Electrophoresis (Journal of Chromatography Library) Volume 18

Citation preview

JOURNAL OF CHROMATOGRAPHY LIBRARY - volume 18

electrophoresis a survey of techniques and applications part A: techniques

JOURNAL OF CHROMATOGRAPHY LIBRARY

Volume 1

Chromatography of Antibiotics by G.H. Wagman and M.J. Weinstein

Volume 2

Extraction Chromatography edited by T. Braun and G. Ghersini

Volume 3

Liquid Column Chromatography. A Survey of Modern Techniques and Applications edited by 2. Deyl, K. Macek and J. J a d k

Volume 4

Detectors in Gas Chromatography by J. hv6;k

Volume 5

Instrumental Liquid Chromatography. A Practical Manual on High-Performance Liquid Chromatographic Methods by N.A. Parris

Volume 6

Isotachophoresis. Theory, Instrumentation and Applications by F.M. Everaerts, J.L. Beckers and Th.P.E.M. Verheggen

Volume 7

Chemical Derivatization in Liquid Chromatography by J.F. Lawrence and R.W. Frei

Volume 8

Chromatography of Steroids by E. Heftmann

Volume 9

HPTLC - High Performance Thin-Layer Chromatography edited by A. Zlatkis and R.E. Kaiser

Volume 10

Gas Chromatography of Polymers by V.G. Berezkin, V.R. Alishoyev and 1.6. Nemirovskaya

Volume 1 1

Liquid Chromatography Detectors by R.P.W. Scott

Volume 12

Affinity chromatography by J. Turkov6

Volume 13

Instrumentation for High-Performance Liquid Chromatography edited by J.F.K. Huber

Volume 14

Radiochromatography. The Chromatography and Electrophoresis of Radiolabelled Compounds by T.R. Roberts

Volume 15

Antibiotics. Isolation, Separation and Purification edited by M.J. Weinstein and G.H. Wagman

Volume 16

Porous Silica. I t s Properties and Use as Support in Column Liquid Chromatography by K.K. Unger

Volume 17

75 Years of Chromatography - A Historical Dialogue edited by L.S. Ettre and A. Zlatkis

Volume 18

Electrophoresis. A Survey of Techniques and Applications. Part A: Techniques edited by 2. Deyl

JOURNAL OF CHROMATOGRAPHY LIBRARY - volume 78

electrophoresis a survey of techniques and applications part A: techniques

editor Z.Deyl Physiological Institute, Czechoslovak Academy of Sciences, Prague

co-editors F.M. Everaerts, Z. Pruslk and P.J. Svendsen

ELSEVIER SCIENTIFIC PUBLISHING COMPANY Amsterdam - Oxford New York 1979

-

ELSEVIER SCIENTIFIC PUBLISHING COMPANY

335 Jan van Galenstraat P.O. Box 21 1, 1000 AE Amsterdam, The Netherlands

Distributors for the United States and Canada: ELSEVIER/NORTH-HOLLAND INC.

52, Vanderbilt Avenue New York, N.Y. 10017

Library OF Congress Cataloging in Publication Data

Main entry under title:

Electrophoresis : a survey of'techniques and applications. ( J o u r n a l of chromatography l i b r a r y ; v. l8- ) CONTEITTPS: pt. A. Techniques. Includes bibliographical references and index. 1. Electropharesis. I. Deyl, Zdenkk. 11. Series.

QD79.344345 541' .37 ISBN 0-444-41721-4 ( p t . A )

79-22525

ISBN 0-444-41721-4(Vol. 18) ISBN 0-444-41616-1(Series)

0 Elsevier Scientific Publishing Company, 1979. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic. mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Scientific Publishing Company, P.O. Box 330, 1000 AH Amsterdam, The Netherlands. Printed in The Netherlands

Contents

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1X XI XV

1. Theory of electromigration processes (J . Vack) . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Equilibria in electrophoretic systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Processes in electrophoretic systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mathematical description of the electrophoretic process . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 8 18 20

2 . Classification of electromigration methods (J . Vaclk) . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zone electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The moving boundary method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lsotachophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Focusing methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Combined methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23 23 24 26 27 29 33 37

3 . Evaluation of the results of electrophoretic separations (J . Vack) . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Determination of mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principles of quantitative evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39 39 40 42

4 . Molecular size and shape in electrophoresis ( 2 . Deyl) . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elimination of charge differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Choice of standard series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular size determination by pore limit electrophoresis . . . . . . . . . . . . . . . . . . Determination of Stokes radii by rheophoresis . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45 45 49 61 62 63 66

5 . Zone electrophoresis (except gel-type techniques and immunoelectrophoresis) (W . Ostrowski) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Paper techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thin-layer electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 . Gel-type techniques (2. Flrkal) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory of gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Starch gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

69 70 96 103 113 113 114 116

VI

CONTENTS Acrylamide gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agarose gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agarose-acrylamide composite gels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..

