmodified) As Substrates
2.2.2. Mature (Modified) tRNA Substrates
599
599
599
2.2.3. Cleavage of a ''tDNA" Sequence
2.2.4. Other (Non-tRNA) RNA Substrates
601
604
2.3. Recognition: Sequence or Structure?
2.3.1. Sequence Preferences Compared to DNA
2.3.2. Single-Stranded vs. Double-Stranded Cleavage
< previous page
596
page_593
604
604
next page >
< previous page
page_594
next page > Page 594
2.3.3. Selectivity: Substrate Preference for DNA or RNA?
605
605
2.3.4. Binding Affinity
606
2.4. Chemistry of Cleavage
606
2.4.1. Cleavage Products
2.4.2. C4'H Abstraction vs. C1'H Abstraction
608
609
3. Therapeutic Relevance
610
4. Conclusions and Outlook
611
Acknowledgment
611
Abbreviations
612
References
1 Introduction 1.1 Chemotherapeutic Role of Bleomycin The bleomycins (BLMs) are a group of glycoprotein antibiotics which are used in the treatment of malignant lymphomas and squamous cell carcinomas [14]. The anticancer potency of BLM has been attributed to its ability to mediate strand scission of DNA [511]. Cells treated with BLM accumulate in the late G2 phase of the cell cycle when treated with BLM under conditions known to induce DNA damage [813]. Inhibition of mitosis has also been observed [14], and recovery from this inhibition is very slow. In spite of all that is known, there is no direct evidence that DNA cleavage is the cause of cell growth inhibition and cell death by BLM. In fact, Berry et al. showed that repair of BLMinduced strand scission occurs at a faster rate than DNA degradation, after an initial 1-hr exposure to the drug [15]. Other factors suggest that DNA cleavage may not be BLM's actual therapeutic locus of action. In particular, there is significant evidence that BLM damages cellular membranes via lipid peroxidation [1620]. Also, studies with cultured human cells treated with various BLM congeners showed that in the presence or absence of dibucaine, a local anesthetic found to enhance the ability of BLM to inhibit cell growth by increasing the fluidity of membranes [21,22], there was no linear cor-
< previous page
page_594
next page >
< previous page
page_595
next page > Page 595
relation between ability to mediate DNA strand scission and ability to inhibit cell growth [23]. Finally, chromosomal DNA is extensively encased within proteins [24] throughout most of the cell cycle. Although DNA undergoing mitosis would be more susceptible to BLM cleavage, this fact alone cannot account for BLM's antitumor selectivity. It also fails to explain why BLM arrests cells in the late G2 phase of the cell cycle, since this phase occurs just before mitosis [24,25]. 1.2 Activation Bleomycin degrades DNA in a reaction that is O2- and metal ion-dependent [2630]. The oxygen activation [3133] and subsequent chemistry of Fe(II)-BLM-mediated cleavage of DNA has been extensively studied and characterized [7,3440]. Scheme 1 outlines the se-
Scheme 1. Activation of Fe-BLM. quence of events proposed to occur during activation of BLM. Initially, Fe(II) reacts with BLM to form the pink Fe(II)-BLM complex [3133]. This electron paramagnetic resonance (EPR)-silent, high-spin Fe(II)-BLM complex then rapidly and reversibly reacts with O2 to form the complex O2-Fe(II)-BLM, which is a short-lived, EPR-silent species (t1/2 = 6 sec in H2O, 2°C) [31]. This transient species, if not destroyed by
< previous page
page_595
next page >
< previous page
page_596
next page > Page 596
deoxygenation by its DNA substrate [41], proceeds to become what is termed ''activated BLM" via a single-electron reduction. Activated BLM is the species responsible for DNA cleavage [31] and differs from the O2-Fe(II)-BLM complex in that it displays an EPR spectrumbut one which is not consistent with that of Fe(III)-BLM. Recent reports using X-ray absorption spectroscopy [42] and electrospray mass spectrometry [43] indicate that activated BLM is a low-spin ferric species, consistent with previous EPR findings, and accurately described by the peroxy-iron(III)-BLM label. FeBLM can also reductively activate O2 and thus facilitate small molecule chemistry with molecules such as cisstilbene [44], N,N-dimethylaniline, or p-deuteroanisole [45]. 1.3 Recognition and Cleavage of DNA Substrates Bleomycin contains two major domains: (1) the metal binding region and (2) the C-terminus + bithiazole region designated the DNA binding domain (Fig. 1A). BLM cleavage of DNA involves both single- and double-strand breaks and is sequence-selective; a strong preference is shown for single-strand cleavage at 5'-GC-3' and 5'-GT-3' sites [46,47]. Several groups have investigated BLM's preferred sites for double-strand cleavage [4850]. These reports indicate that there are five sequence patterns noted for double-strand cleavage by BLM, with the two most predominant sites being shown below:
The DNA substrate itself plays a large role in where and how BLM interacts with DNA. Platination of DNA causes major alterations in the sequence specificity of BLM-mediated cleavage [51,52], and other experiments showed that Fe(II)-BLM consistently cleaved single-nucleotide bulged DNA duplexes at sites near the bulge [53]. Also, methylation of DNA at selected cytosines and adenosines resulted in diminished cleavage at or adjacent to the methylated sites [54]. Significant evidence suggests that BLM binds in the minor groove of DNA [5456]. Minor groove-binding molecules such as distamycin A
< previous page
page_596
next page >
< previous page
page_597
next page > Page 597
Fig. 1. Structure of BLM A2 (A) and gly-BLM analogs (B). Reproduced from Ref. 66 by permission from the American Society for Biochemistry and Molecular Biology, Inc.
< previous page
page_597
next page >
< previous page
page_598
next page > Page 598
and actinomycin D altered the sequence selectivity of BLM cleavage [55], and other studies showed that agents which bind in the major groove of DNA do not alter BLM cleavage specificity [54]. However, BLM may also be a (partial) intercalating agent, based on nuclear magnetic resonance (NMR) and DNA unwinding studies [5763]. The planar, aromatic bithiazole region is thought to be capable of intercalation and has also been associated with the GC and GT specificity of BLM cleavage [55,64]. But other reports indicate that the metal binding region plays a large role in determining sequence specificity [62,65,66]. Carter et al. [66] showed that four variable-length BLM analogs having either 0, 1, 2, or 4 glycine units replacing the normal threonine unit of BLM (see Fig. 1B) exhibited identical sequence-specific cleavage of two DNA substrates. Lastly, the compound 2[(N-(aminoethyl)aminomethyl]-4-(2,4-imidazoyl)ethyl)carbamoyl]-5-bromopyrimidine (PMAH), which models the metal-binding region of BLM, cleaved DNA with the same sequence selectivity as intact BLM [67]. As a final note, scientists have debated for years as to whether a single BLM molecule is responsible for doublestranded cleavage of DNA or whether two BLM molecules are required [6870]. Steighner and Povirk proposed in 1990 [49] that a single molecule of BLM is responsible for the double-strand cleavage observed. Recent quantitative studies by Absalon et al. [71,72] using DNA hairpins as substrates confirmed this original hypothesis. 2 Cleavage of RNA by Fe(II)-BLM 2.1 Historical Perspective For years it was reported that RNA was not a substrate for BLM cleavage. However, early studies did not deal specifically with RNA as the substrate. Some studies investigated BLM cleavage of RNA in the presence of DNA [7375]. Other experiments investigated cleavage of RNA in the absence of added metal ions [76], or by admixing Cu(II)-BLM + Fe(II)-BLM to single- and double-stranded RNA homopolymers [77]. The first study showing RNA to be a relevant target for BLM was that of Magliozzo et al. [78]. They reported minimal degradation of
< previous page
page_598
next page >
< previous page
page_599
next page > Page 599
yeast tRNAPhe by high concentrations (300 µM) of Fe(II)-BLM. TLC analysis indicated the presence of products that comigrated with adenine and uracil. Since then every major class of RNA has proved to be a substrate for degradation by Fe(II)-BLM. These studies are discussed below. 2.2 tRNA As Substrate 2.2.1 tRNA Precursor Transcripts (Unmodified) As Substrates Using a tRNAHis precursor from Bacillus subtilis, Carter et al. [79] reported BLM cleavage at 3 µM Fe(II)-BLM concentration, comparable to that observed for DNA substrates under similar conditions. The major BLM cleavage site on the tRNAHis substrate was identified using RNA enzymatic sequencing reactions [80] and was determined to be U35 in a 5'-GU-3' sequence (Fig. 2A). Other lesser cleavage sites were also observed. Since then, additional tRNA transcripts have been cleaved by Fe(II)-BLM, including an E. coli tRNASeCys precursor and a Schizosaccharomyces pombe amber suppressor tRNASer [81] (Fig. 2B), a yeast cytoplasmic tRNAAsp precursor [82], and an E. coli tRNAAsp precursor [83]. 2.2.2 Mature (Modified) tRNA Substrates In addition to the report by Magliozzo with yeast tRNAPhe, two other groups have reported BLM cleavage of mature tRNA substrates. Holmes et al. [81] showed cleavage of mature E. coli , and Hüttenhofer et al. [83] confirmed Magliozzo's previous report with mature yeast tRNAPhe (Fig. 2D), and also reported cleavage of a mature molecule. Cleavage of the mature tRNAs was similar to that reported for tRNA transcripts, i.e., 5'E. coli GN-3' sites were frequently observed. But other sites, including a 5'-UU-3' sequence and a 5'-UC-3' sequence, were also reported [81] (Fig. 2C). 2.2.3 Cleavage of a ''tDNA" Sequence Holmes and Hecht [84] reported cleavage of a corollary "transfer DNA" sequence corresponding to the sequence for the B. subtilis tRNAHis
< previous page
page_599
next page >
< previous page
page_600
next page > Page 600
Fig. 2. Secondary structures of four RNA substrates for Fe(II)-BLM and their reported cleavage sites. (A) B. subtilis tRNAHis precursor (B) S. pombe amber suppressor tRNASer precursor (C) E. coli mature tRNAHis (D) yeast tRNAPhe. (Reproduced from Ref. 81 by permission of the American Chemical Society.)
< previous page
page_600
next page >
< previous page
page_601
next page > Page 601
precursor. The cleavage occurred at the same major site as in the B. subtilis tRNAHis precursor; specifically, the 5'GU-3' sequence located at the junction between the acceptor stem and D-stem/loop region of the tRNAHis precursor (Fig. 2A) was also cleaved by Fe-BLM in the tDNAHis substrate. In addition, all minor cleavage sites in the tRNA substrate were cleavage sites in the tDNA substrate except one, but the tDNA substrate exhibited numerous additional Fe-BLM cleavage sites that were not observed in the tRNA substrate. 2.2.4 Other (Non-tRNA) RNA Substrates Other RNA substrates were investigated for cleavage by Fe(II)-BLM. One of these was a random E. coli 231-base RNA transcribed from the pSP64 vector, after Pvu II linearization of the plasmid. Cleavage occurred at two major sites in this substrate [79]. Another RNA cleaved by Fe(II)-BLM was a 347-base 5' fragment of HIV-1 reverse transcriptase (HIV-RT) mRNA. Although the exact cleavage sites were not characterized, at least four cleavage sites were observed [79,82]. A messenger RNA substrate, coding for the iron-binding protein ferritin, was treated with Fe(II)-BLM by Dix et al. [85]. Structural changes in the iron regulatory element (IRE) were monitored by changes in sensitivity to BLM cleavage at particular sites. The IRE and flanking region (FL) of the mRNA for ferritin was subjected to treatment with Fe-BLM, and a single BLM cleavage site was observed at U17 in the wild-type transcript. The site was identified as a 5'-GU-3' sequence at the junction of a single-strand/double-strand region (Fig. 3). Various mutations in the wild-type RNA resulted in altered BLM sensitivity. Specifically, a mutation which disrupts a conserved triplet base pair region in the IRE, and thus results in decreased translational regulation by the regulator protein which binds this region, resulted in BLM cleavage at two new sites, A10 and All (Fig. 3). Also, a deletion mutant which removed 12 bases 3' to the IRE resulted in the appearance of a new cleavage site, this time at U-1 in the flanking region. Ribosomal RNA has also been used as a substrate for BLM cleavage. Cleavage of yeast 5S rRNA by Holmes et al. [81] indicated three major BLM cleavage sites in the molecule. All three sites occurred at a U in a 5'-GUA-3' sequence, and all sites were within one or two bases of a one-nucleotide bulged region 3' to the cleavage site. In addition,
< previous page
page_601
next page >
< previous page
page_602
next page > Page 602
Fig. 3. Structure and cleavage sites in the IRE and FL regions of the mRNA for ferritin. *, Enhanced BLM cleavage sites in mutant ferritin mRNAs; , BLM cleavage of wild-type ferritin mRNA. (Redrawn from Ref. 85 by permission of the Academic Press.) these sites are postulated to lie in helical regions of the RNA and to act as stabilizers of tertiary structure [86]. Battigello et al. recently investigated an RNA pseudoknot as a substrate for Fe-BLM cleavage [87]. A proposed mRNA pseudoknot in the coding region for HIV-1 reverse transcriptase [88] was treated with Fe(II)-BLM and a single BLM cleavage site was observed (Fig. 4). Using enzymatic sequencing reactions, the cleavage site was mapped to U18, which is in a 5'-AU-3' sequence, near the junction of two coaxially stacked helices. The cleavage was Mg2+-dependent, with cleavage at U18 eliminated in the absence of Mg2+. In the absence of magnesium ions, the proposed pseudoknot structure would not be expected to be
< previous page
page_602
next page >
< previous page
page_603
next page > Page 603
Fig. 4. Fe(II)-BLM cleavage and structure of the mRNA HIV-1 , proposed pseudoknot. *, Fe(II)-BLM cleavage; neocarzinostatin cleavage. The BLM reaction contained ~10,000 cpm of 5'-end-labeled RNA, 5 mM Mg2+, 5 mM NaH2PO2, pH 7.1, and 40 µM Fe(II)-BLM. (Reproduced from Ref. 87 by permission of Elsevier Science Publishing Co, Inc.) stable; thus BLM cleavage of this RNA substrate is dependent on the RNA tertiary structure itself. Another RNA substrate recently investigated in our laboratory, a 22-mer RNA hairpin [87], was also cleaved by Fe-BLM at several sites, one of which occurs within five bases of the 5' end (unpublished results). And finally, an RNA/DNA heteroduplex was investigated for Fe-BLM cleavage under aerobic conditions. Morgan and Hecht determined
< previous page
page_603
next page >
< previous page
page_604
next page > Page 604
that both the RNA and DNA strands were cleaved at comparable concentrations at a limited number of sites [89]. Interestingly, the cleavage sites determined for the RNA strand of the heteroduplex differed from those observed in the RNA alone, from which the heteroduplex was formed via reverse transcription. This substrate is of particular interest because RNA-DNA heteroduplexes are formed during normal transcription and also during reverse transcription. 2.3 Recognition: Sequence or Structure? 2.3.1 Sequence Preferences Compared to DNA A survey of the RNA substrates cleaved by Fe-BLM indicates that most cleavage occurred at 5'-G-pyr-3' sequences, particularly 5'-GU-3' sequences [82]. However, several Fe-BLM cleavage sites have been reported for RNA substrates that are rarely if ever observed in DNA substrates. Specifically, a 5'-CA-3' sequence, a 5'-UG-3' sequence, a 5'-GG- sequence, and a 5'-GA-3' sequence were all cleaved in the B. subtilis precursor tRNAHis substrate [79] (Fig. 2A); the RNA strand in the RNA/DNA heteroduplex substrate was cleaved by Fe-BLM at a 5'-GG-3', a 5'-CC3', and a 5'-UA-3' [89]; and a 5'-UU-3' site was reported in E. coli mature are regular recognition sequences for Fe-BLM cleavage of DNA substrates.
(Fig. 2C) [81], none of which
Also, all possible 5'-GN-3' sequences have been cleaved in RNA substrates, but curiously, only a small fraction of the possible 5'-Gpyr-3' sites in a given RNA substrate are cleaved. Some RNAs are not cleaved by BLM at all, even under forcing conditions [79,82]. Based on BLM cleavage of RNA to date, sequence in and of itself does not appear to be the primary determinant for where cleavage occurs. 2.3.2 Single-Stranded vs. Double-Stranded Cleavage The 1990 report by Carter et al. [79] identifying the major site of cleavage in the B. subtilis tRNAHis precursor as a single-strand region of RNA was unprecedented. Other lesser cleavage sites were also mapped to single-strand regions of the tRNAHis precursor (see Fig. 2A above). That Fe-BLM cleavage of the tRNAHis precursor occurred primarily at single-strand/double-strand junctions, or in wholly single-strand regions of the RNA molecule, proved to be a general principle rather than
< previous page
page_604
next page >
< previous page
page_605
next page > Page 605
an isolated account. With the exception of the RNA/DNA heteroduplex, every RNA molecule reported to be a substrate for Fe(II)-BLM cleavage shows most cleavage occurring at single-strand/double-strand junctions [79,8185,87]. Although enhanced BLM cleavage of sites in DNA substrates has been reported adjacent to bulged bases [53], no BLM cleavage sites in verified single-strand regions of DNA have been reported. 2.3.3 Selectivity: Substrate Preference for DNA or RNA? Fe(II)-BLM cleavage of RNA substrates was found to be highly selectivein fact, much more selective for RNA than for DNA. For example, a tRNATyr precursor was not a substrate for Fe(II)-BLM A2 under any conditions tried [79,82]. Furthermore, numerous other mature and transcript tRNAs have been investigated for Fe(II)-BLM cleavage, and several were not cleaved [79,81,82]. Besides greater selectivity, BLM cleavage of RNA occurs at a much smaller number of sites than cleavage of corresponding DNA substrates with comparable sequences and lengths [79,81,82,84]. On average, Fe(II)-BLM cleavage of RNA is ~10-fold more selective than Fe(II)-BLM cleavage of DNA [79,82]. The effects of carrier DNA and tRNA on Fe(II)-BLM RNA cleavage were studied using precursor tRNAHis from B. subtilis as the substrate [84]. Addition of unlableled tRNAHis to either the tDNAHis reaction or the tRNAHis reaction significantly diminished BLM cleavage, and for both substrates cleavage diminished more with addition of unlabeled tRNA than with addition of unlabeled tDNA. Also, at comparable concentrations of tDNAHis and tRNAHis, cleavage of the DNA substrate was readily observed at 500 nM Fe(II)-BLM, while a similar degree of cleavage was not observed for the RNA substrate until 2.5 µM Fe(II)-BLM was added. The authors conclude that the difference in BLM cleavage efficiency results because Fe(II)-BLM actually binds RNA better than it does DNA, at least in this case. However, the binding does not translate into more efficient cleavage or a greater number of cleavage sites. 2.3.4 Binding Affinity At several experimental conditions, the binding affinity of Fe(II)-BLM for DNA has been determined to be ~105 M1 [90]. To date, no quantitative binding measurement has been determined for Fe(II)-BLM and
< previous page
page_605
next page >
< previous page
page_606
next page > Page 606
RNA. As such, only general conclusions can be drawn. Holmes et al. [81] and Hecht [82] (vide supra) argue that Fe(II)-BLM actually binds more tightly to RNA than to DNA. Assuming this is true, why is the frequency of BLM cleavage so much less for RNA than for DNA substrates? In a provocative discussion recently put forth [82], Hecht contends that the observed BLM cleavage for RNA substrates actually represents most of the BLM lesions occurring in the RNA molecule [91]. Thus, the lower frequency of cleavage does not result from additional ''unseen" lesions not amenable to detection. Rather, the data imply that the C4' hydrogens in DNA and RNA are not equally accessible to activated BLM, possibly due to the very different nature of the minor groove in RNA; thus, fewer binding sites in RNA actually lead to cleavage [82]. 2.4 Chemistry of Cleavage 2.4.1 Cleavage Products Oxidative cleavage of DNA substrates by Fe(II)-BLM results in production of base-propenals, an oligonucleotide with a 5'-phosphate, and an oligonucleotide with a 3'-phosphoroglycolate [57,90,92]. Scheme 2 shows the products arising from cleavage at C-3 of an octanucleotide DNA duplex by Fe-BLM. BLM first abstracts a hydrogen atom from the C-4' position of the deoxyribose ring at cytidine-3, and a transient C-4' radical rapidly forms a C-4' hydroperoxy radical upon O2 capture. Then a Criegee-type rearrangement [93] results in scission of the C-3'-C-4' deoxyribose bond. Free base- and alkali-labile lesions are also generated during BLM-mediated degradation of DNA, but these lesions do not lead to strand scission unless the DNA duplex is subsequently treated with base [94] and so have not been shown. Unfortunately, all the cleavage products resulting from Fe-BLM-treated RNA substrates have not been determined. The only information available regarding the chemistry of Fe(II)-BLM cleavage of RNA comes from (1) evidence comparing electrophoretic mobilities of Fe(II)-BLM cleavage products to products arising from enzymatic digestion of the same substrate [84]; (2) thin-layer chromatography (TLC) and high-performance liquid chromatography (HPLC) analysis of Fe-BLM cleavage of tRNAPhe indicating release of free bases [78,79,91]; and (3) information obtained using chimeric deoxyoligonucleotide duplexes as substrates for Fe(II)-BLM cleavage [79,80,91] containing either a ribo-
< previous page
page_606
next page >
< previous page
page_607
next page > Page 607
Scheme 2. Pathway for Fe(II)-BLM-mediated cleavage of a DNA chimeric octanucleotide sequence 5'-CGCTAGCG-3', resulting from oxidative cleavage at C3. (Reproduced from Ref. 81 by permission of the American Chemical Society.) cytidine or an aracytidine at position 3 (the predominant site of cleavage for this chimeric sequence when it contains all deoxynucleotides). The use of direct product analysis for elucidating the chemistry of Fe(II)-BLM cleavage of RNA has not been done because to date no RNA substrate was found to undergo BLM cleavage near the ends of the
< previous page
page_607
next page >
< previous page
page_608
next page > Page 608
molecule [83]. Recently, however, Fe(II)-BLM was shown to cleave the proposed HIV-1 RT mRNA pseudoknot RNA (Fig. 4) and also a 22-mer RNA hairpin (vide supra) near their 3' and 5' ends, respectively. Thus they represent substrates which could prove especially useful in working out the chemical products generated upon Fe(II)-BLM cleavage of RNA. 2.4.2 C4'H Abstraction vs. C1'H Abstraction For the above-mentioned chimeric octanucleotide DNA duplex containing the C3-ribo or C3-ara nucleotide (see Fig. 5), evidence for C1' H abstraction and subsequent cleavage at the C3 site was reported by Duff et al. [91]. Because the amount of free cytosine released was greater than would have been expected for C4' H abstraction alone,
Fig. 5. Chimeric octanucleotide duplex substrates for Fe-BLM cleavage containing ara- or ribocytidines at position C3. (left) C3-ribo CGCTAGCG, (right) C3-ara CGCTAGCG.