7 . Quantitative immunoelectrophoresis (P J Svendsen) . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus and accessories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

117 128 129 130 133 133 134 136 139 153

8 . Moving boundary electrophoresis in narrow-bore tubes (F. M . Everaerts and J . L. Beckers) Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principles of moving boundary electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . State of the art and comparison with other techniques. . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 155

9 . Isoelectric focusing (N .Catsimpoolas) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carrier ampholytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Support media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thin layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Density gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Free solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Twodimensional methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lmmunoisoelectric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transient state isoelectric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

167 167 168 170 171 172 173 175 176 176 176 179 190 190

.

155 156 160 162 165

10. Analytical isotachophoresis (J . Vaclk and F M . Everaerts) . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality and quantity in isotachophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

193 194 195 206 208 218 224

11. Continuous flow-through electrophoresis (2. Pruslk) . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrophoretic cells of other shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Factors affecting zone width in flow-through cells ....................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

229 229 231 235 235 250

12. Continuous flow deviation electrophoresis (A . Kolin) ....................... 253 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Theoretical aspects of continuous flow deviation electrophoresis . . . . . . . . . . . . . . . 255 Modes of implementation of continuous flow deviation electrophoresis . . . . . . . . . . . 261 Flat fluid band electrophoresis (“free-flow” electrophoresis) . . . . . . . . . . . . . . . . . 261

VII

CONTENTS Endless fluid belt electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Symbols and units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I3 . Preparative electrophoresis in gel media (2. Hrkal) . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel composition and concentration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel dimensions and sample load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Voltage gradient and the electrophoresis time . . . . . . . . . . . . . . . . . . . . . . . . . . Duration of electrophoresis and flow-rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14 . Preparative electrophoresis in columns (P. J . Svendsen)

......................

267 295 295 296 296 299 299 299 300 300 301 305

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application and procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

307 307 308 315 324

(P. Blanickg) . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preparative isoelectric focusing technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

327 328 328 332 341 342

15 . Preparative isoelectric focusing

.

16. Preparative isotachophoresis (P.J Svendsen)

............................

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

17 . Preparative isotachophoresis on the micro scale (L Arlinger)

345 345 346 348 355 362

. . . . . . . . . . . . . . . . . . . 363

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Small-scale preparative isotachophoretic methods . . . . . . . . . . . . . . . . . . . . . . . Principle of preparative capillary method . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of preparative capillary method . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

363 364 365 370 376

List of frequently occurring symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

379

.................................................

385

Subject index

This Page Intentionally Left Blank

Contributors

Dr. L. Arlinger, Development Department, LKB Producter AB, S-161 25 Bromma 1 , Sweden Dr. P. Blanicky, Research Institute of Child Development, Faculty of Pediatrics, Charles University, Prague 5-Motol, Czechoslovakia Dr. N. Catsimpoolas, Biophysics Laboratory, Department of Nutrition and Food Science, Massachussetts Institute of Technology, 02 139 Massachussetts Avenue, Cambridge, Mass., U.S.A. Dr. Z. Deyl, Institute of Physiology, Czechoslovak Academy of Science, Bud6jovicka 1083, Prague 4-KrE, Czechoslovakia Dr. F. M. Everaerts, Department of Instrumental Analysis, Eindhoven University of Technology, Eindhoven, The Netherlands Dr. 2. Hrkal, Institute of Hematology and Blood Transfusion, U nemocnice 1, 128 20 Prague 2. Czechoslovakia Dr. A. Kolin, School of Medicine, Biophysics Laboratory, The Center for the Health Sciences, Los Angeles, Calif. 90024, U.S.A. Dr. W.Ostrowski, N. Copernicus Academy of Medicine in Krakow, Institute of Medical Biochemistry, ul. Kopernika 7, 31 -034 Krakow, Poland Dr. Z. Prusik, Institute of Organic Chemistry and Biochemistry, Czechoslovak Academy of Science, Flemingovo nam. 2 , Prague 6, Czechoslovakia Dr. P. J. Svendsen, Protein Laboratory, University of Copenhagen, Sigurdsgade 34,2200 Copenhagen, Denmark Dr. J . Vacik, Department of Physical Chemistry, Faculty of Science, Charles University, Albertov, Prague 2 , Czechoslovakia