< previous page
page_608
next page >
< previous page
page_609
next page > Page 609
they postulated that C1' H abstraction followed by loss of cytosine at that position was responsible. Trapping of a quinoxaline intermediate further supported this conclusion [82,91]. 3 Therapeutic Relevance There is now considerable experimental evidence to suggest that RNA could constitute an important therapeutic target for BLM. In particular, (1) RNA resides in the cytoplasm of cells, so that there is one less membrane to be crossed by the drug; (2) RNA has no known repair mechanisms, so that BLM cleavage of RNA is expected to cause irreversible damage; and (3) selective destruction of an RNA molecule that encodes a protein critical for the onset of cell division (and which has a turnover rate much higher than its translation from the messenger RNA) could help explain why BLM arrests cells in the late G2 phase of the cell cycle rather than in the S phase [813]. Proteins such as the cyclins [95,96] meet the criteria for rapid turnover and importance in cell division, and as such represent good molecules to target for defining therapeutic relevance of BLM cleavage of RNA molecules. Cleavage of the HIV-1 RT mRNA fragment [80,83] and the HIV-1 RT proposed mRNA pseudoknot [88] are also consistent with BLM cleavage of RNA having therapeutic relevance. While tRNA molecules may not seem an obvious or even relevant target for anticancer drugs, a recent study by Séror et al. [97] reported that a mutant cysteinyl-tRNA synthetase affects the timing of chromosomal replication in B. subtilis. The authors proposed a general hypothesis that the control of chromosomal initiation may be regulated by different elements of the translational machinery, including aminoacyl transfer RNA synthetases, or tRNAs. Furthermore, they stated that these elements ''may regulate the cell cycle by controlling the level of a subset of proteins involved in sensing cell mass and triggering specific cycle events." It is tempting to speculate that BLM's ability to selectively cleave some tRNAs and not others could produce a result similar to that observed by Séror et al. [97], thus lending greater credence to the idea of RNA as a therapeutic target for chemotherapy drugs such as BLM.
< previous page
page_609
next page >
< previous page
page_610
next page > Page 610
However, Hüttenhofer et al. [83] found that cleavage of yeast tRNAphe is essentially abolished in the presence of and cleavage of an E. coli tRNAAsp physiological levels of Mg2+, as is BLM cleavage of E. coli precursor, thus indicating that the relevance of BLM's cleavage of RNA may be limited at physiological concentrations of Mg2+. This is true for the tRNA molecules they investigated. But not all BLM cleavage of RNA is so adversely affected by the addition of Mg2+. Specifically, Fe-BLM cleavage of yeast 5S rRNA and B. subtilis precursor tRNAHis was not abolished in the presence of 5 mM Mg2+ [82]. Even more convincingly, BLM cleavage of a proposed mRNA pseudoknot from a postulated novel gene in HIV-1 reverse transcriptase requires ~5 mM Mg2+ for cleavage [98]. In the absence of Mg2+, no Fe-BLM cleavage was observed. However, an RNA hairpin corresponding to the 3'-RNA hairpin region of the HIV-1 proposed pseudoknot (see Fig. 4), was cleaved at the equivalent uridine in the presence or absence of Mg2+ (unpublished results). These results imply that Mg2+ stabilization of the tertiary structure of the RNA pseudoknot is required for BLM cleavage. Thus, at least for this proposed RNA pseudoknot, Mg2+ is critical for BLM cleavage of the structurally intact RNA substrate. 4 Conclusions and Outlook Fe(II)-BLM cleavage has been demonstrated with every major class of RNA. The cleavage reported is highly selective, with some RNAs not cleaved at all and others exhibiting several prominent cleavage sites. In addition, BLM cleavage of RNA substrates is ~10-fold more selective than BLM cleavage of DNA substrates [79,82]. Data from the Hecht laboratory suggest that RNA is actually bound more tightly by BLM than DNA [84] but that the C4' H in the minor groove of RNA may be less accessible than the C4' H in the minor groove of DNA. This would result in fewer overall cleavage sites and less efficient cleavage, and concurrently would contribute to the significantly greater BLM cleavage selectivity observed for RNA vs. DNA substrates [82]. Fe(II)-BLM is also proving to be a valuable chemical ribonuclease for examining RNA secondary and tertiary structures, as shown by Dix et al. in the IRE region of the mRNA of ferritin [85]. Relative to existing
< previous page
page_610
next page >
< previous page
page_611
next page > Page 611
enzymatic ribonucleases, Fe(II)-BLM exhibits more selective cleavage of RNA substrates, and it has the advantage of preferring to cleave RNA at single-strand/double-strand junctions. This makes it a uniquely valuable probe for RNA structure. The therapeutic relevance of BLM cleavage of RNA has yet to be definitively established, but logic suggests that RNA damage could contribute to the therapeutic effects established for BLM. Destruction of a critical RNA involved in cellular function could help account for the tumor selectivity exhibited for BLM but not totally explained by existing biological effects. Acknowledgment We thank Dr. Sidney M. Hecht for his continued help and support, Ms. Carine Foucault for the BLM cleavage studies with the pseudoknotderived RNA hairpin structures, and Dr. David Mack for valuable discussions. Abbreviations
BLM
bleomycin
EPR
electron paramagnetic resonance
FL
flanking region
HIV
human immunodeficiency virus
HPLC high-performance liquid chromatography IRE
iron regulatory element
NMR
nuclear magnetic resonance
nt
nucleotide
PMAH 2[(N-(aminoethyl)aminomethyl]-4-[N-(2-(4-imidazolyl)ethyl)- carbamoyl]-5bromopyrimidine pur
purine
pyr
pyrimidine
RT
reverse transcriptase
SeCys selenocysteine TLC
thin-layer chromatography
< previous page
page_611
next page >
< previous page
page_612
next page > Page 612
References 1. H. Umezawa in Bleomycin: Current Status and New Developments (S. K. Carter, S. T. Crooke, and H. Umezawa, eds.), Academic Press, New York, 1978, p. 15f. 2. S. M. Hecht in Bleomycin: Chemical, Biochemical and Biological Aspects (S. M. Hecht, ed.), Springer-Verlag, New York, 1979, p. 1ff. 3. H. Umezawa, Lloydia, 40, 67 (1977). 4. S. T. Crooke in Bleomycin: Current Status and Developments (S. K. Carter, S. T. Crooke, and H. Umezawa, eds.), Academic Press, New York, 1978, pp. 914. 5. H. Suzuki, K. Nagai, H. Yamuki, N. Tanaka, and H. Umezawa, J. Antibiot., 22, 446 (1969). 6. H. Umezawa in Bleomycin: Chemical, Biochemical and Biological Aspects (S. M. Hecht, ed.), Springer-Verlag, New York, 1979, p. 24ff. 7. J. Kross, W. D. Henner, S. M. Hecht, and A. W. Haseltine, Biochemistry, 21, 4310 (1982). 8. S. C. Barlogie, B. Drewinko, J. Schumann, and E. J. Freireich, Cancer Res., 31, 1182 (1976). 9. S. C. Barranco and R. M. Humphrey, Cancer Res., 31, 1218 (1971). 10. W. N. Hittelman and P. N. Rao, Cancer Res., 34, 3433 (1974). 11. J. M. Clarkson and R. M. Humphrey, Cancer Res., 36, 2345 (1976). 12. M. Nagatsu, R. M. Richart, and A. Lambert, Cancer Res., 32, 1966 (1972). 13. M. Watanabe, Y. Takabe, T. Katsumata, and T. Terasima, Cancer Res., 43, 878 (1974). 14. R. A. Tobry, J. Cell. Physiol., 79, 259 (1972). 15. D. E. Berry, L.-H. Chang, and S. M. Hecht, Biochemistry, 24, 3207 (1985). 16. H. Ekimoto, K. Takahashi, A. Matsuda, T. Takita, and H. Umezawa, J. Antibiot., 38, 1077 (1985). 17. J. M. C. Gutteridge and X.-C. Fu, FEBS Lett., 123, 71 (1981). 18. H. Kikuchi and T. Tesuka, J. Antibiot., 45, 548 (1992). 19. R. Nagata, S. Morimoto, and I. Saito, Tetrahedron Lett., 31, 4485 (1990).
< previous page
page_612
next page >
< previous page
page_613
next page > Page 613
20. R. J. Guajardo and P. K. Mascharak, Inorg. Chem., 34, 802 (1995). 21. S. Mizuno and A. Ishida, Biochem. Biophys. Res. Commun., 105, 425 (1985). 22. S. Mizuno and A. Ishida, Biochem. Biophys. Res. Commun., 107, 1021 (1982). 23. D. E. Berry, R. E. Kilkuskie, and S. M. Hecht, Biochemistry, 24, 3214 (1985). 24. E. J. Gardiner and D. T. Snustad (eds.), Principles of Genetics, 6th ed., John Wiley and Sons, New York, 1985. 25. A. W. Murray and M. W. Kirschner, Nature, 339, 275 (1989). 26. R. Ishida and T. Takahashi, Biochem. Biophys. Res. Commun., 66, 1432 (1975). 27. E. A. Sausville, J. Peisach, and S. B. Horwitz, Biochemistry, 17, 2740 (1978). 28. N. J. Oppenheimer, C. Chang, L. O. Rodriguez, and S. M. Hecht, J. Biol. Chem., 256, 1514 (1981). 29. G. M. Ehrenfeld, L. O. Rodriguez, S. M. Hecht, C. Chang, J. V. Basus, and N. J. Oppenheimer, Biochemistry, 24, 81 (1985). 30. G. M. Ehrenfeld, N. Murugesan, and S. M. Hecht, Inorg. Chem., 23, 1496 (1984). 31. R. M. Burger, J. Peisach, and S. B. Horwitz, J. Biol. Chem., 256, 11636 (1981). 32. H. Kuramochi, K. Takahara, T. Takita, and H. Umezawa, J. Antibiot., 34, 576 (1981). 33. R. M. Burger, S. B. Horwitz, J. Peisach, and J. B. Wittenberg, J. Biol. Chem., 254, 12299 (1979). 34. S. M. Hecht in Bleomycin: Chemical, Biochemical and Biological Aspects (S. M. Hecht, ed.), Springer-Verlag, New York, 1979, p. 1ff. 35. M. Takeshita and A. P. Grollman in Bleomycin: Chemical, Biochemical and Biological Aspects (S. M. Hecht, ed.), Springer-Verlag, New York, 1979, p. 207ff. 36. L. Giloni, M. Takeshita, F. Johnson, C. Iden, and A. P. Grollman, J. Biol. Chem., 256, 8608 (1981). 37. N. Murugesan, C. Xu, G. M. Ehrenfeld, H. Sugiyama, R. E. Kilkuskie, L. O. Rodriguez, L.-H. Chang, and S. M. Hecht, Biochemistry, 24, 5735 (1985).
< previous page
page_613
next page >
< previous page
page_614
next page > Page 614
38. H. Sugiyama, R. E. Kilkuskie, and S. M. Hecht, Nucl. Acids Res., 12, 1581 (1984). 39. J. C. Wu, J. W. Kozarich, and J. Stubbe, J. Biol. Chem., 258, 4694 (1983). 40. H. Sugiyama, C. Xu, N. Murugesan, and S. M. Hecht, J. Am. Chem. Soc., 197, 4104 (1985). 41. R. M. Burger, J. Peisach, W. E. Blumberg, and S. B. Horwitz, J. Biol. Chem., 254, 10906 (1979). 42. T. E. Westre, K. E. Loeb, J. M. Zalaski, B. Hedman, K. O. Hodgson, and E. I. Solomon, J. Am. Chem. Soc., 117, 1309 (1995). 43. J. W. Sam, X.-J. Tang, and J. Peisach, J. Am. Chem. Soc., 116, 5250 (1994). 44. D. C. Heimbrook, R. L. Mulholland, and S. M. Hecht, J. Am. Chem. Soc., 108, 7839 (1986). 45. N. Murugesan and S. M. Hecht, J. Am. Chem. Soc., 107, 493 (1985). 46. M. Takeshita, A. P. Grollman, E. Ohtsubo, and H. Ohtsubo, Proc. Natl. Acad. Sci., USA, 75, 7983 (1978). 47. A. D. D'Andrea and W. Haseltine, Proc. Natl. Acad. Sci., USA, 75, 3608 (1978). 48. C. K. Mirabelli, A. Ting, C. H. Huang, S. Mong, and S. T. Crooke, Cancer Res., 42, 2779 (1982). 49. R. J. Steighner and L. F. Povirk, Proc. Natl. Acad. Sci., USA, 112, 3196 (1990). 50. L. F. Povirk, Y.-H. Han, and R. J. Steighner, Biochemistry, 28, 5808 (1989). 51. P. K. Mascharak, Y. Sugiura, J. Kuwahara, T. Suzuki, and S. J. Lippard, Proc. Natl. Acad. Sci., USA, 80, 6795 (1983). 52. R. Hertzberg, M. J. Caranfa, and S. M. Hecht, Biochemistry, 24, 5285 (1985). 53. L. D. Williams and I. H. Goldberg, Biochemistry, 27, 3004 (1988). 54. R. P. Hertzberg, M. J. Caranfa, and S. M. Hecht, Biochemistry, 27, 3164 (1988). 55. Y. Sugiura and T. Suzuki, J. Biol. Chem., 257, 10544 (1982). 56. J. Stubbe and J. W. Kozarich, Chem. Rev., 87, 1107 (1987). 57. T. E. Boothe, T. Sakai, and J. D. Glickson, Biochemistry, 22, 4211 (1983).
< previous page
page_614
next page >
< previous page
page_615
next page > Page 615
58. D. M. Chem, T. T. Sakai, J. D. Glickson, and D. J. Patel, Biochem. Biophys. Res. Commun., 92, 197 (1980). 59. S. Y. Lin, and A. P. Grollman, Biochemistry, 20, 7589 (1981). 60. L. F. Povirk, M. Hogan, and N. Dattagupta, Biochemistry, 18, 96 (1979). 61. L. F. Povirk, M. Hogan, and N. Dattagupta, and M. Buechner, Biochemistry, 20, 665 (1981). 62. M. Levy and S. M. Hecht, Biochemistry, 27, 2647 (1988). 63. R. A. Manderville, J. F. Ellena, and S. M. Hecht, J. Am. Chem. Soc., 116, 10851 (1994). 64. J. Kuwahara and Y. Sugiura, Proc. Natl. Acad. Sci. USA, 85, 2459 (1988). 65. H. Sugiyama, R. E. Kilkuskie, L.-H. Chang, L.-T. Ma, S. M. Hecht, G. A. van der Marel, and J. H. van Boom, J. Am. Chem. Soc., 108, 3822 (1986). 66. B. J. Carter, V. S. Murty, K. S. Reddy, S.-N. Wang, and S. M. Hecht, J. Biol. Chem., 265, 4193 (1990). 67. R. J. Guajardo, S. E. Hudson, S. J. Brown, and P. K. Mascharak, J. Am. Chem. Soc., 115, 7971 (1993). 68. C. W. Haidle, R. S. Lloyd, and D. L. Robertson in Bleomycin: Chemical, Biochemical, and Biological Aspects (S. M. Hecht, ed.), Springer-Verlag, New York, 1979, pp. 222-243. 69. R. S. Lloyd, C. W. Haidle, and R. R. Hewitt, Cancer Res., 38, 3191 (1978). 70. R. S. Lloyd, C. W. Haidle, and D. L. Robertson, Biochemistry, 17, 1890 (1978). 71. M. J. Absalon, J. W. Kozarich, and J. Stubbe, Biochemistry, 34, 2065 (1995). 72. M. J. Absalon, J. W. Kozarich, and J. Stubbe, Biochemistry, 34, 2076 (1995). 73. C. W. Haidle and J. Bearden, Jr., Biochem. Biophys. Res. Commun., 65, 815 (1975). 74. C. R. Krishnamoorthy, D. E. Vanderwall, J. W. Kozarich, and J. Stubbe, J. Am. Chem. Soc., 110, 2008 (1988). 75. M. Hori in Bleomycin: Chemical, Biochemical, and Biological Aspects (S. M. Hecht, ed.), Springer-Verlag, New York, 1979, p. 195ff.
< previous page
page_615
next page >
< previous page
page_616
next page > Page 616
76. H. Suzuki, K. Hagai, E. A. Kutsu, H. Yamaki, N. Tanaka, and H. Umezawa, J. Antibiot., 23, 473 (1970). 77. C. W. Haidle, M. T. Kuo, and K. K. Weiss, Biochem. Pharmacol., 21, 3308 (1972). 78. R. S. Magliozzo, J. Peisach, and M. R. Ciriolo, Mol. Pharm., 35, 428 (1989). 79. B. J. Carter, E. de Vroom, E. C. Long, G. A. van der Marel, J. H. van Boom, C. Debouck, and S. M. Hecht, Proc. Natl. Acad. Sci. USA, 87, 9373 (1990). 80. J. D'Alessio, RNA Sequencing, in Gel Electrophoresis of Nucleic Acids: A Practical Approach, (D. Rickwood and B. D. Hanes, eds.), IRL Press Ltd., Oxford, 1982, pp. 173197. 81. C. E. Holmes, B. J. Carter, and S. M. Hecht, Biochemistry, 32, 4293 (1993). 82. S. M. Hecht, Bioconj. Chem., 5, 513 (1994). 83. A. Hüttenhofer, S. Hudson, H. F. Noller, and P. K. Mascharak, J. Biol. Chem., 267, 24471 (1992). 84. C. E. Holmes and S. M. Hecht, J. Biol. Chem, 268, 25909 (1993). 85. D. J. Dix, P.-N. Lin, A. R. McKenzie, W. E. Walden, and E. C. Theil, J. Mol. Biol., 231, 239 (1993). 86. J. McDougall and R. N. Nazar, J. Biol. Chem., 258, 5256 (1983). 87. J.-M. Battigello, M. Cui, S. Roshong, and B. J. Carter, Enediyne-Mediated Cleavage of RNA, in Bioorganic and Medicinal Chemistry Symposium on the Recent Advances in DNA Binding Agents (R. S. Coleman, ed.), Vol. 3, Part 6, 1995, pp. 839849. 88. E. W. Taylor, C. S. Ramanathan, R. K. Jalluri, and R. G. Nadimpalli, J. Med. Chem., 37, 2637 (1994). 89. M. Morgan and S. M. Hecht, Biochemistry, 33, 10286 (1994). 90. A. Natrajan and S. M. Hecht in Molecular Aspects of Anticancer Drug-DNA Interactions (S. Neidle and M. Waring, eds.), Macmillan, London, 1993, pp. 197242. 91. R. J. Duff, E. de Vroom, A. Geluk, S. M. Hecht, G. A. van der Marel, and J. H. van Boom, J. Am. Chem. Soc., 115, 3350 (1993). 92. J. Stubbe and J. W. Kozarich, Chem. Rev., 87, 1107 (1987). 93. I. Saito, T. Morii, and T. Matsuura, Nucl. Acids Res., 12, 95 (1983). 94. L. E. Rabow, J. Stubbe, J. W. Kozarich, and J. A. Gerlt, J. Am. Chem. Soc., 107, 4104 (1986).
< previous page
page_616
next page >
< previous page
page_617
next page > Page 617
95. T. Evans, E. T. Rosenthal, J. Youngblom, D. Distel, and T. Hunt, Cell, 33, 389 (1983). 96. A. W. Murray and M. W. Kirschner, Nature, 339, 275 (1989). 97. S. J. Séror, S. Casarégola, F. Vannier, N. Zouri, M. Dahl, and E. Boye, EMBO J., 13, 2472 (1994). 98. M. Cui, Characterization of RNA pseudoknots, masters thesis, University of Toledo, 1995.
< previous page
page_617
next page >
< previous page
page_619
next page > Page 619
21 Metallobleomycin-DNA Interactions: Structures and Reactions Related to Bleomycin-Induced DNA Damage David H. Petering,1 Qunkai Mao,1 Wenbao Li,1 Eugene DeRose,1 and William E. Antholine2 1Department of Chemistry, University of WisconsinMilwaukee, Milwaukee, WI 53201-0413, USA 2National Biomedical ESR Center, Medical College of Wisconsin, Milwaukee, WI 53226, USA
620
1. Introduction
622
2. Cellular DNA Strand Scission by Bleomycin
624
3. Cellular Metal Ion Requirement for Bleomcyin
4. Reactions of Bleomycin in Nuclear and Plasmid Systems
625
5. In Vitro Mechanisms of DNA Damage
6. Comparative Chemistry of Cobalt and Iron Bleomycin in the Absence of DNA
7. Self-Inactivation Chemistry of Activated Iron Bleomycin
629
633
634
8. Reactions of DNA-Bound Cobalt Bleomycin
635
9. Reactions of DNA-Bound Iron Bleomcyin
< previous page
624
page_619
next page >
< previous page
page_620
next page > Page 620 637
10. Interactions between FeBlm, Other Ligands, and DNA
638 11. Binding of
A2 and Co(III)Blm A2 to DNA Oligomers and Site Selectivity 640
12. Interaction of ZnBlm with DNA
641
13. Site Selectivity of DNA Binding and Damage
641
14. Mechanism of Double-Strand Cleavage
642
Acknowledgments
643
Abbreviations
643
References
1 Introduction Among the common agents used to treat human cancer, which number about 40, at least three require metals for their activity, i.e., cis-diamminedichloroPt(II), bleomycin (Blm), and adriamycin [1,2]. In cancer cells each appears to have as its target the DNA molecule. The simple platinum drug is essentially an electrophilic reagent which binds covalently to DNA bases [3]. Adriamycin forms noncovalent adducts with DNA, inhibits the DNA replication enzyme topoisomerase II, complexes a variety of metal ions including Fe(III), and participates in redox reactions with cellular reductants and 02 [48]. Of the three, bleomycin has the most elaborate structure and perhaps the most complicated chemistry in its interactions with DNA. As shown in Fig. 1, its one-dimensional sequence is made up of a metal binding domain to which is attached a disaccharide group, a DNA binding domain involving a bithiazole moiety with two planar rings and a positively charged R group, and a peptide linking these two parts of the structure. Although there is some difference of opinion about the ligand structure of various metallobleomycins, the five nitrogens labeled in the figure likely compose most if not all of the binding site for various metal ions, including Fe, Cu, Co, Zn, and Cd [913]. Under different conditions Fe, Cu, Co, and Mn complexes of
< previous page
page_620
next page >
< previous page
page_621
next page > Page 621
Fig. 1. Covalent structure of bleomycin A2 and B2. Metal coordination sites shown with dots. bleomycin can cause DNA strand scission [14,15]. Most attention has focused on the solution mechanism of activation of FeBlm to cause such damage. Addition of Fe2+ to Blm and O2 rapidly produces an electron spin resonance (ESR)-active species, which when added to DNA cleaves its backbone or releases free bases starting with the abstraction of the C4' hydrogen of deoxyribose (see Fig. 2 below) [1520]. The same intermediate can be generated by the addition of H2O2 to Fe(III)Blm [16]. The presumed pathway of activation includes the following Eqs. (1)(4).
If these reactions are carried out in the absence of DNA, the activated species decays quickly, losing its capacity to damage DNA with the same rate constant that it breaks down [17]. This observation indicates
< previous page
page_621
next page >
< previous page
page_622
next page > Page 622
that
or another species that rapidly forms from it initiates the reaction of the drug with DNA.
During the decay reaction, reacts with itself, producing an ill-defined modification of the drug that greatly reduces its capacity to cause DNA damage in subsequent reactions (1)(3) [17,21]. The nature of the activated product in reactions (3) and (4) has been shown to be the peroxyl adduct of Fe(III)Blm. It clearly contains a low-spin Fe(III) according to ESR and Mössbauer studies [17,22]. The broadening of its ESR spectrum when 17O2 is used in reaction (2) shows that at least one oxygen atom is part of the activated complex [23]. Recent electrospray mass spectrometry on the adduct indicates that it has a mass consistent with the presence of . This conclusion has been affirmed by extended X-ray absorption fine structure (EXAFS) experiments which concluded that a peroxide group is bound to iron in the activated species [24]. DNA damage can result from reactions (1)(3), reaction (4), and by the reaction of Fe(III)Blm with reducing agents such as ascorbate and 2-mercaptoethanol and dioxygen [reactions (5)(7))] [2527].
Generally, attention has focused on the mechanisms by which single strands of DNA are cleaved and how the site specificity of the drug for 5'-G-pyrimidine-3' sites in the DNA breakage reaction occurs [15,2830]. Since Blm is an effective double-strand cleavage agent, a principal goal is to understand how FeBlm causes double-strand cleavage of DNA. In the sections below, details of the cellular and chemical interactions between metallobleomycin and DNA will be described which center on the mechanism of action of the drug to cause DNA damage. 2 Cellular DNA Strand Scission by Bleomycin Concentration-dependent inhibition of cell proliferation of Ehrlich cells by Blm or its Fe, Cu, or Zn complexes occurs over the range 02 µM [31].
< previous page
page_622
next page >
< previous page
page_623
next page > Page 623
Larger concentrations gradually reduce cell viability, leading to the characteristic ''dog leg" dose-response curve for the drug. Despite the large sensitivity of cells to Blm, most of the early experiments characterizing Blm-induced DNA damage have utilized orders of magnitude more drug than are necessary to halt cell division. When single- and double-strand DNA cleavage is quantified over the same range of Blm concentrations that produce inhibition of cell proliferation, only the appearance of double-strand damage shows a similar concentration dependence, demonstrating that inhibition of cell proliferation could be the result of this type of DNA damage [32,34,35]. The importance of double-strand cleavage has been underscored in studies of the repair of single- and double-strand damage produced by exposure to Blm [33,34]. While single-strand scission is rapidly repaired over the course of 60 min in cells treated with 1,10-phenanthroline (OP) to inhibit the ongoing action of Blm, double-strand breaks are repaired more slowly, leaving about 20% of the fragmented DNA unrepaired after 5 hr [33]. If OP is not applied to these cells, net single and double strand damage increases over time, indicating that continued cleavage occurs which more than offsets repair. The efficacy of Blm as a cleavage reagent for DNA is striking when one considers that only about 1% of the drug in the external medium reaches the interior of the cell [32]. Its poor uptake into cells is presumably due to the abundant polar and charged groups in its glycopeptide structure (Fig. 1). Moreover, only about 3% of the intracellular drug distributes into the nucleus despite the presence of a DNA binding domain in its structure, which has an association constant with DNA on the order of 105 M1 at concentrations of drug that are highly cytotoxic [36,37]. Two estimations of the ratio of base pairs to Blm in the nucleus are 108:1 and 105:1 [31,38]. Because of these huge ratios, once Blm reaches the nucleus it will exist as a bound complex. The fact that strand scission reactions continue over time in the presence of an overwhelming excess of DNA suggests strongly that it is FeBlm bound to DNA that undergoes activation to initiate the cleavage reactions. The alternatives, that unbound Fe(III)Blm picks up H2O2 [reaction (4)] or that unbound Fe(II)Blm reacts as in Eqs. (1)(3) to generate the activated species, appear to be inconsistent with the likelihood that virtually all nuclear Blm is bound to DNA in the steady state. Furthermore, the short half-life of
< previous page
page_623
in solution at 37°C,
next page >
< previous page
page_624
next page > Page 624
estimated to be 15 sec on the basis of its half-life of 120 sec at 6°C, would permit self-inactivation to be a major competing reaction in the cell [17]. 3 Cellular Metal Ion Requirement for Bleomycin Blm, Fe(III)Blm, Cu(II)Blm, and ZnBlm display similar antitumor action and cytotoxic activity against cells in culture [31]. Recent studies with HL-60 and Euglena gracilis cells grown in defined media, which can be varied in nutrient Fe content without recourse to chelating agents, show that iron deficiency lowers or abolishes the effects of Blm on cell viability and DNA strand scission, that CuBlm is similarly affected, and that FeBlm remains fully active under this condition [1]. These results clearly show that FeBlm is the cytotoxic species which inhibits cell proliferation and causes DNA damage. The other metallobleomycins must undergo conversion to FeBlm to become active. Importantly, in the Euglena system it was clear that the loss of drug action occurs only after >90% of the control level of cellular iron is lost from the cells, including the two major cytosolic storage pools [39]. Thus, Blm has a remarkable capacity to compete for scarce cellular iron and appears to be the cytotoxic species. 4 Reactions of Bleomycin in Nuclear and Plasmid Systems The nucleus is a useful model of intermediate complexity between the cell and isolated DNA, which permits the examination of single- and double-strand damage to intact chromosomal DNA by the same alkaline elution and neutral filter elution techniques used in whole cells. At ratios of base pairs to drug of 20200:1, metal-free Blm is almost without effect on nuclear DNA, indicating that the isolated nucleus does not contain available iron to activate Blm [40]. In contrast, Fe(III)Blm is highly active in causing single- and double-strand breaks, which produce DNA fragments on the order of 106107 bases or 105 base pairs in size, respectively. A comparison between these numbers and the con-
< previous page
page_624
next page >
< previous page
page_625
next page > Page 625
centration ratios of base pairs to FeBlm listed above reveals that only a few of the drug molecules need to be activated to cause the observable double-strand damage, similar to that found in cells. Therefore, FeBlm need not function repeatedly as a catalytic site for dioxygen activation to achieve substantial DNA damage. No reductant is added to the reaction mixture of nuclei and Fe(III)Blm to achieve DNA breakage [40]. While about 50% of the damage can be inhibited with N-ethylmaleimide (NEM), which implicates endogenous thiol groups as one source of reducing equivalents, the nature of the other reductants remains unclear. Exogenous NADH and NADPH can stimulate strand scission, possibly through the activity of the cytochrome P-450 reductase and cytochrome b5 reductase in the nuclear envelope to generate nucleotides is an attractive hypothesis.
[41]. In cells, activation of FeBlm by these
The reaction of FeBlm with DNA plasmids such as pBR322 bears similarities to the properties described above for its reaction with nuclei [42,43]. First, both single- and double-strand breakage appear in concert as a function of drug concentration and are formed with the same kinetics. This behavior may be distinguished from that of Fe2+ in which single-strand breaks occur exclusively at lower concentrations. Double-strand damage appears only at larger concentrations, indicating that it results from an accumulation of single-strand scissions that eventually overlap to give the appearance of double-strand cleavages [43]. From studies that have examined the sites of cleavage, it seems clear that double-strand breaks compose a subset of the single-strand sites of reaction [4447]. 5 In Vitro Mechanisms of DNA Damage Numerous studies have investigated aspects of the mechanisms of DNA damage which are initiated by activated iron bleomycin. Figure 2 summarizes two hypothetical pathways which result in either single-strand cleavage (path a) or base release (path b). It is generally accepted that both pathways are started by the specific reaction of with the C4' hydrogen of deoxyribose located in the backbone linkage between the two bases of the specific site of cleavage, 5'-G-pyrimidine-3'
< previous page
page_625
next page >
< previous page
next page >
page_626
Page 626
Fig. 2. Pathways of DNA degradation initiated by
< previous page
page_626
.
next page >
< previous page
page_627
next page > Page 627
[4850]. In path a, the resultant radical reacts with O2 to form a peroxy radical, which undergoes reduction to the C4' peroxide. Eventually this species degrades to the 5'-phosphoglycolate, 3'-phosphate, and a base propenal [26,28,29,51,52]. The other pathway is dioxygen-independent, proceeding from the C-4' carbon radical by oxidation to the C4' hydroxide [53]. Ultimately, this rearranges with the loss of the nucleotide base and formation of an abasic, ring-opened sugar with a ketone in the C4' position [26,28,54,55]. Under basic conditions this can undergo a backbone cleavage reaction. Beyond the initial abstraction of the C4' hydrogen, FeBlm participates in the organic pathways at least in the first reaction of pathway b, which effectively involves the oxidation of C4' radical together with the hydroxylation of the resultant carbonion ion. It has recently been shown that the oxygen of the C4' hydroxide is derived from bulk water, not the peroxide of , and that the oxygens of the activated species do not undergo rapid exchange with bulk water [53,56]. Together they imply that some version of the following reactions occur in this step:
FeBlm may also mediate the reduction of in pathway a. However, it has been suggested that glutathione, which is not kinetically competent to initiate strand scission through the reduction of Fe(III)BlmDNA, can provide electrons to this intermediate [57]. Detailed examination of the site distribution of products of the reaction of activated FeBlm with DNA, namely, single- and double-strand cleavage and base release, has revealed the following rules for combinations of reactions [45,46]:
Primary strand 1: (5'-G-pyrimidine-3')
Secondary strand 2:
1. Single-strand cleavage
No reaction Single-strand cleavage (double-strand cleavage); base release
2. Base release
No reaction
Single-strand cleavage or base release, paths a and b, may occur on the site-specific strand 1 without reaction on the other strand. Or, single-strand cleavage on strand 1 may be accompanied by cleavage of the other strand directly opposite or offset by one nucleotide to produce a
< previous page
page_627
next page >
< previous page
page_628
next page > Page 628
double-strand break. Finally, single-strand scission on 1 may be paired with base release on strand 2. Recent experiments complicate this set of rules by showing that the strand bearing the 5'-G-pyrimidine-3' sequence may be attacked either first or second during double-strand cleavage so that there can be random order strand cleavage at a doublestrand breakage site [47,58]. The stoichiometry of reaction of with C4'H may be considered in light of these rules [15,43]. At the outset of the base release path, two-electron oxidation reduction takes place:
For single-strand cleavage a one-electron redox process occurs which produces a species, (OH)Fe(III)Blm, which may also be described as O=Fe(IV)Blm:
In principle, therefore, double-strand scission can be initiated by two rounds of reaction (11):
Lacking in these reactions is a means to reduce (OH)Fe(III)Blm or O=Fe(IV)Blm to Fe(III)Blm and OH when only single-strand breakage occurs or when it is coupled with base release on the other strand. Significant concentrations of hydroxyl radical are not detected during the reaction of FeBlm with DNA, indicating that it is not simply released as a product of reaction (10) [59]. According to these reactions, single-strand breakage followed by base release on the other strand would require reactivation of Fe(III)Blm. A different scheme for the strand cleavage reactions links the first hydrogen abstraction reaction to the need for another electron in path a to reduce the peroxy radical to C4'O2H [58]. Assuming that the actual species which reacts with the C4' hydrogen is a perferryl species:
The ferryl species, O=Fe(IV)Blm, equivalent to (OH)Fe(III)Blm, then provides the required electron and in the process is oxidized back to the perferryl structure:
< previous page
page_628
next page >
< previous page
page_629
next page > Page 629
Pathway b is initiated as in reaction (11) through a combination of reactions (13) and (16):
If only single-strand breakage occurs at the site, extra oxidizing equivalents in the form of O=Fe(V)Blm are left over and unaccounted for as above. If double-strand cleavage is to take place, it is ready to initiate reaction on the second strand. Nevertheless, after repetition of reactions (14) and (15), again the reactive perferryl species is formed. While this set of reactions provides a mechanism for the catalytic cleavage of the DNA backbone, the evidence to date is that Fe2+, Blm, and O2 or Fe(III)Blm and H2O2 react stoichiometrically with DNA to cause strand scission [58,60]. Both sets of reactions show that, in principle, double-strand scission may occur at a site as it interacts with a single molecule of . Nevertheless, the fact that several products may result from such reactions indicates that other processes at the site contribute to the actual distribution of products. The nature of the actual form of FeBlm, which attacks DNA, is not known. The only observable species that is kinetically competent to damage DNA is , as described above [17]. That this adduct is first converted to a ferryl or perferryl compound, which then reacts with DNA, has been hypothesized [14,15,58,61]. However, recent EXAFS evidence that has a rather long FeO bond has suggested that the structure is not poised to carry out homo- or heterolytic cleavage of peroxide as in reaction (13) and may react directly with DNA as the peroxide [24,61]. 6 Comparative Chemistry of Cobalt and Iron Bleomycin in the Absence of DNA Excellent studies have described the formation of activated FeBlm in solution and its capacity to initiate single-strand cleavage when added to DNA [17,62]. However, until recently little work has inquired into the details of these reactions when DNA is present.
< previous page
page_629
next page >
< previous page
page_630
next page > Page 630
The reaction of Blm, Co2+, and O2 has been examined in the absence and presence of DNA as a model for corresponding reactions involving FeBlm. In the absence of polymer, the following pathway of reaction has been defined [63]:
The dioxygenated product of reaction (18) can be observed by ESR spectroscopy [64]. The dimer intermediate formed in reaction (19) is detected by time-dependent 1H nuclear magnetic resonance (NMR) spectroscopy and O2 analysis of the reaction mixture. Finally, the two products. Forms I and II, have been separated and characterized to show that peroxide is stoichiometrically present in Form I. The sequence of reactions following the mixing of Fe2+ with Blm and O2 occurs much more rapidly than those described above. Only the
< previous page
page_630
next page >
< previous page
page_631
next page > Page 631
Fig. 3. Three-dimensional conformations of Co(II)Blm: (a) Form I, A2 (b) Form I, rotated about 180° from view in (a); (c) Form II, Cod(II)Blm A2. proposed intermediate, O2-Fe(II)Blm, has been detected by stopped flow spectrophotometry [16]. In contrast to products of reaction of Fe2+, Blm, and O2, which are both paramagnetic and include the highly reactive activated form, the cobalt complexes, Form I and Form II, are diamagnetic and can be examined in detail by NMR spectroscopy. A two-dimensional NMR structural analysis of both products has been done [65]. The threedimensional structures of Forms I and II are shown in Fig. 3.
< previous page
page_631
next page >
< previous page
page_632
next page > Page 632
There are several unexpected features of these structures. It had been thought, on the basis of NMR studies of Blm, ZnBlm, and COFe(II)Blm, that metallobleomycins exist in an extended structure that can be conveniently divided into three parts: the metal and DNA binding domains and the peptide which links the two [9,12,66]. In this model, the efficiency of FeBlm as a DNA cleavage reagent results simply from the proximity of the metal binding site to DNA that is achieved through association between DNA and the other end of the drug molecule. However, in Forms I and II it is evident that both assume similar folded structures involving the cobalt coordination center and the peptide linker region. In addition, in Form I the bithiazole moiety is folded back across the metal binding domain, achieving a compact conformation (Fig. 3a, b). The importance of this finding is considered below in the mechanism of self-inactivation of
and in the reactions of Co- and FeBlm that occur in the presence of DNA.
Second, it is appreciated that the metal coordination site of Blm can adopt two chiral configurations (Fig. 4) [65]. From molecular dynamic calculations it is hypothesized that Form I exists in one of these configurations and Form II in the other. They do not interconvert because of the presence of Co(III), which is kinetically inert to ligand dissociation. In the first, the DNA-binding portion of the molecule can fold back across the metal coordination site, placing the bithiazole moiety over the peroxide group that is bound in the axial, sixth position around the Co(III) ion (Fig. 3a, b). In Form II, the bithiazole retains the
Fig. 4. Metal-ligand configurations of Co(III)Blm in Form I (a) and Form II (b) according to molecular dynamics calculations. Alaninamide (A) primary (1°) and secondary (2°) amine nitrogens; pyrimidine (P) N5 nitrogen; histidine amide (HN) and imidazole (H) N1 nitrogens.
< previous page
page_632
next page >
< previous page
page_633
next page > Page 633
flexibility to fold over the metal domain but can only do so above the fifth coordination site bearing the axial primary amine nitrogen (Fig. 3c and 4b). Therefore, as the drug molecule wraps around a metal ion, it establishes two different configurational relationships between the metal and DNA domains of Blm. This implies that even if the bithiazole groups of Forms I and II are both in folded or both extended conformations upon interaction with DNA, their metal centers will adopt different structural orientations with respect to DNA. What that might imply about reactions of FeBlm when bound to DNA is discussed below. 7 Self-Inactivation Chemistry of Activated Iron Bleomycin One of the intriguing features of FeBlm is that its activated form undergoes self-degradation in the absence of a DNA substrate for reaction. This reaction is first order in , indicating that a single reactive molecule modifies itself [17]. On the basis of altered ultraviolet and fluorescence spectra of redox-inactivated FeBlm (FeRIBlm), it was postulated that the bithiazole group is altered in this reaction [67]. Indeed, the conformation of Form I,
A2, if analogous to that of the corresponding iron species, suggests that the activated
peroxide in could react with the bithiazole moiety in the folded structure (Fig. 3). Other data show that RIBlm reacts efficiently with Fe2+ and O2 to produce the activated species as observed by the appearance of its unique ESR spectrum [68]. This result suggests that there is no modification of the iron coordination site. Ongoing one-dimensional and two-dimensional 1H NMR studies indicate that the bithiazole residue is indeed modified in an unspecified way [43]. The only other alteration in the entire structure is the loss of one of the methyl groups of the dimethylsulfonium tail (Fig. 1). That this part of Blm can interact with the metal center of the drug is shown by the finding that the 113Cd chemical shifts of CdBlm A2 and B2 differ by 2 ppm [13,43]. Since the only distinction between these two species resides in the R group, the chemical environment of the Cd center must include this region of the structure. On the basis of this provisional information, it seems that RIFeBlm displays substantially reduced
< previous page
page_633
next page >
< previous page
page_634
next page > Page 634
DNA strand scission activity simply because it cannot bind to DNA in the absence of an intact bithiazole and positively charged R group. 8 Reactions of DNA-Bound Cobalt Bleomycin The fully characterized reaction pathway for the oxidation of Co(II)Blm by O2 provided the basis for the examination of the same redox process in the presence of DNA [63]. Essentially the same pathway of reaction occurs as summarized by reactions (17)(20), although the dimer product of reaction (19) has not been observed [69]. Nevertheless there are key differences that are fundamentally important to understanding how FeBlm causes DNA strand breakage. First, as the base-pair-to-drug ratio increases, the rate of oxidation of Co(II) declines precipitously such that at a ratio of 10 base pairs to 1 drug molecule a species identified as O2-CoBlm is stable for hours. Since the size of the binding site for Form I or II is 23 base pairs, it is clear that the dioxygenated species cannot readily dissociate once it is bound and thus cannot rapidly or efficiently reorganize itself along the polymer to undergo bimolecular reaction required for its oxidation [reaction (19)]. Thus it is hypothesized that it is also the stable, random binding of FeBlm to DNA sites at a distance from one another which causes the well-known ''inhibition" of DNA cleavage by Fe(II)Blm as the base-pair-to-Blm ratio increases [70]. The stability of the dioxygenated adduct of Co(II)Blm bound to DNA is shown by the inability to remove dioxygen from this species by extensive incubation under vacuum [69]. Like other Co(II)-dioxygen complexes, the CoBlm adduct displays an ESR spectral signature of [64]. By binding this structure to oriented DNA fibers and examining its ESR spectrum as a function of orientation of the fibers in the magnetic field, it has been shown that the dioxygen ligand is rigorously confined to a plane perpendicular to the DNA polymer axis [71]. By inference, the Co binding domain of the drug is fixed with respect to the DNA structure. In contrast, the metal centers of Fe(III)and Cu(II)Blm retain partial rotational freedom [7274]. This result strongly suggests that both the DNA and metal binding domains of O2-Co(II)Blm strongly interact with the DNA structure. Because O2-Co(II)BlmDNA undergoes
< previous page
page_634
next page >
< previous page
page_635
next page > Page 635
oxidation-reduction with Fe(II)BlmDNA bound to adjacent sites on the polymer, it is hypothesized that the structures of DNA-bound CoBlm species resemble those of the corresponding Fe(II)BlmDNA structures [69]. According to these results, the model is oversimplified such that only the DNA domain tethers the drug to DNA to bring the reactive metal domain into the proximity of the polymer in order to raise the efficiency of the DNA damage reaction. That the metal domain of key forms of CoBlm is intimately associated with DNA also applies to Form I, . In isolation the peroxide is lost from the structure under a variety of treatments, including lyophilization. However, when bound to DNA oligomers, Form I is stabilized, suggesting that its interaction with DNA prevents the dissociation of the peroxide ligand [75]. 9 Reactions of DNA-Bound Iron Bleomycin It is hypothesized that the reactions of Co(II)Blm associated with DNA are an excellent model for the behavior of DNA-bound Fe(II)Blm. Whereas the detailed mechanism of reaction of Fe2+, Blm, and O2 has not been resolved because of its fast kinetics, the behavior of these reactants in the presence of DNA is substantially slower and can be studied with conventional kinetic methods. In qualitative detail, the reaction pathway is identical to that of Co(II)Blm, except that the product proceeds to react with the DNA backbone [76]. Fe(II)BlmDNA first reacts with O2 to form a dioxygen adduct, which is increasingly stable to oxidation as the average distance between base pairs increases [reaction (21)] [76].
As with the analogous cobalt structure, this dioxygen species is highly stable. Dissociation of the complex occurs at a rate constant of 0.16 min1, which is two orders of magnitude smaller than rate constants for the same process involving oxy- α- or β-hemoglobin or the model heme complex, dioxygenated pyrroheme-N-[3-(1imidazoyl)propyl]amide Fe(II) [76]. Furthermore, ligand substitution with 4,7-phenylsulfonyl-1,10-phenanthroline (BPS), which occurs within the time of mixing with Fe(II)Blm or Fe(II)BlmDNA, takes place much more slowly when O2 is
< previous page
page_635
next page >
< previous page
page_636
next page > Page 636
bound to the iron center. Either the dioxygen adduct is uniquely stable in these two reactions or, more likely, the metal coordination domain interacts with DNA, thereby modifying the rates of these processes. The rate and extent of reaction of Fe(II)BlmDNA and O2 to cause DNA strand scission as measured by the formation of base propenals and free bases is inversely related to the base-pair-to-drug ratio as hypothesized on the basis of the reaction of Co(II)BlmDNA and O2 [76]. Indeed, the reaction of two molecules of O2-Fe(II)BlmDNA to generate the activated form [reaction (22)] occurs with DNA-dependent secondorder constants that are similar to the rate constants for base propenal formation.
When the activated species is generated by reaction (4), which does not require electron transfer between two molecules of FeBlmDNA, there is no effect of base-pair-to-drug ratio on production of DNA cleavage products [76]. However, in reaction (22), as the base-pair-to-FeBlm ratio increases, the dioxygen adduct becomes stable and no strand scission occurs. Since in the cellular situation very large ratios of DNA to drug exist as described above, it is evident that reductants other than Fe(II) itself must participate to activate the drug. From the experiments with nuclei described above, NADH- or NADPH-dependent redox reactions effectively stimulate DNA strand scission by FeBlm [40]. That superoxide anion might be the reducing species is supported by models studied with microsomes, showing that NADPH-stimulated activation of FeBlm is inhibited by superoxide dismutase [77]. With ratios of base pairs to Fe(III)Blm of at least 2550:1, ascorbate is also an effective physiological reductant [78]. Thiols such as 2mercaptoethanol may also provide reducing equivalents but the prevalent cellular thiol glutathione is not an efficient reducing agent [25,57]. The one-electron reduction potential of Fe(III)Blm in the absence of DNA is 129 mV vs. the standard hydrogen electrode [79]. Thus all of the reductants mentioned above can reduce FeBlm in solution under equilibrium conditions. Perhaps glutathione is not an effective reductant because its steric bulk may retard the kinetics of sulfhydryl interaction with FeBlm. Whether binding to DNA affects the redox properties of FeBlm is not yet established.
< previous page
page_636
next page >
< previous page
page_637
next page > Page 637
10 Interactions between FeBLM, Other Ligands, and DNA There are several reactions involving Fe(III)BlmDNA which suggest that the metal domain interacts intimately with DNA and that this interaction varies depending on the base-pair-to-drug ratio. First, upon acidification in nonphosphate buffer, Fe(III)Blm undergoes a transition from low-spin to high-spin iron, which involves the dissociation of one nitrogen ligand, perhaps the axial nitrogen from the five-coordinate Fe(III) center [80]:
Upon addition of DNA at pH 4, the low-spin form is restored in a reaction most simply understood as the rebinding of the axial amine:
Second, the azide group of
is displaced upon binding to DNA [80]:
Third, in phosphate buffer Fe(III)Blm is high spin; upon reaction with DNA it is converted to a low-spin form [17]:
Finally, upon addition of ON-Fe(II)Blm to DNA, the ESR spectrum of the complex is perturbed to a more rhombic form [81]:
In all of these examples, it is evident that DNA markedly alters some structural features of the metal coordination site. The mechanisms of these effects of DNA on metal domain properties are not yet known. Indeed, at least some of the reactions are more subtle than indicated above. In a recent reexamination of reaction (26), it was appreciated that the high-spin/low-spin transition which occurs upon titrating Fe(III)Blm with DNA takes place while the bithiazole group is bound to DNA according to fluorescence spectroscopy [78]. The same situation probably applies to reaction (27). Furthermore, the rate constant for reaction of BPS with O2-Fe(II)BlmDNA is also related to
< previous page
page_637
next page >
< previous page
page_638
next page > Page 638
the base-pair-to-drug ratio [76]. It is hypothesized that the binding of various forms of the FeBlm to DNA constrains the conformations of both drug and DNA, and that such interactions change as the sites of binding become isolated from one another. Supporting this view is a study showing that the binding of CuBlm to DNA causes long-distance conformational effects in the polymer [82]. 11 Binding of
A2 and Co(III)BLM A2 to DNA Oligomers and Site Selectivity
The close analogy between the mechanisms of reaction of Co(II)Blm and Fe(II)Blm with dioxygen in the absence and presence of DNA, the availability of three-dimensional structures of Forms I and II Co(III)Blm, and the observation that Form I cleaves DNA photochemically with the same site specificity as activated FeBlm, have encouraged the examination of the structures of Forms I and II bound to DNA oligomers [63,65,69,75,83]. Although work on such structures is still in progress, important features of these complexes are emerging. From NMR analyses of the interaction of A2 with oligomers, d(CCAGGCCTGG)2 (DNAa), d(GGAAGCTTCC)2 (DNAb), and d(AAACGTTT)2 (DNAc), containing preferential cleavage sites, 5'-GC-3' or 5'-GT-3', it is evident that Form I binds to each in slow exchange on the NMR time scale [75,83]. The B5 and B5' protons of the bithiazole moiety intercalate in the minor groove between base pairs including the 3' side of the two sites of reaction on DNA (Figs. 1 and 3a). These include C6G15-C7G14, C6G15-T7A14, and T6A11-T7A10, showing that several base pairs on the 3' side of the cleavage site are compatible with a similar binding interaction for the bithiazole. Close association of the positively charged dimethylsulfonium group with the backbone has also been observed in the Form I-DNAa complex [83]. Almost all of the intramolecular 1H nuclear Overhauser effects (NOEs) that exist in Form I in solution remain in the DNA complex, except for the one between B5 and the pyrimidine methyl (PCH3) that is lost as the bithiazole slides between DNA base pairs (Fig. 3a) [65,75]. This is replaced by a new intermolecular NOE between this methyl group and a G5 2-amino proton in the minor groove of each complex
< previous page
page_638
next page >
< previous page
page_639
next page > Page 639
[75,83]. This result demonstrates that the metal domain is closely associated with the 5' end of the cleavage site. That can only occur if a compact, highly folded structure of Form I exists when bound to DNA that closely resembles its conformation in solution. In this structural relationship between the drug and DNA, the peroxide is placed between the two base pairs of the cleavage site in a favorable location to interact with the C4' hydrogens. Thus, Form I associates with either selective site of cleavage with a similar overall conformation. If an analogous structure of were present, site-selective cleavage would be expected. Considering that the metal domain-linker comprises a closely packed unit in Form I (Fig. 3a, b), in which the DNA domain folds over the metal domain to create an enclosed space for the peroxide ligand, it is not difficult to imagine that the bithiazole might be shifted from its solution conformation without significantly disturbing the rest of the structure as it inserts between DNA base pairs. If that is the case, then there must be other stabilizing interactions between the metal domain and DNA to define the compact, folded structure of Form I bound to DNA proposed above. Otherwise, a more extended conformation of Form I might exist when it is bound to DNA, e.g., a conformation like the one Form II adopts in solution (Fig. 3c). These findings strengthen the hypothesis that when various forms of CoBlm and possibly FeBlm interact with DNA, both the DNA and metal domains participate in the binding process [65,81]. Nevertheless, the placement of a highly folded metallobleomycin in the minor groove of DNA may necessitate substantial deformation of the polymer. In BDNA the phosphorus-to-phosphorus distance across the double helix in the minor groove is about 11 Å [84]. The smallest dimension for the metal domain plane containing PCH3 is 1011 Å. Considering that the phosphorusphosphorus distance between the nearest van der Waals surfaces is only 6 Å in the minor groove, significant metal domain-DNA interactions would seem to require some reorganization of the DNA structure. That such is the case is consistent with several findings described above, showing that properties of the metal domain are dependent on the base-pair-to-drug ratio. NMR spectra of the Form II-DNAb complex are strikingly different from those for the analogous Form I adduct [75]. Form II binds in fast exchange to DNAb, shows no conclusive evidence of intercalation of
< previous page
page_639
next page >
< previous page
page_640
next page > Page 640
its bithiazole moiety into the base pair structure, and does not display an NOE between a G5 2-amino proton and PCH3. Nevertheless it retains virtually all of its intramolecular solution NOEs. The radically different DNA binding behavior of Forms I and II must result from the different metal domain structures of these molecules because each has the same DNA domain. Possibly intercalation requires a particular structure for the metal domain. For example, according to molecular dynamics calculations on Forms I and II, their metal domains have different ligand chiralities about Co(III) [75]. Secondarily, these impose distinctly different stereochemical relationships between metal and DNA domains, which were mentioned above. Consequently, it may be impossible for Forms I and II to adopt the same folded conformation when associated with DNA. 12 Interaction of ZnBLM with DNA Information on the binding of ZnBlm to d(CGCTAGCG)2 recently appeared which indicates that ZnBlm and this 8mer duplex bind in fast exchange on the NMR time scale [85]. It is concluded that the drug lies in the minor groove in a fully extended conformation. Again, as with Form II-DNAb, there is no evidence for intercalation of the bithiazole into the stacked base pairs. A possible explanation stems from comparative studies of ZnBlm and CdBlm [13]. The one-dimensional 13C spectra of Zn- and 113CdBlm are virtually the same and undergo similar temperature-dependent chemical exchange processes, suggesting that they assume similar conformations. 13C-113Cd couplings observed as splittings in the 13C spectrum of 113CdBlm only detect stable metal ligation by nitrogens of the primary amine, the pyrimidine, and imidazole groups as well as temperature-dependent binding by the secondary amine nitrogen. Evidently, in comparison with Form I or II, CdBlm has only a partially formed metal domain. By inference, ZnBlm also may not adopt the same metal coordination geometry as the Co(III)Blm species. Similarly, the metal domain of ZnBlm, which is probably three- or four-coordinate, not five-coordinate as in Forms I and II, may not support folding of the linker against the metal domain that is a fundamental part of the Form I and Form II structures. That this may be correct is suggested by the
< previous page
page_640
next page >
< previous page
page_641
next page > Page 641
NMR analysis of ZnBlm, which concluded that the structure exists in a fully extended form without much metal domain-linker folding [12]. 13 Site Selectivity of DNA Binding and Damage The common but highly selective cleavage and base release damage at 5'-G-pyrimidine-3' sites was originally attributed to unknown, specific interactions between the DNA domain of Blm and DNA [86]. Several evaluations of the role of the DNA domain in site selectivity have provided conflicting results. Bleomycin and phleomycin, a closely related drug with a partially reduced, nonplanar bithiazole moiety, display the same site selectivity, suggesting that a modified bithiazole ring structure which cannot intercalate into DNA has little to do with site specification [87,88]. This conclusion has been bolstered by the finding that the DNA domain tethered to FeEDTA demonstrates no preference for G-pyrimidine base sequences [89]. Nevertheless, when the bithiazole is replaced by a monothiazole ring in a synthetic Blm molecule, all selectivity is lost [90]. Another series of synthetic bleomycins have been constructed in which the metal and DNA domains are attached through a series of peptide linkers differing in the number of glycine residues [91]. All of the molecules display normal site specificity, leading the authors to conclude that the metal domain, not the DNA domain, is primarily responsible for this property of reaction. Emerging from the NMR structural studies of DNA complexes is the model that both metal and DNA domains interact with the minor groove of the DNA polymer, the metal domain with guanine and the DNA domain with the pyrimidine [75,83]. This strongly suggests that binding specificity and hence site specificity for DNA damage resides in the molecule as a whole. 14 Mechanism of Double-Strand Cleavage A detailed picture of the mechanism of DNA double-strand cleavage is not yet in hand. Some of the aspects of this problem are considered here.
< previous page
page_641
next page >
< previous page
page_642
next page > Page 642
A fundamental question is whether a single molecule of FeBlm acts to cleave both strands or whether each strand can be cleaved independently with FeBlm dissociating from the site between these events [43,46,58,92]. Data showing that the ratio of single- to double-strand cleavage is independent of FeBlm concentration and that both types of damage occur with the same kinetics suggest that they result from the action of one molecule bound at the site [43,47,58]. Similarly, in cells where the ratio of DNA to nuclear Blm is at least 105:1, it is highly unlikely that FeBlm, having cleaved one strand and dissociated from it, could rebind to the same site during random collisions with DNA [32]. A second question which follows is whether FeBlm at the site of reaction must be activated once or twice to cause DNA damage. Reactions (12), (14), and (15) show that it is possible to initiate breakage of the DNA backbone on both strands with a single molecule of but do not reveal the chemistry which partitions the damage at a site between single- and double-strand scission. Perhaps an event occurs after the first cleavage, such as dissociation of the complex from the site, which competes with reaction on the other strand. However, the inability to account for all the oxidizing equivalents in either set of reactions leaves open the possibility that FeBlm is activated twice as it causes double-strand scission. A key issue is how the metal domain is able to interact with stereochemical precision on the two strands to extract both C4' hydrogens, which in B-DNA are 18 Å apart if related to the same base pair or greater if offset by one base pair [58]. It is tempting to consider the intercalated bithiazole as the anchor for the complex and the folded linkermetal domain as a structural unit that can move with respect to the DNA domain. Possibly it can orient the peroxide and whatever species reacts with the other strand, sequentially, into close proximity to the two C4' hydrogens. Acknowledgments The authors appreciate the support of NIH grant CA-22184 and American Cancer Society grant DHP-31.
< previous page
page_642
next page >
< previous page
page_643
next page > Page 643
Abbreviations
Blm
bleomycin
BPS
4,7-phenylsulfonyl-1,10-phenanthroline
EDTA
ethylenediaminetetraacetic acid
ESR
electron spin resonance
EXAFS
extended X-ray absorption fine structure
A2
Form I Form II
Co(III)Blm A2
NADH
reduced nicotinamide adenine dinucleotide
NADPH
reduced nicotinamide adenine dinucleotide phosphate
NEM
N-ethylmaleimide
NMR
nuclear magnetic resonance
NOE
nuclear Overhauser effect
OP
1,10-phenanthroline
RIBlm
redox-inactivated bleomycin
References 1. K. Radtke, F. A. Lornitzo, R. W. Byrnes, W. E. Antholine, and D. H. Petering, Biochem. J., 302, 655 (1994). 2. S. Nyayapati, G. Afshan, F. Lornitzo, R. W. Byrnes, and D. H. Petering, Fr. Rad. Biol. Med., accepted (1995). 3. S. L. Bruhn, J. H. Toney, and S. J. Lippard, Prog. Inorg. Chem., 38, 477 (1990). 4. A. H.-J. Wang, G. Ughetto, G. J. Quigley, and A. Rich, Biochemistry, 26, 1152 (1987). 5. A. H. Corbett and N. Osheroff, Chem. Res. Toxicol., 6, 585 (1993). 6. H. Beraldo, A. Garnier-Suillerot, L. Tosi, and F. Lavelle, Biochemistry, 24, 284 (1985). 7. L. Gianni, J. L. Zweier, A. Levy, and C. E. Myers, J. Biol. Chem., 260, 6820 (1985).
8. H. G. Keizer, H. M Pinedo, G. J. Schuurhuis, and H. Joenje, Pharmac. Ther., 47, 219 (1990). 9. M. A. J. Akkerman, E. W. J. F. Neijman, S. S. Wijmenga, C. W. Hibers, and W. Bermel, J. Am. Chem. Soc., 112, 7462 (1990).
< previous page
page_643
next page >
< previous page
page_644
next page > Page 644
10. Y. Iitaka, H. Nakamura, T. Nakatani, Y. Muraoka, A. Fujii, T. Takita, and H. Umezawa, J. Antibiot., 31, 1070 (1978). 11. J. C. Dabrowiak and M. Tsukayama, J. Am. Chem Soc., 103, 1543 (1981). 12. M. A. J. Akkerman, C. A. G. Haasnoot, and C. W. Hilbers, Eur. J. Biochem., 173, 211 (1988). 13. J. D. Otvos, W. E. Antholine, S. Wherli, and D. H. Petering, Biochemistry, accepted (1995). 14. J. Stubbe and J. W. Kozarich, Chem. Rev., 1107 (1987). 15. D. H. Petering, R. W. Byrnes, and W. E. Antholine, Chem.-Biol. Interact., 73, 133 (1990). 16. R. M. Burger, S. B. Horwitz, J. Peisach, and J. B. Wittenberg, J. Biol. Chem., 254, 12299 (1979). 17. R. M. Burger, J. Peisach, and S. B. Horwitz, J. Biol. Chem., 256, 11636 (1981). 18. H. Kuramochi, K. Takahashi, T. Takita, and H. Umezawa, J. Antibiot., 34, 576 (1981). 19. H. Suzuki, K. Nagai, H. Yamaki, N. Tanaka, and H. Umezawa, J. Antibiot., 22, 446 (1969). 20. C. W. Haidle, K. K. Weiss, and M. T. Kuo, Mol. Pharmacol., 8, 531 (1972). 21. R. M. Burger, J. Peisach, and S. B. Horwitz, J. Biol. Chem., 257, 3372 (1982). 22. R. M. Burger, T. A. Kent, S. B. Horwitz, E. Münck, and J. Peisach, J. Biol. Chem., 258, 1559 (1983). 23. J. Sam, X. Tang, and J. Peisach, J. Am. Chem. Soc., 116, 5250 (1994). 24. T. E. Westre, K. E. Loeb, J. M. Azleski, B. Hedman, K. O. Hodgson, and E. I. Solomon, J. Am. Chem. Soc., 117, 1309 (1995). 25. E. A. Sausville, J. Peisach, and S. B. Horwitz, Biochemistry, 17, 2740 (1978). 26. E. A. Sausville, J. Peisach, and S. B. Horwitz, Biochemistry, 17, 2746 (1978). 27. G. R. Buettner and P. L. Moseley, Biochemistry, 31, 9784 (1992). 28. A. D. D'Andrea and W. A. Haseltine, Proc. Natl. Acad. Sci. USA, 75, 3608 (1978).
< previous page
page_644
next page >
< previous page
page_645
next page > Page 645
29. M. Takeshita, A. P. Grollman, E. Ohtsubo, and H. Ohtsubo, Proc. Natl. Acad. Sci. USA, 75, 5983 (1978). 30. L. F. Povirk, W. Köhnlein, and F. Hutchinson, Biochem. Biophys. Acta, 521, 126 (1978). 31. E. A. Rao, L. A. Saryan, W. E. Antholine, and D. H. Petering, J. Med. Chem., 23, 1310 (1980). 32. R. W. Byrnes, J. Templin, D. Sem, S. Lyman, and D. H. Petering, Cancer Res., 50, 5275 (1990). 33. R. W. Byrnes and D. H. Petering, Biochem. Pharmacol., 41, 1241 (1991). 34. R. W. Byrnes and D. H. Petering, Rad. Res., 134, 343 (1993). 35. R. W. Byrnes and D. H. Petering, Rad. Res., 137, 162 (1994). 36. M. A. Chien, A. P. Grollman, and S. B. Horwitz, Biochemistry, 16, 3641 (1977). 37. L. F. Povirk, M. Hogan, N. Dattagupta, and M. Buechner, Biochemistry, 20, 665 (1981). 38. S. N. Roy and S. B. Horwitz, Cancer Res., 44, 1541 (1982). 39. S. Lyman, P. Taylor, F. Lornitzo, A. Weir, D. Stone, W. E. Antholine, and D. H. Petering, Biochem. Pharmacol., 38, 4273 (1989). 40. R. W. Byrnes and D. H. Petering, Biochem. Pharmacol., 48, 575 (1994). 41. I. Mahmutoglu and H. Kappus, Mol. Pharmacol., 34, 1302 (1988). 42. L. F. Povirk, Y.-H. Han, W. Köhnlein, and F. Hutchinson, Nucl. Acids Res., 4, 2573 (1977). 43. D. H. Petering, P. Fulmer, W. Li, Q. Mao, and W. E. Antholine, in Genetic Response to Metals (B. Sarkar, ed.), Marcel Dekker, New York, 1995, Chapter 12. 44. M. J. Mclean, A. Sar, and M. J. Waring, J. Mol. Recog., 1, 184 (1989). 45. L. F. Povirk, Y.-H. Han, and R. J. Steighner, Biochemistry, 28, 5808 (1989). 46. R. J. Steighner, and L. F. Povirk, Proc. Natl. Acad. Sci. USA, 87, 8350 (1990). 47. M. J. Absalon, J. W. Kozarich, and J. Stubbe, Biochemistry, 34, 2065 (1995).
< previous page
page_645
next page >
< previous page
page_646
next page > Page 646
48. J. C. Wu, J. W. Kozarich, and J. Stubbe, J. Biol. Chem., 258, 4694 (1983). 49. J. C. Wu, J. W. Kozarich, and J. Stubbe, Biochemistry, 24, 7562 (1985). 50. J. C. Wu, J. W. Kozarich, and J. Stubbe, Biochemistry, 24, 7569 (1985). 51. C. W. Haidle, K. K. Weiss, and M. T. Kuo, Mol. Pharmacol., 8, 531 (1972). 52. M. T. Kuo and C. W. Haidle, Biochem. Biophys. Acta, 335, 109 (1974). 53. L. Rabow, G. H. McGall, J. Stubbe, and J. W. Kozarich, J. Am. Chem. Soc., 112, 3203 (1990). 54. L. Rabow, J. Stubbe, J. W. Kozarich, and J. A. Gerlt, J. Am. Chem. Soc., 108, 7130 (1986). 55. L. F. Povirk, W. Wübker, W. Köhnlein, and F. Hutchinson, Nucl. Acids Res., 4, 3573 (1977). 56. J. W. Sam and J. Peisach, Biochemistry, 32, 1488 (1993). 57. W. E. Antholine, D. H. Petering, L. A. Saryan, and C. E. Brown, Proc. Natl. Acad. Sci. USA, 78, 7517 (1981). 58. M. J. Absalon, W. Wu, J. W. Kozarich, and J. Stubbe, Biochemistry, 34, 2076 (1995). 59. L. Rodriguez and S. M. Hecht, Biochem. Biophys. Res. Commun., 104, 1470 (1982). 60. D. H. Petering, W. E. Antholine, and L. A. Saryan, in Anticancer and Interferon Agents. Drugs and the Pharmaceutical Sciences, Vol. 24 (R. M. Ottenbrite and G. B. Butler, eds.), Marcel Dekker, New York, 1984, p. 203. 61. A. L. Feig and S. J. Lippard, Chem. Rev., 94, 759 (1994). 62. A. Natrajan, S. M. Hecht, G. A. van der Marel, and J. H. van Boom, J. Am. Chem. Soc., 112, 3997 (1990). 63. R. X. Xu, W. E. Antholine, and D. H. Petering, J. Biol. Chem., 267, 944 (1992). 64. Y. Sugiura, J. Am. Chem. Soc., 102, 5216 (1980). 65. R. X. Xu, D. Nettesheim, J. D. Otvos, and D. H. Petering, Biochemistry, 33, 907 (1994). 66. C. A. G. Haasnoot, U. K. Pandit, C. Kruk, and C. W. Hilbers, J. Biomol. Struct. Dynam., 2, 449 (1984).
< previous page
page_646
next page >
< previous page
page_647
next page > Page 647
67. M. Nakamura and J. Peisach, J. Antibiot., 41, 638 (1988). 68. J. Templin, L. Berry, S. Lyman, R. W. Byrnes, W. E. Antholine, and D. H. Petering, Biochem. Pharmacol., 43, 615 (1992). 69. R. X. Xu, W. E. Antholine, and D. H. Petering, J. Biol. Chem., 267, 950 (1992). 70. R. M. Burger, J. Peisach, W. E. Blumberg, and S. B. Horwitz, J. Biol. Chem., 254, 10906 (1979). 71. M. Chikira, W. E. Antholine, and D. H. Petering, J. Biol. Chem., 264, 21478 (1989). 72. H. Shields, C. McGlumphy, and P. J. Hamrick, Jr., Biochem. Biophys. Acta, 697, 113 (1982). 73. H. Shields and C. McGlumphy, Biochem. Biophys. Acta, 800, 277 (1984). 74. M. Chikira, T. Sato, W. E. Antholine, and D. H. Petering, J. Biol. Chem., 266, 2859 (1991). 75. Q. Mao, P. Fulmer, E. DeRose, and D. H. Petering, submitted. 76. P. Fulmer and D. H. Petering, Biochemistry, 33, 5319 (1994). 77. M. R. Ciriolo, R. S. Magliozzo, and J. Peisach, J. Biol. Chem., 262, 6290 (1987). 78. W. Li, P. Fulmer, W. E. Antholine, and D. H. Petering, submitted. 79. D. L. Melnyk, S. B. Horwitz, and J. Peisach, Biochemistry, 20, 5327 (1981). 80. J. P. Albertini and A. Garnier-Suillerot, Biochemistry, 23, 47 (1984). 81. W. E. Antholine and D. H. Petering, Biochem. Biophys. Res. Commun., 91, 528 (1979). 82. M. J. Levy and S. M. Hecht, Biochemistry, 27, 2647 (1988). 83. W. Wu, D. E. Vanderwall, J. Stubbe, J. W. Kozarich, and D. J. Turner, J. Am. Chem. Soc., 116, 10843 (1994). 84. W. Saenger, Principles of Nucleic Acid Structure, Springer-Verlag, New York, 1984. 85. R. A. Manderville, J. F. Ellena, and S. M. Hecht, J. Am. Chem. Soc., 116, 10851 (1994). 86. S. A. Kane and S. M. Hecht, in Progress in Nucleic Acid Research and Molecular Biology, (W. E. Cohn and K. Moldave, eds.), Academic Press, San Diego, 1994, p. 313.
< previous page
page_647
next page >
< previous page
page_648
next page > Page 648
87. J. Kross, W. D. Henner, S. M. Hecht, and W. A. Haseltine, Biochemistry, 21, 4310 (1982). 88. L. F. Povirk, M. Hogan, M. Buechner, and N. Dattagupta, Biochemistry, 20, 665 (1981). 89. N. Hamamichi, A. Natrajan, and S. M. Hecht, J. Am. Chem. Soc., 114, 6278 (1992). 90. S. A. Kane, A. Natrajan, and S. M. Hecht, J. Biol. Chem., 269, 10899 (1994). 91. B. J. Carter, V. S. Murty, K. S. Reddy, S.-N. Wang, and S. M. Hecht, J. Biol. Chem., 265, 4193 (1990). 92. T. J. Keller and N. M. Oppenheimer, J. Biol. Chem., 262, 15144 (1987).
< previous page
page_648
next page >
< previous page
page_649
next page > Page 649
Subject Index A AAS, see Atomic absorption spectroscopy Absorption bands and spectra (see also Infrared spectroscopy and UV absorption), 4648, 181, 209, 211, 213215, 234, 239, 262, 264, 370, 378, 379, 384, 387, 401, 493, 494 polarized, 188190 transient, 190, 347, 348, 350360 Acetazolamide, 3437 Acidity constants (see also Equilibrium constants), 3234, 3639, 42, 44, 60, 62, 111, 112, 281, 284, 285, 571, 572 Acquired immunodeficiency syndrome, see AIDS Acridine(s), 97, 230, 440, 441 9-amino-, 97, 200 9-anilino-, 356 bis-, 202 9-chloro-, 414 hybrid, 441 nitro-, 328 orange, 254265, 329, 356, 386 residue, 4346 4'-(9-Acridinylamino)methanesulfon-m-anisidine, see Amsacrine Acridione, 417 Actinomycin, 220, 225, 227, 477 C, 204 D, 200, 239, 598 Activation parameters (see also Enthalpy and Entropy), 94
< previous page
page_649
next page >
< previous page
page_650
next page > Page 650
Active sites of (see also Coordination spheres) enzymes, see Enzymes Adenine(s) (and residues), 43, 114117, 123, 129, 148, 154, 226, 256, 280, 310, 311, 406, 461, 463, 477, 517, 529, 548, 599 9-methyl-, 64, 67, 119, 121, 127, 128 Adenosine, 311, 596 2'-deoxy-, 39, 40, 183, 184, 279, 280, 305 Adenosine monophosphate, see AMP Adenosine triphosphate, see ATP Adriamycin, 58, 620 Affinity constants, see Stability constants Aggregation DNA salt-induced, 385, 386 self-, see Self-aggregation AIDS, 368, 390 Alaninamide, 632 Albumin, 543 serum, 433, 444 Alcohol(s) (see also Hydroxyl groups and individual names), 410, 572 allyl, 108 Alcoholic groups (see also Alcohols, Hydroxyl groups, and Phenolates), 412 Aldehyde groups, 407412, 531 salicyl-, 519 Alkali metal ions (see also individual names), 150, 151, 157, 160 Alkaline earth metal ions (see also individual names), 157, 160, 167 Alkaline phosphatase, 408, 409, 411, 491 Alkoxide, 565, 571, 572, 580 Ames test, 275, 285 Amide(s) (see also individual names), 527, 575, 577, 635
alanin-, 632 azetazol-, 3437 deprotonated, 36, 433 hydrolysis, 570 nitrogen, 632 Amine(s) (see also Amino groups and individual names), 278, 519, 553555, 571, 572, 632, 633, 640 alkyl-, 584 butyl-, 553 cyclohexyl-, 109 heterocyclic, 89, 9698, 100 methyl-, see Methylamine poly-, see Polyamines tetra-, see Tetraamines tris(aminoethyl)-, 556 Amino acids (see also individual names), 125, 129 derivatized, 430432 Amino groups, 40, 65, 160, 226, 431, 433 deprotonation, 124 2-Amino-6-oxopurine, see Guanine Ammonia, 109, 112, 130, 568, 572 cobalt complexes, see Cobalt(III) platinum complexes, see Cisplatin and Transplatin ruthenium complexes, see Ruthenium(III) zinc complexes, see Zinc(II) Ammonium ion, 151 AMP 5'-d-, 63, 67, 68, 71, 204 Amsacrine, 329, 356 [12]aneN3, see 1,5,9-Triazacyclododecane [12]aneN4, see Cyclen [14]aneN4, see Cyclam
Aniline(s), 554 N,N-dimethyl-, 596 Anisole p-deutero-, 596 Anthracyclines, 201 Antibiotics (see also individual names), 406, 413, 477
< previous page
page_650
next page >
< previous page
page_651
next page > Page 651
[Antibiotics] enediyne, 408 glycoprotein, 594 Antibodies, 115 Anticancer drugs (see also individual names), 107, 128, 162, 400, 406, 416, 420, 446, 447, 609 cisplatin, see Cisplatin platinum, see Platinum anticancer drugs transplatin, see Transplatin Antigene strategy, 101, 161, 162, 168, 230, 415, 419, 420 Antisense strategy, 101, 161, 162, 168, 230, 415, 419 Antitumor activity (of), 271 metallocene complexes, 5760, 7477 platinum complexes, 17, 161 transplatin, see Transplatin Antiviral agents, 400, 420 Ascorbate (or ascorbic acid), 165, 403, 404, 406, 432, 433, 444, 456, 457, 491, 497, 622, 636 Association constants (see also Dissociation constants and Stability constants), 195, 275, 276, 285, 623 Atomic absorption spectroscopy, 273 ATP, 76, 487, 504 γ-32P, 77 Autoradiography, 538, 554 Azide(s), 57, 637 sodium, 402 Azidothymidine, 3943 B Bacillus subtilis, 599601, 604, 605, 609, 610 Bacteriophages 434 repressor, 469, 470 λ genome, 164, 469
λ repressor, 455, 469 T4, 224, 225 Barium ion, 152, 157 Base pairs or sequences, 5, 11, 38, 40, 46, 49, 92, 94, 95, 97, 99, 150, 163, 193, 194, 204, 216, 220, 228, 231, 232, 240, 255, 261, 263266, 543, 638640 AT, 126, 146, 160, 187, 191, 203, 225227, 257, 264266, 311, 373, 375, 378380, 407, 408, 410, 412, 413, 416, 419, 445, 465, 467, 470, 522, 525, 527, 528, 532, 534 AU, 602 CA, 604 CC, 604 CG, 256, 410 GA, 604 GC, 17, 115, 146, 187, 194, 225227, 229, 256, 264266, 311, 373, 375, 378, 379, 407, 408, 410, 416, 465, 467, 527, 528, 541, 542, 548, 551, 596, 598 GG, 17, 115, 542, 551, 604 GT, 596, 598 GU, 551, 599, 601, 604 Hoogsteen, see Hoogsteen metal-modified, 125, 126 mismatch, 541, 542, 551 TA, 470 TG, 225 UA, 604 UC, 599 UG, 604 UU, 599, 604 Watson-Crick, see Watson-Crick Benzene(s) 1,2-diamino-, see 1,2-Diaminobenzene 1,2,4-trimethoxy-, 274 iodosyl-, see Iodosylbenzene Bicarbonate, see Hydrogen carbonate
< previous page
page_651
next page >
< previous page
page_652
next page > Page 652
Binding constants (see also Association constants, Dissociation constants, Equilibrium constants, and Stability constants), 194198, 208221, 300, 301, 306, 308, 309, 314, 316, 346, 354 apparent, 188, 193, 194 salt effects, 195, 196 2,2'-Bipyrazine, 183, 230 ruthenium complexes, see Ruthenium(II) 2,2'-Bipyridyl, 57, 166, 201, 204, 209, 210, 212, 213, 215, 217, 218, 220, 221, 231, 233, 274, 275, 283, 311, 312, 316, 318, 337, 356 copper complexes, see Copper(II) 4,4'-dimethyl-, 318 diphenyl-, 336 iron complexes, see Iron(II) osmium complexes, see Osmium(II) ruthenium complexes, see Ruthenium(II) structures, 182 titanium(IV) complexes, 57 Bismuth(III), 565 Bleomycin, 406, 413, 414, 416, 420, 446448 cadmium, see Cadmium ion chemotherapeutical role, 594, 595 cobalt, see Cobalt(II) and Cobalt(III) copper, see Copper(II) DNA interaction, 619642 iron- (see also Iron(II) and Iron(III)), 309, 310, 456, 531534, 593610 metallo-, 518, 619642 model, 598 nickel, 543, 544, 555557 O=Fe(IV), 628, 629 O=Fe(V), 628, 629 structures, 597, 621
zinc, see Zinc(II) Bond angle energy, 14, 15 calculated lengths, 14 distances, 40, 44, 108, 125 experimental lengths, 14 ''Borderline" species, see "Hard and soft" species Bromide as ligand, 57, 58 Buffer(s), 201, 203, 339, 351, 432, 489, 565, 603 cacodylate, 210, 211, 214, 437 oxalate, 210, 211 phosphate, 210213, 215, 223, 226, 275, 284, 313, 317, 637 sucrose, 233, 236 Tris, 48, 210216, 223, 353, 355 Butylamine, 553 C Cacodylate buffer, see Buffer Cadmium ion, 154, 254, 620, 633 113Cd-NMR, see NMR bleomycin, 633, 640 Calcium ion, 2, 76, 150152, 154, 157160, 487, 566 homeostasis, 506 phosphate, 461, 473 Calf thymus DNA, see DNA Calicheamicin, 477 Calorimetry, 197, 198 Cancer (see also Tumors) bladder, 401 ovarian, 88 testicular, 88 Cancerogenicity, 254
Carbamezapine, 411 Carbon 13C-NMR, see NMR radical, see Radicals Carbonic anhydrase, 36, 37 B, 234, 235 Carbonium ion, 495 Carbonyl groups (or carbonyl oxygen), 40, 43, 44, 46
< previous page
page_652
next page >
< previous page
page_653
next page > Page 653
Carboxylates (or carboxylic acids) (see also individual names), 57, 58, 401, 411, 444, 576, 577 Carcinogenesis, 328, 539 Carcinomas (see also Cancer and Tumor), 594 colon 38 adeno-, 57 human colon, 57 human gastrointestinal, 57 Lewis lung, 57 murine mammary adeno-, 57 Carcinostatic activity of cisplatin, 58 metallocenes, see Metallocenes Catalase, 414, 491, 497 Catalysis by metal ions, 561585 Catalytic sites, see Active sites CD, see Circular dichroism Cells chloride concentration, 271 death, 506, 509, 517, 594 growth, 594, 595 mouse fibroblast, 402 TA3Ha, 61 Cesium ion, 151, 156 Charge transfer, 304, 305 ligand-to-metal, 337 metal-to-ligand, 187, 233, 234, 237, 329, 333, 338, 348, 350 Chemotherapy (see also Drugs), 59, 77, 88, 270 photo-, 269291 Chirality of complexes (see also Enantioselectivity), 106, 107, 179, 204, 205, 232, 238, 329, 332, 517, 518 Chlorambucil, 417
Chloride as ligand, 57, 58, 60, 61, 89, 90, 93, 94, 110, 114, 125, 131, 270, 271, 273291, 494, 495 platinum chloro complexes, see Cisplatin and Transplatin Chromatography, 207, 288 affinity, 192 high-performance liquid, see High-performance liquid chromatography Sephadex G-25, 286, 287, 289 size exclusion, 272, 286 thin layer, see Thin layer chromatography Chromium (oxidation state undefined), 168, 254 Chromium(II), 290 tris(1,10-phenanthroline) complex, 184 Chromium(III), 284290 bis(1,10-phenanthroline)(L) complexes, 284290 tris(1,10-phenanthroline) complex, 231, 285, 331, 332 Chromium(VI), 56 Chromomycin, 220, 518 A3, 229 Chromosomes, 155, 168, 473, 543, 609 Circular dichroism, 20, 115, 188, 189, 205207, 229, 233, 238, 239, 373, 378, 383, 385, 386, 474 induced, 374, 375, 382, 386, 387, 389 magnetic, see Magnetic circular dichroism Cisplatin (see also cis-Diammineplatinum(II)), 2, 17, 19, 42, 54, 5763, 71, 75, 78, 8892, 96101, 105132, 161, 162, 201, 270, 271, 274, 276, 517, 620 antitumor activity, 161 hydrolysis, 90 photoeffective, 271273, 275, 290 Clostridium perfringens, 226 CMP 5'-d-, 63, 67, 68, 119 Cobalt(II), 12, 154, 254, 255, 262265, 404, 542, 580, 620, 634 bleomycin, 413, 629, 630, 634636, 638, 641
< previous page
page_653
next page >
< previous page
page_654
next page > Page 654
[Cobalt(II)] tris(2,2'-bipyridyl) complex, 210, 299, 300 tris(1,10-phenanthroline) complex, 211, 217, 299, 300 Cobalt(III), 12, 158160, 164, 167, 444, 568, 569, 572, 620, 640 bleomycin, 630635, 638640 hexaamine complex, 158160, 164, 167, 204, 208, 228, 229, 385, 387, 388 porphyrins, 371, 378, 379, 383, 404, 406 tris(2,2'-bipyridyl) complex, 210, 231, 299, 300 tris(1,10-phenanthroline) complex, 184, 212, 217, 231, 299, 300, 331 Colon 38 adenocarcinoma, 57 Colorimetry (see also Absorption bands and spectra and Spectrophotometry), 273 Complexes inner-sphere, see Inner-sphere mixed-ligand, see Ternary complexes outer-sphere, see Outer-sphere ternary, see Ternary complexes Conductivity measurements, 202, 203 Conformation(al) anti, 91, 118, 148 change, 97 syn, 91, 118, 148 Cooperative binding effects, 193, 194, 226 anti-, 256 Coordination spheres (or geometry) (see also Active sites), 12, 16, 22, 23, 30, 31, 55, 123, 124, 126, 131, 148, 205, 224227, 233235, 238241, 379, 576, 640 Copper(I) DNA interaction, 165, 199, 200 (1,10-phenanthroline) complexes, 165, 225, 226, 230, 408, 413, 470, 485509
Copper(II), 123, 124, 154, 161, 254256, 261265, 430, 433, 434, 436, 439441, 444, 446, 489, 542, 565, 569572, 577, 578, 580, 582584, 598, 620, 622 bis(2,2'-bipyridyl) complex, 185 bis(1,10-phenanthroline complex, 165, 185 bleomycin, 598, 624, 638 porphyrins, 371, 372, 374, 377381, 383, 385, 386, 389, 402 Cotton effects negative, 276 Crithidia fasciculata, 461 Crosslinking, 114, 230, 417, 418 DNA-DNA, 539 DNA-protein, 115, 129, 539 interstrand, 19, 8896, 98, 100, 101, 115, 118, 160, 161, 291 intrastrand, 1719, 8890, 9296, 100, 101, 161, 162, 517 1,2-intrastrand, 131 1,3-intrastrand, 91, 9396, 98, 116118 1,4-intrastrand, 93 Cryoscopy, 67 Crystal structures, see X-ray crystal structures Crystal violet, 187, 201 5'-CTP, 505 Cyanide as ligand, 89, 275 Cyclam, 31, 32, 540, 556, 577 hexamethyl-, 556 Cyclen, 31, 32, 3843, 4649, 577 Cyclic phosphate ester, 563, 566, 583 Cyclic voltammetry (see also Voltammetry), 191, 192, 210212, 215, 300305, 307, 309, 312318 staircase, 303305 Cyclins, 609
< previous page
page_654
next page >
< previous page
page_655
next page > Page 655
Cyclohexylamine, 109 Cyclopentadienyl compounds, 5378 hafnium, 56, 60, 74 iron, 54 molybdenum, 56, 57, 5978 niobium, 57, 74 rhenium, 57 titanium, 5762, 69, 74, 75 vanadium, 57, 6063, 7478 zirconium, 56, 5961, 74 Cysteamine, 162 Cysteine (and residues), 430, 433, 472, 473 Cytidine, 63, 311, 606608 2'-deoxy-, 39, 40 Cytidine monophosphate, see CMP Cytidine 5'-triphosphate, see 5'-CTP Cytochrome b5 reductase, 625 Cytochrome c electron transfer, 326, 357 Cytochrome P450, 404, 405, 411, 412 reductase, 625 Cytoplasm, 402, 609 Cytosine (and moiety), 18, 117, 118, 129, 146, 166, 256, 310, 529, 544, 546, 548, 596, 608, 609 1,5-dimethyl-, 119, 123 1-methyl-, see 1-Methylcytosine Cytotoxicity (see also Toxicity), 128, 414, 416, 420, 506, 623, 624 cisplatin, 88, 107, 113 metallocenes, 55, 61 transplatin, 88, 107, 113
D Daunomycin, 197, 198, 227, 328 cis-DDP, see Cisplatin trans-DDP, see Transplatin Densitometry, 457, 466, 526, 532, 533 Deoxyribonucleic acid, see DNA Deoxyribonuclease, see DNase Deoxyribose, 165, 405408, 456, 460, 486, 488, 491, 500, 519, 529, 530, 606, 621 Deprotonation constants, see Acidity constants Depurination, 254, 517, 553, 554 Deuterium, 505 Dextran, 417 Dialysis, 188, 189, 206, 207, 210215, 226, 273, 286289 competition, 227, 229 equilibrium, see Equilibrium dialysis 1,2-Diaminobenzene, 291 1,2-Diaminoethane, see Ethylenediamine 2,6-Diaminohexanoate, see Lysine 2,5-Diaminopentanoate, see Ornithine cis-Diamminedichloroplatinum(II), see Cisplatin trans-Diamminedichloroplatinum(II), see Transplatin cis-Diammineplatinum(II) (see also Cisplatin), 1822, 58, 62, 63, 75, 76 Diastereomeric complexes, 232-240 Dibucaine, 594 Dichroism circular, see Circular dichroism electric linear, see Electric linear dichroism flow linear, see Flow linear dichroism linear, see linear dichroism magnetic circular, see Magnetic circular dichroism Dien, see Diethylenetriamine
Diethylenetriamine and derivatives, 115, 119, 125, 162 Diethylenetriamine-N,N,N',N'',N"-pentaacetate, 478, 576 Diethylpyrocarbonate, 548 Diffusion constant rotational, 198
< previous page
page_655
next page >
< previous page
page_656
next page > Page 656
Dihydronicotinamide adenine dinucleotide, see NADH Dihydronicotinamide adenine dinucleotide phosphate, see NADPH Dimethylamine, 109 Dimethylarsinic acid, see Cacodylate N,N'-Dimethyl-2, 7-diazapyrenium, 329, 339, 341 Dimethylformamide, 109, 112, 437 Dimethyl sulfate, 503, 538, 539 Dimethylsulfoxide, 109, 553 Dinucleotides (and units thereof), 113, 119, 228, 272, 562, 565567, 571573, 576, 578, 580 ApA, 504 CpG, 204, 378, 379 d(ApG), 88, 90, 91, 96 d(GpA), 88 d(GpC), 88, 92 d(GpG), 20, 88, 90, 91, 96 r(GpG), 2022 UpA, 573 Diol (see also Hydroxyl groups, Sugars, and individual names) 5,6-cis-, 166 Dioxygen, 223, 288, 309, 310, 403406, 413, 414, 416, 430, 432, 444, 489, 528, 529, 532, 595, 596, 606, 620622, 625, 627, 629631, 633638 17O2, 622 radical, see Radicals Dipeptides (see also Peptides and individual names), 125, 129 Diphosphate, see Pyrophosphate Dipole-dipole interactions, 264 Dipyrido(3,2-a:2',3'-c)phenazine, 181, 183, 335, 343 osmium complexes, see Osmium(II) ruthenium complexes, see Ruthenium(II) N,N'-Disalicylidene-ethylenediamine, see Salen
Dismutase superoxide, see Superoxide dismutase Dissociation constants (see also Association constants, Equilibrium constants, and Stability constants), 195, 338, 354 Distamycin, 439, 440, 446, 477, 527 A, 596 Dithionite sodium, 414 Dithiothreitol, 404, 414, 432, 489 DNA (see also Nucleic acids), 11, 42, 46, 113, 119, 582, 583, 606 A-, 91, 149, 216, 360, 361, 498, 503 affinity cleavage, 434, 439 aptamers, 154156 artificial nucleases (see also Nucleases, artificial), 399420, 427448 atmospheric interactions, 150 B-, 49, 91, 117, 149, 159, 166, 179, 195, 202, 205208, 216, 220, 228, 291, 327, 329, 331, 334, 355, 359361, 374, 378, 408, 457, 463, 493, 498503, 505, 517, 528, 541, 542, 546, 639, 642 B to Z transition, 228, 539 backbone (see also Sugar, phosphate backbone), 43, 71, 91, 116, 157, 164, 165, 167, 228, 354, 358, 388, 405, 412, 441, 444, 468, 476, 479, 492, 621, 629, 635, 638, 642 bacterial, 78 bent, 92, 118, 158, 159, 203, 461464 binding domains of proteins, 429439 calf thymus, 71, 184, 186, 194, 205, 208, 210215, 222, 223, 226, 227, 237, 275, 286, 301303, 305, 307, 311, 313, 316318, 338, 339, 341, 353, 355, 388, 402, 404, 406, 407, 529
< previous page
page_656
next page >
< previous page
page_657
next page > Page 657
[DNA] catalytic activity, 87101 chromosomal, 595, 624 circular, 200, 203, 228, 314, 373, 402 cleavage (see also strand scission), 97, 98, 100, 163, 164, 166, 167, 225, 229, 273, 309314, 336338, 342344, 351, 354, 355, 390, 399421, 427448, 460463, 477, 486, 487, 489, 492, 494, 496, 497, 503505, 515534, 546, 595598, 604, 627, 638, 641, 642 clips, 255, 256 Clostridium perfringens, 226 crosslinks, see Crosslinks cruciform, 91, 229 damage, 190, 254, 271, 285, 298, 360, 390, 417, 444, 594, 619642 denatured, 276, 277 depurination, 254, 517 diffusion, 301306, 308, 312 double-stranded (see also duplex), 38, 43, 9396, 98, 114, 197, 203, 205, 221, 230, 412, 414, 416, 419, 420, 457, 460, 461, 473, 499, 540, 627 drug interactions, 190, 191, 196203, 367391, 429, 477, 478 duplex (see also double-stranded), 92, 95, 98, 99, 101, 115, 145, 151, 157, 161, 162, 164, 168, 328, 330, 336, 342, 345, 346, 361, 371382, 384, 386, 403, 413, 432, 441, 461, 466, 500, 539, 541, 542, 596, 606 dye interactions (see also Dyes), 189, 191, 253260 electrochemistry, 297319 electron transfer, 180, 187, 190, 231, 232, 256258, 260, 262, 265, 325361 enantioselective complex binding, list of, 206, 207 fibers, 446, 634 footprinting, 165, 336, 378, 379, 383385, 446, 455458, 467469, 472478, 495, 497, 503, 506, 507 four-way junctions (see also Holliday junction), 158160, 166168, 383385, 464466 groove binding (see also major groove and minor groove), 179, 187, 189191, 199201, 203, 220225, 275, 276, 328, 338341, 373, 375, 376, 378382, 416, 439441, 446, 517, 518 H-, see intramolecular triplex hairpin, 155, 499, 500, 541543, 598 Holliday junction (see also four-way junction), 158160, 164, 464, 465, 467, 503, 509
homeodomains, 470, 471 hybrid, 95, 97, 99 hydrolysis, 564, 570 in the nucleosome, 473475 intercalators, see Intercalation intramolecular triplex, 145149, 152, 154, 166 J-, 166 junctions, 464468 kinking, 1719, 117, 131, 200, 201, 203, 378 left-handed, 22, 49 linear, 203, 402 M. lysodeikticus, 226 major groove, 17, 75, 92, 92, 145, 150, 164, 167, 191, 216, 220, 224, 225, 227229, 284, 330, 333, 335, 336, 340, 354, 379, 380, 406, 434, 441, 442, 470, 471, 473, 475, 498, 499, 503, 517, 518, 539, 541, 598 melting temperature, 373 mesojunction, 165 metal ion interactions, 150154, 177240
< previous page
page_657
next page >
< previous page
page_658
next page > Page 658
[DNA] metallated, 272, 273, 283, 290, 596 metallobleomycin interactions, 619642 metallocene interactions, 5378 minor groove, 117, 160, 165, 187, 191, 198, 201, 204, 216, 220, 225, 228, 229, 235, 239, 283, 329, 330, 378, 379, 388, 402, 405408, 410, 416, 434, 446, 463, 468, 470, 471, 474, 477, 498503, 509, 517, 518, 525, 527529, 532, 534, 596, 610, 638641 multiple stranded, 143168 neutron bombardment, 328 nicked, 314, 404, 519, 520 nodular, 166, 167 oxidative cleavage, 273, 403, 408, 529531 Oxytricha, 155 photocleavage, 167, 336338, 342344, 351, 354, 355, 390, 403, 441, 638 photolytic covalent binding of metal complexes, 269291 plasmid, 71, 78, 152, 167, 310, 314, 402, 405, 519, 520, 522, 526, 532, 533, 624, 625 porphyrin interactions, 367391, 399421 probes, 177240 processing enzymes, 5378 protein interactions, 427448, 455, 468477 quadruplex, 144, 151, 154158, 168 radiation damage, 298 radiolabeled, 411, 419, 432, 445, 457459, 532, 533, 538 recognition, see Molecular recognition repair, 107, 122, 131 replication, 2, 58, 76, 77, 88, 113, 131, 145, 271, 272, 620 right-handed, 22, 49, 353 -RNA hybrids, see RNA rod-like, 199, 201, 202 salmon testis, 74
salt-induced aggregates, 385, 386 selective photodamage, 390 single-stranded, 92, 9599, 114, 161, 162, 166, 203, 221, 223, 230, 343, 402, 403, 417419, 460, 473, 498, 499, 502, 505, 509, 520, 540, 541, 596, 627 stability of complexes, 254, 255 strand scission (see also cleavage), 336, 456, 495, 497, 500, 519, 538540, 544546, 553, 555, 594, 595, 621624, 628, 629, 634, 636 structure, 158, 327, 330, 337, 340, 345, 361, 376, 432, 434, 435, 443, 461468, 498, 540547 supercoiled, 146, 148, 152, 166, 200202, 221, 314, 373, 402, 404, 414, 416, 444, 477, 519, 520 superstructures, 158 surface binding, 378, 386, 388, 446 synthesis, see Synthesis t-, 599, 605 T4-, 224, 225 tDNAHis, 601, 605 telomeres, 154157, 168, 543 thermostability, 131, 256 three-way junction, 383385, 402, 403, 464 transcription, 76, 88, 161, 468, 475, 476, 503509 transition metal ion interactions, 177240 triple-helical, 333 triplex, 91, 101, 126, 144154, 161, 163168, 189 unwinding, 92, 118, 150, 200203, 221, 224, 225, 329, 373, 598
< previous page
page_658
next page >
< previous page
page_659
next page > Page 659
[DNA] Z-, 49, 91, 184, 189, 204208, 216, 227, 228, 241, 360, 361, 378, 498, 499, 542 Z to B transition, 206, 208, 228, 229, 378 zinc fingers, see Zinc fingers DNA polymerases, 2, 7578, 554, 555 1,10-phenanthroline inhibition, 489491 E. coli, 489 I, 487, 488 mammalian, 487 viral, 487 DNase(s), 487491, 506, 507 I, 97, 455, 474 DPNH, see NADH Drugs (see also individual names), 62, 78, 91, 101, 195, 220, 239, 271, 328, 368, 373, 377, 379, 381, 383385, 387, 389, 406, 411, 444, 477, 478, 594, 609, 623, 625, 635638, 641 antitumor, 328, 329, 406 DNA interactions, see DNA metalloporphyrin-based, 420 photosensitizing, 271 resistance, 76, 77 DTPA, see Diethylenetriamine-N,N,N',N'',N"-pentaacetate Dyes (see also individual names), 181, 187, 190192, 202, 204, 210215, 220, 221, 224, 228, 230, 291 fluorescence, 253266 Dysprosium(III), 575, 578 E E. coli, see Escherichia coli EDTA, see Ethylenediamine-N,N,N',N'-tetraacetate Ehrlich ascites tumor, 5559
cells, 622 Electric linear dichroism, 185, 189, 224 Electrochemical studies, 217, 297319, 404 Electrode(s) graphite, 310 material, 298 mercury, 310 surface, 310, 311 tin-doped indium oxide, 301, 304 Electrode potentials, see Redox potentials Electron density, 44, 381, 385, 571 Electronic spectra, see Absorption bands and spectra Electron microscopy, 115, 198 Electron nuclear double resonance, see ENDOR Electron paramagnetic resonance, see EPR Electron spin resonance, see EPR Electron transfer, 254, 274, 281, 282, 290, 304, 305, 317, 318, 402, 403, 417, 529531, 553, 636 direct spectral evidence for, 347349 Förster-type, see Förster mechanism long-range, 231, 232, 332, 390 photoinduced, 231, 232, 248, 328361 quenching, 331, 332, 347, 360 radiationless electron excitation, 256258, 260, 262, 265 rate, 331, 359 through DNA, see DNA through proteins, 326 Electrophoresis, 75, 78, 159, 186, 200203, 505, 529 gel, see Gel electrophoresis Electrostatic interactions (see also Territorial binding), 5, 7, 911, 1416, 38, 179, 197, 217, 218, 220, 221, 224, 272, 275, 276, 299, 300, 309, 315, 316, 338, 373, 379, 383, 410, 446, 493, 503, 517, 568
< previous page
page_659
next page >
< previous page
page_660
next page > Page 660
[Electrostatic interactions] long-range, 11, 23 Electrostatic potential, 91, 180, 402, 498 Ellipticine, 414, 416 Emission anisotropy, 188190, 233, 235237, 239 lifetimes, 222, 223 Emission spectroscopy, 181, 188190, 209, 211, 213215, 344, 346, 355 Enantioselectivity, 191, 205208, 220, 224229, 231233, 235239, 241, 330, 331, 334, 337, 353, 354, 517, 518, 524, 525 Endonucleases, 7578, 506, 546, 550 VII, 468 restriction, see Restriction enzymes ENDOR spectroscopy, 281, 377 Energy free, see Gibbs free energy transfer, 280, 282, 291, 377, 390 Enthalpy (see also Activation parameters), 94, 196198 Entropy (see also Activation parameters), 7, 93, 94, 196198 Enzymes (see also individual names) active sites, 30, 36, 478, 487, 509 DNA processing, see DNA excision repair, 473 resolving, 467, 468 restriction, see Restriction enzymes Epoxidation catalyst for, 518, 519 olefin, 518520, 524, 527 EPR, 281, 377, 446, 595, 621, 622, 630, 634, 637 Equilibrium constants (see also Acidity constants, Binding constants, and Stability constants), 93, 110, 284, 381, 383
Equilibrium dialysis (see also Dialysis), 192, 216, 275 Escherichia coli, 113, 166, 167, 475, 476, 487489, 503, 504, 551, 599601, 604, 610 ESR, see EPR Ethanol (see also Alcohol), 199, 200, 207, 286, 333 Ethidium bromide, 96, 187, 197202, 221, 230, 231, 254256, 258265, 329, 339, 341, 356358, 377, 383, 500 photodestruction, 254 N,N'-Ethylenebis(salicylidene-aminato), see Salen Ethylenediamine, 31, 32, 184, 186, 196, 291 Ethylenediamine-N,N,N',N'-tetraacetate, 158, 163165, 413, 430434, 439, 494, 576 derivatives, 430, 532534 S-(2-pyridylthio)cysteaminyl-, 430, 433 tricyclohexyl ester, 430, 431 N-Ethylmaleimide, 625 Euglena gracilis, 624 Eukaryotes (see also individual names), 78, 113, 473, 503, 504, 582, 584 Europium(III) (see also Lanthanides), 570, 574580, 582, 583 EXAFS, 622, 629 Exonuxleases 3'-, 488 3',5'-, 489 III, 455, 491 Extended absorption fine structure spectroscopy, see EXAFS F FAB-MS, see Fast atom bombardment mass spectrometry
< previous page
page_660
next page >
< previous page
page_661
next page > Page 661
Fast atom bombardment mass spectrometry, 65, 279 Fenton reactions, 164, 455460, 493, 555 Ferritin, 601, 602, 610 Ferrocyanide, see Hexacyanoferrate(II) Filtration methods, 192, 207, 210 Flow linear dichroism, 185, 189, 224, 225, 375 Fluorescein, 230 Fluorescence, 115, 187, 207, 230, 309, 368, 371, 375, 402, 416 decay, 356 excitation, 258265 of dyes, 253266 quenching, 257265, 280, 281, 288, 329, 356, 377 resonance energy transfer, 230 spectroscopy, 633, 637 Fluoride as ligand, 57, 58 Fluorometry, 273 Fluorophores (see also individual names), 187, 188, 230 Footprinting studies of DNA, see DNA Force field, 19, 20, 22, 23 calculations (see also Molecular mechanics), 1216 Förster mechanism, 256, 257, 260, 336 Formation constants, see Binding constants, Equilibrium constants and Stability constants Frenkel's adsorption model, 255 Ftorafur, 39, 40, 4244 Furanone 5-methylene-2-, see 5-Methylene-2-furanone Furfural, 406408, 411, 531 G Gadolinium(III) (see also Lanthanides), 578
Gel electrophoresis, 18, 19, 200203, 343, 405, 529, 538 agarose, 203 high-resolution, 566 polyacrylamide, 203, 411, 454, 457, 458, 460, 554 Gel mobility studies, 118, 200, 464 Gene artificial repressors, 168 expression, 49, 145, 168, 404, 550, 582 human U1, 166 immunoglobulin, 156 sea urchin histone, 166 Gibbs' free energy, 196, 217, 381 Glutamic acid, 414 γ-L-Glutamyl-L-cysteinylglycine, see Glutathione Glutathione, 100, 114, 131, 627, 636 Glycerol, 231, 332, 374 Glycine (and residues), 436438, 597, 598, 641 Glycolate, 540 phospho-, 491, 493, 505, 555, 606, 627 Glycosidic bond (or conformation), 17, 20, 69, 91, 155 Glycylglycylhistidine, 433436, 438440, 444448, 543545 Glyoxalate, 498 5'-GMP, 478 Gold(III) porphyrin complexes, 371, 374, 386 GTP, 504 Guanine (and moiety), 42, 69, 114119, 123, 129, 148, 150, 151, 154, 160, 162, 228, 256, 280, 310, 311, 314, 318, 402, 403, 418, 517, 527, 529, 546, 548, 553557, 641 7,9-dimethyl-, 119, 121 9-ethyl-, 119122, 183, 184, 204 9-methyl-, 119, 121, 126, 129, 163 8-oxo-, 314
probes for, 537-557 quartets, 155158, 160, 544
< previous page
page_661
next page >
< previous page
page_662
next page > Page 662
[Guanine (and moiety)] radical, see Radicals Guanosine, 63, 71, 276, 281, 283, 284, 288, 312, 478, 548 2'-deoxy-, 39, 40, 46, 183, 184, 204, 277281, 288 deoxyguanosine radical, see Radicals 8-hydroxydeoxy-, 281, 283 Guanosine 5'-monophosphate, see 5'-GMP Guanosine triphosphate, see GTP Gyrases, 476, 477 H Hafnium(IV), 56, 57, 60, 69, 74 Halides (see also individual names), 274 Halogens (see also individual elements), 109, 129, 130 ''Hard and soft" species, 36, 60, 63, 89, 574, 576 Heavy metals, see Transition metal ions and individual elements Helix-turn-helix motif, 429, 430, 432, 435, 468471 Hemin, 404, 418 Hemoglobin, 12, 635 Hepatotoxicity, 57 Heteronuclear complexes, 123125 Hexacyanoferrate, 347 High-mobility group proteins, 474, 475 High-performance liquid chromatography, 108, 116, 276, 277, 407, 566, 583, 606 Hin recombinase, 430, 432, 434, 435, 446, 544, 545 Histidine (and residues), 433, 436438, 472, 544, 632 glycylglycyl-, see Glycylglycylhistidine Histones, 166, 473, 474, 477 HIV, 155, 390, 416, 418, 501, 506, 603 -1 reverse transcriptase, 476, 601, 602, 604, 608610
Holliday junctions in DNA, see DNA Homeostasis of calcium, 506 Hoogsteen base pairs, 126, 127, 146, 148, 163 hydrogen bonds, 145, 148, 150, 154, 155 modified ligands, 121 reverse base pairs, 148 HPLC, see High-performance liquid chromatography Human immunodeficiency virus, see HIV Hydrogen bonds, 4, 5, 20, 22, 38, 40, 43, 44, 46, 91, 115118, 125128, 156, 159, 160, 162, 180, 204, 209, 276, 336, 338, 378, 379, 381, 517, 518, 524, 527, 542, 548, 567, 568 Hoogsteen, see Hoogsteen Hydrogen carbonate, 37, 494 Hydrogen peroxide (see also Peroxides), 164, 165, 404, 405, 408, 414, 417, 418, 433, 444, 456, 457, 479, 486509, 520, 529, 556, 621623, 629 reactivity, 493, 494 Hydrogen transfer, 311, 312 Hydrolysis of amide, see Amides cisplatin, see Cisplatin DNA, see DNA metallocenes, see Metallocenes phosphodiester, see Phosphodiester RNA, see RNA transplatin, see Transplatin Hydrophobic interactions, 38, 180, 205, 209, 216, 217, 221, 228, 272, 276, 300, 329, 527 pocket, 543 Hydroquinones, 328 Hydroxide(s), 564, 565, 571573
< previous page
page_662
next page >
< previous page
page_663
next page > Page 663
[Hydroxide(s)] ion, 488 Hydroxo complexes, 30, 33, 34, 111, 112, 570 2-Hydroxyethanethiolate, 182, 195, 196, 204, 227 Hydroxyl groups (see also Alcohol, Diol, Phenolates, and Sugars), 401, 414, 488, 529, 531, 563, 567, 571, 572, 581, 627 radicals, see Radicals 2-Hydroxypropyl-4-nitrophenylphosphate, 566569, 572, 578 Hydroxyquinoline, 567 Hyperchromism, 187 Hypochromism, 181, 187, 233, 234, 329 Hypoxanthine (and moiety) 7,9-dimethyl-, 120, 121 9-methyl, 183, 184 quartet, 161 I Ihremycin, 200 Imaging agents, 270 Imidazole (and moieties) (see also Histidine), 357, 359, 544, 571573, 631, 632, 640 -like nitrogen, 433 Iminodiacetate, 582, 583 Immunoglobulins, 156 Indium oxide electrode tin-doped, 301 Indole 4',6-diamidino-2-phenyl-, 187, 191, 198 Inductively coupled plasma spectroscopy, 74 Infrared spectroscopy, 108, 149, 190, 262, 377, 519 Inner-sphere coordination, 150, 312, 314
Inosine, 39, 40, 42, 44 Intercalation or intercalators (see also Stacking), 87101, 179, 187, 189191, 195, 196, 198205, 209, 216218, 220225, 227232, 235, 238240, 254265, 272, 300302, 328361, 368, 369, 371, 373, 375386, 388, 390, 400, 402, 405, 413417, 477, 500, 517, 518, 598, 638641 Introns, 229 Iodide as ligand, 57, 58 Iodosylbenzene, 403, 404, 406, 518, 520 Ionic radii, 578 Ionization constants, see Acidity constants and Equilibrium constants IR, see Infrared spectroscopy Irehdiamine, 201 Iridium(IV), 543 Iron(II), 12, 314, 447, 456, 504, 509, 593611, 620, 625, 629, 633 bleomycin (see also Bleomycin), 310, 406, 593610, 621623, 631, 632, 634638 EDTA, 163165, 230, 413, 430, 432, 434, 439, 456, 457, 478, 479, 493, 502, 532, 533, 534, 641 tris(2,2'-bipyridyl) complex, 179, 185, 205, 206, 210, 212 tris(1,10-phenanthroline) complex, 206, 208, 211, 229, 299 Iron(III), 54, 309, 456, 620 bleomycin (see also Bleomycin), 310, 595, 596, 621629, 632634, 636, 637, 639, 642 low-spin, 622 porphyrins, 371, 373, 378, 379, 401, 403406, 414418 tris(2,2'-bipyridyl) complex, 210, 318 tris(1,10-phenanthroline) complex, 299
< previous page
page_663
next page >
< previous page
page_664
next page > Page 664
Iron regulatory element, 601, 602, 610 Irradiation, 271, 272, 279, 281, 333, 342, 343, 402 UV, 167, 284, 336 K Kethoxal, 548 Ketones (see also individual names), 209, 627 Keto groups, see Carbonyl groups Kinases, 7578 Kinetic studies (see also Rate constants and Rates of reactions), 110112, 114, 195, 379382, 579, 583 Stern-Volmer, see Stern-Volmer plot L Lac repressor, 430, 455 Lactone derivatives, 531 Lanthanides(III), 565, 572, 577 macrocyclic complexes, 573578 Lanthanum(III) (see also Lanthanides), 565, 567, 571, 575584 Laser excitation, 380 induced luminescence spectroscopy, 578 Lead(II), 254, 562, 565, 567, 571, 572, 580 Leucine zipper motif, 430, 443 Leukemia L1210, 113, 414, 416 P388, 58, 113 Ligases, 412 Light scattering, 307 dynamic, 303 resonance, see Resonance light scattering Linear dichroism, 188190, 232235, 237241
electric, see Electric linear dichroism flow, see Flow linear dichroism Lithium ion, 151, 156 Luminescence, 108, 220, 230, 234, 274, 329, 333, 334, 338, 342, 343, 348, 377 chemi-, 210, 211 laser-induced spectroscopy, 578 quenching, 187, 221224, 231, 239, 333, 335, 338341, 344, 346355 Luthetium(III) (see also Lanthanides), 576579 Lymphomas (see also Tumors) malignant, 594 Lyophilization, 286, 288 Lysine (and residues), 431, 446 M M. lysodeikticus, 226 Macrochelates, 19 Macroconstants, see Acidity constants and Stability constants Macrocycles (see also individual names), 128, 540, 541, 543, 544, 546552, 555557, 567, 569, 571583 polyamines, see Polyamines Magnesium ion, 2, 34, 150154, 157159, 167, 201, 202, 221, 227, 403, 404, 464, 478, 487, 518, 520, 526, 540, 544, 546548, 566, 584, 602, 603, 610 Magnetic circular dichroism, 108 Magnetic resonance electron para-, see EPR nuclear, see NMR Magnetic susceptibility measurements, 519
< previous page
page_664
next page >
< previous page
page_665
next page > Page 665
Manganese(II), 2, 154, 262264, 266, 309, 565, 566, 581, 620 Manganese(III) porphyrins, 371, 378380, 383, 385, 400, 401, 403413, 415, 416, 418420, 532, 534 Salen, 518534 Manganese(IV), 529, 531 Manganese(V), 405, 411, 518, 527530 Marcus plot, 318 Mass spectrometry, 122, 519, 583 electrospray, 596, 622 fast atom bombardment, see Fast atom bombardment mass spectrometry MCD, see Magnetic circular dichroism Mechanisms excision, 122 of reactions, see Reaction mechanisms SOS-type, 122 Melanoma B16, 57 Melting temperature, 48, 115, 117, 118 Mercaptoethanol, 489, 622, 636 3-Mercaptopropionic acid, 490, 495, 497 Mercury(II), 120, 122, 124, 125, 154, 160 199Hg, 125 Metallocenes, 5378 carcinostatic activity, 5659, 78 dihalides, 271 DNA interactions, see DNA hydrolysis, 5962 synthesis, see Synthesis Metallopeptide DNA interactions, 427448 Metalloporphyrins (see also individual metals), 367391, 518
Metastasis, 76, 77 Methylamines, 109 di-, see Dimethylamine tri-, see Trimethylamine 4-Methoxypyridine, 312314 1-Methylcytosine, 64, 66, 109, 119126, 129, 130, 163 N-Methyl-2,7-diazapyrenium, 9699 Methylene blue, 187, 290, 291, 375 5-Methylene-2-furanone, 406408, 416, 491, 495, 496, 509, 531 Methyl green, 187 9-Methylpurine alkylated, 63 1-Methylpyrimidine, 63 5-Methyluracil, see Thymine Methylviologen, 231, 329, 339, 341 pyrene derivatives, 356 Mice athymic, 57 fibroblast cells, see Cells leukemic, see Leukemia mammary adenocarcinoma, 57 Micelles (see also Vesicles), 180, 354 Michaelis-Menten reaction, 95, 98 Microchip, 298 Microscopy electron, see Electron microscopy optical, 198 Microsomes, 636 Mitochondria, 506 Mitosis, 595 inhibition, 58, 76, 78, 594
Mixed ligand complexes, see Ternary complexes Mixed solvents, 199, 200, 374 Mobility shift assays, 203, 497 Molecular light switch, 333336, 417 Molecular mechanics (see also Force field, calculations), 4, 19, 118, 166 energy minimization, 59, 16, 22 molecular dynamics simulations, 511, 16, 2023, 117, 149, 151, 224, 227, 381, 632
< previous page
page_665
next page >
< previous page
page_666
next page > Page 666
Molecular modeling, 75 history, 3 nucleic acids, 123, 148, 149, 152, 157, 167, 379, 381, 390, 500 oligonucleotides, 3, 711, 1622 Molecular orbital calculations extended Hückel, 108, 124 Molecular recognition DNA, 37, 145, 166, 428, 444, 448, 468, 471, 519, 547, 596598, 604606 motifs, see individual names nucleic acid, 336, 429, 570, 581 nucleobases, 3746 RNA, 37, 593611 Molybdenum(IV), 57, 5978 cyclopentadienyl complexes, see Cyclopentadienyl complexes Molybdenum(VI), 56, 57 Mononucleotides (see also individual names), 71, 113, 119, 380, 383, 384 Mössbauer spectroscopy, 622 Mouse, see Mice Mutagenesis, 63, 88, 168, 254, 275, 285, 433, 460 experiments, 471, 472 site-directed, 430 transplatin, 113, 131 N NADH, 491, 625, 636 NADPH, 625, 636 Neocarcinostatin, 408, 532, 533, 603 Neocuproine, 494, 504506, 509 Neodymium(III) (see also Lanthanides), 578 Netropsin, 198, 311, 417, 439441, 477, 527, 532, 533
hybrid, 441 Neutron bombardment of DNA, 328 Nickel(II), 154, 158, 254256, 262265, 404, 416, 433437, 444447, 517, 543, 565, 571, 580 complexes as probes, 537557 porphyrins, 371, 374, 375, 378380, 383, 386, 401, 402 Nickel(III), 184, 541, 543, 544, 553, 555557 Nicotinamide adenine dinucleotide (reduced), see NADH Nicotinamide adenine dinucleotide phosphate (reduced), see NADPH Niobium(IV), 57, 74 Niobium(V), 57 Nitrite, 494 Nitrogen 15N label, 107 15N NMR, see NMR dioxide, 209 nitrogen monoxide, 637 NMR, 8, 11, 1720, 93, 95, 155, 168, 224, 272, 330, 354, 375, 378, 379, 383, 454, 473, 477, 542, 543, 598, 632, 638641 1H, 44, 59, 6264, 67, 69, 72, 73, 116, 120, 122, 123, 149, 157, 184, 278, 279, 630, 633 13C, 184, 279, 640 113Cd, 633, 640 15N, 108, 111 31P, 47, 62, 67, 69, 71, 74, 407, 566 195Pt, 108, 114, 125 one-dimensional, 184, 191, 633 two-dimensional, 185, 191, 225, 335, 336, 631, 633 NOESY, 9 Nogalamycin, 477 Nuclear magnetic resonance, see NMR Nuclear Overhauser effect, 8, 9, 11, 335, 638, 640
two-dimensional spectroscopy, see NOESY Nuclear quadrupole resonance spectroscopy, 108 Nucleases, 446
< previous page
page_666
next page >
< previous page
page_667
next page > Page 667
[Nucleases] artificial, 163, 179, 230, 241, 270, 390, 399420, 485509, 546 deoxyribo-, see Deoxyribonucleases endo-, see Endonucleases exo-, see Exonucleases metalloporphyrin, 390, 399420 micrococcal, 491 P1, 408, 409 photo-, 400403 ribo-, see Ribonucleases S1, 166 T1, 539 Nucleic acids (see also DNA, Polynucleotides, and RNA), 123, 59, 62, 71, 221, 230 backbone (see also Sugar phosphate backbone), 7, 388 based diseases, 328, 390 copper(I) complexes, 485509 folding, 537557 molecular modeling, see Molecular modeling oxidation, 297319 porphyrin interactions, 367391 targeting agents, 2949 targeting by iron complexes, 454479 transplatin interactions, 105132 Nucleobases (see also individual names), 107, 109, 110, 114116, 118130, 159, 166, 181, 221, 255, 298, 310, 311, 314, 445, 519, 530, 540, 554, 555 alkylated, 63, 67, 553555 Hoogsteen, see Hoogsteen metallocene binding, 6270 oxidation, 553, 554 quartets, 126, 128, 155158, 160, 162
recognition, see Molecular recognition triplets, 146154, 164, 167 Watson-Crick pairs, see Watson-Crick Nucleosides, see individual names Nucleoside monophosphates (see also individual names) 2'-deoxy-, 62, 63 Nucleoside phosphates, see Nucleotides and individual names Nucleoside triphosphates (see also individual names), 487, 488 Nucleosomes, 455 DNA in, 473475, 477 Nucleotides (see also individual names), 5, 70, 120, 129, 131, 154, 166, 187, 312, 411, 457, 463, 471, 473, 477, 504, 525, 543, 553, 554, 565, 625, 627 di-, see Dinucleotides metallocene binding, 6270 mono-, see Mononucleotides oligo-, see Oligonucleotides poly, see Polynucleotides O Oligonucleotides (see also Nucleotides and Polynucleotides), 9, 11, 38, 92, 93, 95, 101, 107, 116119, 155, 156, 158, 159, 161165, 191, 204, 272, 305, 307, 309, 314, 330, 339, 341346, 357, 359, 390, 400, 408411, 413, 415, 417420, 505, 540543, 548, 550, 551, 553, 554, 566, 571, 575, 581, 583, 584, 606608, 638640 antisense, 563, 570, 582, 584, 585 double-stranded, 116118 metallocene interactions, 7075 modeling, see Molecular modeling platinum complexes, 1622 single-stranded, 116, 118 tethered complexes, 230 Oligopeptides, 427448 minor-groove binding, 439441
< previous page
page_667
next page >
< previous page
page_668
next page > Page 668
Oligosaccharides, 477 Oncogenes, 161, 390 Ornithine (and residues), 436438 Orthophosphate, see Phosphate Osmate ester, 166 Osmium(II), 215, 302 bis(2,2'-bipyridyl)(L) complexes, 300302, 304307, 333, 335337 bis(1,10-phenanthroline)(L) complexes, 349, 352354, 356 dipyridophenazine complex, 300302, 304307, 312, 333, 335337, 349, 352354, 356 tris(2,2'-bipyridyl) complex, 210, 308, 309 Osmium(III), 301, 302 tris(2,2'-bipyridyl) complex, 210 Osmium tetroxide, 165167 Outer-sphere (see also Inner-sphere) coordination, 150, 157, 160, 179, 312314, 358, 373, 376 Oxalate buffer, see Buffer Oxidase xanthine, 491, 494 Oxene, 495 Oxone, see Peroxosulfate 6-Oxopurine, 154 Oxo transfer, 311, 312, 314 Oxygen (see also Dioxygen) singlet, 401403, 416 Oxytricha, 155, 157 P P22 repressor, 441, 469, 470 Palladium(II), 109, 110, 124, 125 cis complexes, 110
porphyrins, 339, 341, 371, 378, 379, 418 trans complexes, 110, 130 Paramagnetic resonance, see EPR and NMR Peptide(s), 10 di-, see Dipeptides glyco-, 623 metallo-, see Metallopeptides nickel complexes, 543546 poly, 385, 436 synthesis, see Synthesis synthetic, 430 tri-, see Tripeptides Peroxide(s), 565, 629631, 633, 635, 639, 642 hydrogen, see Hydrogen peroxide tert-butylhydro-, 520 Peroxodisulfate, 190 Peroxosulfate potassium, 400, 403407, 409414, 416, 420, 433, 444, 520, 540544, 546550, 552, 553, 555, 556 Pfeiffer effect, 205, 206 Pharmaceuticals (see also Drugs), 54, 56 9,10-Phenanthrenequinone diimine, 183, 232 rhodium complexes, see Rhodium(II) ruthenium complexes, see Ruthenium(II) 1,10-Phenanthroline (and derivatives), 57, 191, 199, 200, 204209, 211214, 216220, 225241, 274291, 308, 312, 314, 318, 337, 441, 442, 504, 556, 623 N-alkyl-, 225 5-(aminoglutaryl)-, 441, 442 -2-carbinol, 487, 488 5-chloro-, 314, 318 copper complexes, see Copper(I) and Copper(II) 2,9-dimethyl-, 486, 493, 494, 501, 503507
4,7-dimethyl-, 347349, 353 4,7-diphenyl-, 182, 205210, 216, 228, 274 5-(iodoacetamido)-, 430
< previous page
page_668
next page >
< previous page
page_669
next page > Page 669
[1,10-Phenanthroline (and derivatives)] iron complexes, see Iron(II) 4-phenyl-, 504, 506 5-phenyl-, 504, 507 4,7-phenylsulfonyl-, 635, 637 rhodium complexes, see Rhodium(II) ruthenium complex, see Ruthenium(II) structures, 182 5'-succinylamide, 493 3,4,7,8-tetramethyl-, 182, 206, 208, 209, 216, 501 titanium(IV) complexes, 57 Phenazine dipyrido(3,2-a:2',3'-c)-, see Dipyrido(3,2-a:2',3'-c)phenazine pyrido-, 208 structures, 183 Phenolates (or phenols, and phenolic groups) (see also individual names), 32, 33, 58, 555 N-methyl-o-aminothio-, 57 seleno-, 58 thio-, see Thiophenolates Phenylalanine poly-, 49 Phenylthiolate, 182, 195, 196, 198, 202 Phleomycin, 641 Phosphatases, 529 alkaline, see Alkaline phosphatase Phosphate (including hydrogen and dihydrogen phosphate), 409411, 494 diethyl-, 578 Phosphate ester (see also Phosphodiester), 407, 505, 531, 565, 568, 569, 572, 574, 578, 581 bis(4-nitrophenyl)-, 569
cyclic, see Cyclic phosphate ester hydrolysis, 570 transesterification, 567, 572 Phosphates (and groups) (see also individual names), 5, 62, 69, 71, 74, 75, 150, 151, 157, 159, 160, 167, 199, 200, 228, 255, 256, 264, 265, 369, 405408, 414, 487, 488, 529531, 606, 627 α-Phosphatidyl-L-serine, 76 Phosphodiester(s), 20, 455, 563, 572, 578, 580 backbone, 408, 454, 492, 493, 498, 501, 502, 562, 564, 585 bond, 68, 504, 505 cleavage, 493, 498 complexation, 567569 hydrolysis, 570 ''pseudohydrolysis", 400, 412, 413 transesterification, 563565, 567, 569 Phosphonate ester, 569, 570 methyl-, 230 Phosphonoformic acid, 76 Phosphorothioates, 162, 230 Phosphorus 32P label, 77, 314, 343, 411, 419, 432, 445, 533, 541, 554 31P NMR, see NMR Phosphorus(V), 567, 569 pentacoordinate intermediate, 571 porphyrin, 403 Phosphoryl transfer, 562, 569 Photochemical reaction type II, 401 Photocleavage of DNA, see DNA Photofrin, 401 Photolysis, 269291 DNA, see DNA, photocleavage
Photosensitizers, 400, 401, 403 Phototherapy, 368, 400, 401, 417 Phthalate monoperoxo-, 403405, 433, 520, 526, 540, 544, 546, 553 Phthalocyanines, 401 stacked, 326, 360
< previous page
page_669
next page >
< previous page
page_670
next page > Page 670
Piperidine, 100, 166, 314, 417, 418, 437, 553, 554 Plasmids, 229, 230, 312, 601 DNA, see DNA Platinum (different oxidation states), 107, 168 195Pt, 120 195Pt NMR, see NMR cis complexes (see also Cisplatin), 13, 14, 89, 97100, 124, 130, 162 oligonucleotide complexes, see Oligonucleotides trans complexes (see also Transplatin), 92, 96, 122127, 129, 130 Platinum(II), 13, 14, 71, 87101, 105132, 162, 163, 184186, 194196, 198, 202, 204, 220, 225, 227, 541 cis-diamminedichloro-, see Cisplatin porphyrins, 371 trans-diamminedichloro-, see Transplatin Platinum(IV), 118, 129 Platinum anticancer drugs (see also individual names), 100, 204 cisplatin, see Cisplatin transplatin, see Transplatin PMR, see NMR Polyamines macrocyclic, 3049 Polymerases β, 17 DNA, see DNA polymerases RNA, see RNA polymerases Polynucleotides (see also DNA, Oligonucleotides, and RNA), 7075, 149, 151, 154, 160, 161, 168, 194, 256, 276, 277, 288, 290, 335, 390, 406, 408, 570 poly(A), 46, 151, 276, 277, 290 poly(A)·poly(U), 382 poly(A-U), 4648 poly(C), 47, 153, 276, 277, 290
poly(dA), 406 poly(dA)·poly(dT), 187, 358 poly(dA-dT), 221223, 226, 227, 313, 314, 355, 356, 359, 374, 375, 382, 386, 388, 402, 409, 410, 416, 489 poly(dG)·poly(dC), 216, 358 poly(dG-dC), 208, 222, 223, 226229, 313, 314, 355, 356, 359, 374, 375, 380, 382384, 388, 402, 416 poly(dG-dC)·poly(dA-dT), 378 poly(G), 47, 153, 155, 276, 277, 290 poly(I), 154, 161 poly(I)·poly(C), 382 poly(U), 4649, 151, 160, 276, 290 poly[d(A-T)]·poly[d(A-T)], 410 poly[d(G-C)]·poly[d(G-C)], 49, 410 Polypyridyl complexes, 179, 181183, 212, 229, 271, 273284, 349 rhodium, see Rhodium(III) ruthenium, see Ruthenium(II) Porphyrins, 309, 367391 aggregation on DNA, 386389 deutero-, 414 hemato-, 401, 416 metallo-, see Metalloporphyrins and individual metals photoactivatable, 400403 proto-, see Protoporphyrins self-aggregation, 370, 371, 373, 386 solution properties, 370, 371 structures, 369, 372, 401, 415 Potassium ion, 130, 151, 156158 peroxosulfate, see Peroxosulfate Potentials, see Redox potentials Potentiometric titrations, 40, 43, 111 chloride, 60 Proflavin, 96, 97, 199, 201, 254265, 375, 381, 383
di-tert-butyl-, 187, 199, 201
< previous page
page_670
next page >
< previous page
page_671
next page > Page 671
Prokaryotes, 113, 503, 504 Proline poly-, 359 Propenal, 406, 414, 606, 627, 636 Protein(s) (see also individual names), 10, 23, 30, 168, 255, 358, 360, 609 CAP, 430 cro, 430, 468, 469 -DNA interactions, see DNA electron transfer, 326 folding, 190 glyco-, 594 leucine zipper, see Leucine zipper motif -RNA complexes, 478 synthesis, see Synthesis tethered iron complexes, 479 Protonation constants, see Acidity constants Protoporphyrin IX, 401, 404, 414 Pulse radiolysis, 493 Purine(s) (see also individual names), 62, 63, 67, 69, 76, 88, 93, 94, 100, 116, 120, 121, 126, 128, 145, 146, 148154, 164, 165, 227, 229, 276, 279, 280, 283, 288, 336, 517, 542, 543, 554 depurination, see Depurination 9-methyl-, see 9-Methylpurine nucleotides (see also individual names), 19 6-oxo-, see Hypoxanthine Pyrazine(s), 234 2,2'-bi-, see 2,2'-Bipyrazine 2,3-di-2-pyridyl-, 183 4',7'-phenanthrolino-,5',6':5,6-, 183, 213, 217 structures, 183 Pyridine(s), 109, 166, 312, 357, 359, 555
4-methoxy-, 312314 Pyrimidine(s) (see also individual names), 43, 44, 46, 62, 63, 67, 93, 94, 100, 109, 145, 146, 148152, 154, 164, 165, 227, 229, 336, 446, 477, 517, 598, 622, 625, 627, 628, 630, 632, 638, 640, 641 1-methyl-, see 1-Methylpyrimidine Pyropheophorbides, 401, 417 Pyrophosphate bond cleavage, 487 Q Quantum efficiency, 274, 276, 280, 283, 285, 291 Quenching electron transfer, see Electron transfer fluorescence, see Fluorescence luminescence, see Luminescence Quinoline hydroxy-, 567 Quinones (see also individual names) hydro-, see Hydroquinone 9,10-phenanthrene-diimine, see 9,10-Phenanthrenequinone diimine semi-, see Semiquinones Quinoxaline, 609 R Radicals, 234, 281, 285, 432, 434, 446 alkyl, 529, 531, 553 , 281 carbon, 627, 628 deoxyguanosine, 281283 diffusible, 444446, 493, 564 free, 328 guanine, 314 hydroxyl, 164, 165, 404, 405, 414, 454479, 491, 492, 495, 525, 555, 628, 629
< previous page
page_671
next page >
< previous page
page_672
next page > Page 672
[Radicals] nondiffusible, 434, 436, 446 oxygen, 285, 564 peroxy, 404, 606, 627, 628 phenoxy, 553 Radiolysis pulsed, see Pulse radiolysis Raman spectroscopy, 108, 190, 377 Rate constants, 110, 111, 114, 222, 223, 231, 304, 317, 318, 350355, 357, 380, 381, 402, 572, 583, 584, 621, 635637 pseudo-first order, 89, 578580 Rates of reactions (see also Rate constants), 9194, 100, 131 Reaction mechanisms (see also Kinetic studies), 99, 405, 407, 523, 607, 626 Reaction rates, see Rates of reactions and Rate constants Recognition, see Molecular recognition Redox potentials, 273, 280, 299301, 305, 309, 312314, 317, 318, 335, 338, 353, 555, 556, 636 Reductases cytochrome b5, 625 cytochrome P450, 625 Reduction potentials, see Redox potentials Repressors bacteriophage, see Bacteriophage lac, see Lac repressor P22, see P22 repressor prokaryotic, 468477 Resonance light scattering, 386388 Resonance Raman spectroscopy, see Raman spectroscopy Restriction enzymes, 38, 75, 78, 412, 413, 524, 544 Rhenium(V), 57, 312, 313, 314 Rhenium(VI), 312, 314
Rhodamin, 230 Rhodium(II), 274, 282, 283, 347, 349 Rhodium(III), 184186, 201, 207, 208, 215, 218, 229, 232, 402 phenanthrenequinonediimine complexes, 185, 186, 201, 218, 232, 335341, 343, 344, 346356, 441, 442, 517 polypyridyl complexes, 229, 273284, 347 tris(1,10-phenanthroline) complex, 212, 231, 273, 275, 331, 346, 347 Riboflavin, 281 Ribonucleases, 101, 583, 610, 611 A, 563 T1, 552 T2, 552 Ribonucleic acid, see RNA Ribose (and moiety), 311, 456, 540, 571, 572 deoxy-, see Deoxyribose Ribosomes, 478, 479, 550 Ribozymes, 2, 38, 562, 565, 584 active sites, 548, 562 group I intron, 547, 549 hairpin, 547, 548 hammerhead, 565 selfsplicing, 548 Tetrahymena thermophila, 478, 547, 548 RNA (see also Nucleic acids), 11, 23, 42, 46, 59, 161, 168, 195, 310, 336, 400402, 420, 454, 456, 457, 460, 461, 505, 538 A-, 206, 208, 334, 502 antisense, 161, 230, 390, 551, 563 catalytic, see Ribozymes cleavage (see also strand scission), 486, 501, 593611 damage, 610 -DNA hybrids, 216, 581, 603605 double-helical, 333, 334 double-stranded, 581, 598
duplex, 216, 382, 383, 551, 552 hairpin, 603, 608, 610 hydrolysis, 554 hydrolytic cleavage, 561585
< previous page
page_672
next page >
< previous page
page_673
next page > Page 673
[RNA] loops, 501503 m-, 49, 101, 420, 478, 479, 550, 552, 563, 582, 584, 601603, 608610 metal-catalyzed cleavage, 561585 minor groove, 606, 610 -protein complexes, 478 pseudoknot, 550, 551, 602, 603, 608610 r-, 610 recognition, see Molecular recognition retroviral, 550 ribosomal, 472, 473, 477, 601 secondary structure, 580, 583, 610 self-cleaving (see also Ribozymes), 565, 580, 581 single-stranded, 230, 461, 501, 502, 509, 581, 582, 598 strand scission (see also cleavage), 336, 503, 504, 553, 555 structure, 478, 502, 549, 552, 600 t-, 159, 382384, 478, 479, 502, 580, 581, 599605, 609 tertiary structure, 546552, 580, 583, 602, 603, 610 tetraplex, 158 translation, 161, 582 triplet base pair, 601 tRNAAsp, 599, 610 tRNAHis, 599601, 604, 605, 610 tRNALeu, 599, 610 tRNAPhe, 116, 383, 546548, 562, 581, 599, 600, 606, 610 tRNASeCys, 599 tRNASer, 599, 600 tRNATyr, 605 RNA polymerases, 505507, 509, 554, 555
II, 439 E. coli, 475, 476, 503, 504 RNases, see Ribonucleases Rubidium ion, 151, 156, 157 Ruthenium(II), 300, 301, 311, 312, 402 bis(2,2'-bipyridyl)(L) complexes, 184, 186, 191, 206, 207, 212, 213, 218, 232, 234240, 359 bis(9,10-phenanthrenequinone-diimine)(L) complexes, 215, 338, 344, 346 bis(1,10-phenanthroline)(L) complexes, 181, 185, 186, 196, 198, 199, 204, 213, 214, 218, 225, 230, 231, 275, 333336, 339, 343, 352356, 517 dimethylsulfoxide complexes, 271 dipyridophenazine complexes, 181, 185187, 191, 196, 198, 199, 201, 204, 207, 213215, 217, 218, 220, 225, 230, 232, 234240, 333336, 338341, 343, 344, 346356 tris(2,2'-bipyrazine) complex, 187 tris(2,2'-bipyridyl) complex, 187, 190, 192, 201203, 210, 217, 220222, 231, 233, 239, 241, 316, 317, 329, 339, 341 tris(1,10-phenanthroline) complex, 181, 184187, 196, 201203, 205208, 211, 220, 221, 223229, 275, 329332, 347 tris(4,7-diphenyl-1,10-phenanthroline) complex, 205210, 216, 228 tris(1,4,5,8-tetraazaphenanthrene) complex, 183, 190, 203, 209 tris(3,4,7,8-tetramethyl-1,10-phenanthroline) complex, 206, 208, 209 Ruthenium(III), 215, 357 hexaamine complex, 204, 338341, 349 imidazole complexes, 357 tris(2,2'-bipyridyl) complex, 317, 318
< previous page
page_673
next page >
< previous page
page_674
next page > Page 674
S S. pombe, see Schizosaccharomyces Saccharides (see also Sugar and individual names) oligo-, see Oligosaccharide Salen, 518534 derivatives, 521, 524, 525, 553555 Salmon testis DNA, see DNA Sarcoma 180, 57 Scatchard plots, 192194, 210215, 256 Schiff bases, 519, 580, 581 Schizosaccharomyces pombe, 599, 600 Sea urchin, 166 Sedimentation studies, 186, 192, 198, 200 Self-aggregation of porphyrins, see Porphyrins Semiquinones, 328 Serine phosphatidyl-, see Phosphatidylserine Serum albumin, see Albumin Silver(I), 123, 124, 160, 161 porphyrins, 371 Single photon counting, 344, 347, 350, 356 Skin malignant diseases, 254 Sodium ion, 9, 34, 151, 156158, 167, 192, 197, 227, 229, 291, 381 poly(acrylate), 222, 223 poly(styrenesulfonate), 222, 223 Sodium dodecyl sulfate, 354 ''Soft" species, see "Hard and soft" species Solid phase synthesis, 430, 436, 437
Solvent polarity, 181 Sonication, 199 Soret band, 370, 371, 373375, 382, 385, 402 Spectrophotometry (see also Absorption bands and spectra, Infrared spectroscopy, and UV), 382 Spermidine, 150 Spermine, 151, 159, 164, 228, 415, 420 Stability constants (see also Association constants, Binding constants, Dissociation constants, and Equilibrium constants), 31, 32, 34, 36, 37, 40, 4244, 255, 331, 336, 379, 402, 410, 414, 415, 433, 444, 497, 501, 570, 573, 576 Stacking, 43, 44, 46, 157, 158, 165, 166, 181, 197, 198, 209, 220, 221, 231, 281, 326328, 333, 334, 341, 344, 345, 347, 350, 353355, 357, 358, 360, 361, 380, 381, 464, 465, 542, 543, 602, 640 Standard potentials, see Redox potentials Stereoselectivity, 179 Steric hindrance, 75, 220, 225, 274, 500, 501 interactions, 76, 525 repulsion, 518 Stern-Volmer plot, 338, 341, 352 cis-Stilbene, 596 Stopped flow methods (see also Kinetic studies), 195, 317, 318, 631 Strontium ion, 157 Structure-activity relationships, 442 platinum drugs, 100 Sucrose buffer, 233, 236 Sugars (see also individual names), 150, 167, 255, 312, 407, 408, 410, 411, 414, 416, 476, 555, 627 phosphate backbone (see also DNA, backbone), 231, 405, 517, 518, 581, 627 pucker, 69
< previous page
page_674
next page >
< previous page
page_675
next page > Page 675
Sulfate(s), 494 dimethyl, see Dimethylsulfate peroxo-, see Peroxosulfate Sulfonamides (see also individual names) aromatic, 3437 4-nitrobenzene-, 35 p-toluene-, 35 Sulfur donor ligands (see also Mercapto--, Thiols, and individual names), 58 platinum complexes, 89, 122 Superoxide, 403406, 414, 489, 491, 494, 520, 595, 625, 636 Superoxide dismutase, 414, 636 Supramolecular assemblies, 390 Suramin, 76 Surface binding, 299, 329332 Synthesis (of) DNA, 59, 75, 78, 113, 130 metallocene, 5456 peptide, 430, 436, 437, 439 protein, 59, 433 solid phase see Solid phase synthesis transplatin, 107, 108 Synthetases aminoacyl-tRNA, 609 cysteinyl-tRNA, 609 T Tamoxifen, 76, 77 Techniques (see also individual methods) for DNA-transition metal interaction studies, 181204 list of, 182184
Temperature jump studies, 380 Terbium(III) (see also Lanthanides), 578 Ternary complexes, 39, 40, 44, 47, 49, 125, 129, 130, 182184, 254, 256, 258265, 404, 495 chromium, see Chromium(III) osmium, see Osmium(II) rhodium, see Rhodium(III) ruthenium, see Ruthenium(II) 2,2',2''-Terpyridyl, 182, 184186, 191, 195, 196, 198, 202, 204, 215, 218, 220, 227, 311, 312, 572, 574, 582, 584 Territorial binding, 300, 308, 379 Tetraamines (see also individual names) macrocyclic, 3034, 38 1,4,5,8-Tetraazaphenanthrene ruthenium complexes, see Ruthenium(II) 1,4,7,10-Tetraazacyclotetradecane, see Cyclen 1,4,8,11-Tetraazacyclotetradecane, see Cyclam Tetrahymena thermophila, 478, 547 ribozyme, see Ribozymes Texapyrin, 574, 582584 Theophylline, 69 Thiazole bi-, 447, 596, 598, 620, 630634, 637642 Thin-layer chromatography, 599, 606 Thiocyanate, 57, 58 Thiols (and thiolate groups) (see also Mercapto- and individual names), 165, 404, 408, 489, 490, 495, 497, 501, 532, 625, 636 butane-, 202 2-hydroxyethane-, see 2-Hydroxyethanethiolate phenyl-, see Phenylthiolate Thionine, 187 Thiophenols, 58, 437 amino-, 57, 58
Thiourea, 92, 108, 457 Threonine, 598 Thrombin aptamers, 156 Thulium(III) (see also Lanthanides), 578
< previous page
page_675
next page >
< previous page
page_676
next page > Page 676
Thymidine, 39, 40, 4244, 46, 311, 465, 466, 477, 544 azido-, see Azidothymidine Thymidine monophosphate, see TMP Thymidine triphosphate, see TTP Thymine (and moiety), 3744, 46, 49, 124, 165167, 256, 310, 402, 529, 544, 584 1-methyl-, 44, 45, 119121, 123, 127 quartet, 158 Titanium(III), 69 Titanium(IV), 5662, 69, 74 TMP 5'-d-, 63, 67, 68 Topoisomerase, 7578, 186, 200, 201, 375 II, 77, 78, 620 Torsion energy, 15 Toxicity (of) cyto-, see Cytotoxicity hepato-, see Hepatotoxicity metallocene complexes, 57 porphyrins, 401 Transcription of DNA, see DNA Transcription factors IIIA, 472, 473 Sp1, 544 Transesterification phosphate ester, see Phosphate ester phosphodiester, see Phosphodiester Transfer charge, see Charge transfer electron, see Electron transfer
hydrogen, 311, 312 phosphoryl, see Phosphoryl transfer Transferase DNA, 487 Transition metal ions (see also individual elements), 123, 42, 152154, 160, 163, 168, 254, 256, 260265, 270, 273, 332, 371, 404, 413, 428, 439 complexes, 1216, 22, 42, 329 substitution-inert complexes, 177240 Transplatin, 19, 88101, 105132, 161 antitumor activity, 113 cytotoxicity, see Cytotoxicity hydrolysis, 89, 110112, 114, 131 mutagenicity, see Mutagenesis properties, 108 synthesis, see Synthesis 1,5,9-Triazacyclododecane, 3234, 36, 38 Triazacyclononane, 572 1,4,7-Triazaheptane, see Diethylenetriamine Trien, see Triethylenetetramine Triethylenetetramine, 31, 32, 572 dimethyl-, 336 Trimethylamine, 109 Trimethylphosphine, 13, 14 Trinucleotides, 119 d(GCG), 116 d(GpTpG), 90, 91 d(GTG, 116 UUU, 581 Tripeptides (see also individual names), 429 Gly-Gly-His, 433436, 438440, 444448, 543545 Gly-His-Gly, 446 Gly-His-Lys, 440, 441, 446
Orn-Gly-His, 436438 Tris buffer, see Buffer Tritium label, 59 Tumor (see also Cancer and individual names), 401 Ehrlich ascites, see Ehrlich human epidermoid, 57, 61 human, 100 mammalian, 77 murine, 100 repair mechanism of cells, 59 solid, 55 TTP, 533 Tungsten(VI), 56
< previous page
page_676
next page >
< previous page
page_677
next page > Page 677
Two-dimensional nuclear Overhauser enhancement spectroscopy, see NOESY U Ubiquitin, 11 Ultraviolet, see UV Uracil (and moiety), 43, 4749, 124, 130, 166, 599 1,3-dimethyl-, 119121 1-methyl-, 119, 123, 125, 129, 130 5-methyl-, see Thymine Uranyl ion, 167 Uric acid, 280 Uridine, 39, 40, 42, 43, 46, 123, 311, 610 6-aza-, 39, 44 bromo-, 39, 4244 Uridine 5'-triphosphate, see 5'-UTP 5'-UTP, 504 UV absorption (or spectra) (see also Absorption bands and spectra), 34, 108, 115, 181, 234, 387, 474, 494, 519, 633 circular dichroism, see Circular dichroism irradiation, see Irradiation studies, 47, 59 V Vanadium(IV) and/or vanadyl ion, 56, 5963, 7478, 371 Van der Waals interactions, 4, 226, 380, 383, 384, 517, 518, 527 radii, 525 surfaces, 639 Vesicles (see also Micelles), 180 Viologen, 232 methyl, see Methyl viologen
Viruses (see also individual names), 504, 509 human immunodeficiency, see HIV mouse mammary tumor, 550 retro-, 390 SV40, see Simian virus 40 tobacco ringspot, 548 Viscosity studies, 185, 198203, 217, 224, 225, 375 Visible spectra, see Absorption bands and spectra Vitamin C, see Ascorbate Voltammetry (see also Electrochemistry), 298 cyclic, see Cyclic voltammetry differential pulse, 310, 311 normal pulse, 303308, 315 oxidation kinetics, 314318 square-wave, 315, 317, 318 W Watson-Crick base pairing, 63, 69, 126, 127, 145, 146, 148, 163, 378, 541 hydrogen bonds, 11, 63 modified ligands, 121 X Xanthine, 491, 494 oxidase, see Oxidases Xenopus, 472, 473 X-ray absorption spectroscopy, 596 X-ray crystal structures, 16, 34, 36, 39, 41, 44, 45, 62, 6467, 69, 70, 108, 109, 111, 112, 118121, 123, 125127, 130, 155, 157, 159, 168, 184, 204, 205, 216, 240, 272, 468, 546 (fiber) diffraction, 148, 149, 155, 161, 184, 204
< previous page
page_677
next page >
< previous page
page_678 Page 678
Y Yeasts (see also individual names), 599601, 610 proteins, 430, 442, 443 Ytterbium(III), (see also Lanthanides), 578 Yttrium(III), 575, 577 Z Zeolites, 180 Zinc(II), 2949, 151, 152, 154, 160, 186, 201, 255, 263, 264, 266, 404, 416, 436, 487489, 565, 567, 568, 571573, 580, 620, 622 bleomycin, 624, 632, 640, 641 porphyrins, 371, 372, 375, 379, 380, 383, 385, 390, 402, 403, 416, 417 tetraamine complex, 3032, 38 tris(1,10-phenanthroline) complex, 206 Zinc amalgam, 288 Zinc fingers, 436, 471473, 544, 546 TFIIIA, see Transcription factors Zirconium(IV), 56, 57, 5961, 69, 74
< previous page
page_678