This Page Intentionally Left Blank

Preface

The separation and characterization of the individual components of mixtures has been a prerequisite for progress in chemistry since the infancy of the science. It is therefore most fitting that the Dutch word for the entire science of chemistry, Scheikunde, means the art of separation. Until the beginning of this century, the most commonly used separation methods were filtration, distillation and crystallization. Pioneering studies on electrophoresis were published in 1892 by Picton and Linder, in 1897 by Kohlrausch and in 1899 by Hardy. Tswett’s first work on chromatography was published in 1903. Among the earliest centrifugation experiments one can mention the attempt by De Lava1 in 1879 to separate cream from milk. Electrophoresis, chromatography and centrifugation have since become the most widely used and effective methods available for the separation and identification of all kinds of substances. Later studies have revealed several fundamental similarities among these three methods. A new electrophoresis method is thus rapidly followed by its counterpart in, for example, the field of chromatography. An example is displacement chromatography, which is analogous t o the displacement electrophoresis developed earlier by Kohlrausch. The salient features of both methods are: (1) that the solute zones follow each other like carriages in a train, (2) that the height of a zone (i.e., the maximum concentration of the solute) is a characteristic of the particular solute making up the zone and (3) that the width of the zone is proportional to the amount of solute present. Further, (4) the leading electrolyte used in displacement electrophoresis corresponds to the medium with which the chromatographic column is equilibrated initially and ( 5 ) the terminal electrolyte of the electrophoresis corresponds to the eluting medium in displacement chromatography. In order t o emphasize the similarities among different methods and thereby to increase our understanding of them, we should try to use analogous terminology as much as possible, for example, displacement electrophoresis (instead of the commonly used term isotachophoresis) and displacement Chromatography. Another illustration of the similarities among electrophoresis, chromatography and centrifugation is the fact that each of these methods can be operated in a frontal (moving-boundary electrophoresis, chromatographic frontal analysis and sedimentation-velocity ultracentrifugation), zonal (zone electrophoresis, chromatographic elution development and gradient centrifugation) or displacement (displacement electrophoresis, chromatography and centrifugation) mode. When Arne Tiselius did his pioneering work on moving-boundary electrophoresis, his interest was focused mainly on the characterization of proteins, and one should remember that his teacher, The Svedberg, was at the same time engaged in similar studies using

XI1

PREFACE

the ultracentrifuge. Tiselius’ laboratory notebooks reveal that he had accomplished a zone separation of proteins in a gelatin gel as early as 1927. Had he not underestimated the importance of these experiments and therefore declined to publish them - a misjudgement which he contemplated with regret in a later biography - the big breakthrough in biochemistry would have come much earlier than it did. As it was, more than 20 years elapsed before he and other research groups began to study zone electrophoresis in earnest and finally developed this efficient and mild method for the purification of biologically important molecules. In order to perform a meaningful investigation of the properties or function of a substance, it is often of decisive importance that the substance is pure. In all branches of chemistry one must therefore have access to high-resolution separation methods for both analytical and preparative purposes. In biochemistry and other biological sciences, separation work is often complicated by the structural lability of the substances of interest, which might prohibit large variations in pH or ionic strength. Further, the crude starting material is often extremely complex with regard t o both the number and the nature of the components present and the interactions among them, and may contain only a few micrograms of the substance that one wishes to isolate. Isolation work is thus very often the “bottleneck” in research projects within the biological sciences. Increased knowledge of different fractionation methods can greatly facilitate the choice of an appropriate method for the isolation or analysis of a particular substance and can provide the basis for many valuable inferences regarding the properties of a substance as a “bonus” to the isolation work. In connection with structural studies, one often needs to know the,principle on which the separation is based. For example, knowledge of the sieving properties of polyacrylamide gels was a prerequisite for their use in the determination of the nucleotide sequences of DNA molecules. Many papers on various biopolymers are of dubious value because the studies were carried out with heterogeneous material, and it is often not a straightforward matter to infer the properties and activities of individual components from the properties of the mixture. It is therefore understandable that modern biochemical, medical and biological research training includes exercises in separation methodology. As the treatment is often hasty and superficial, a book such as this can serve the very useful purpose of providing a good theoretical background to various electrophoresis methods and illustrating their practical applications. As a teacher in the annual Uppsala Separation School on modern biochemical separation methods I am often asked the question, “Which separation method is the best?” There is unfortunately no simple answer to this question. One method might be superior for a particular separation problem, whereas an entirely different method might be preferable in another situation. One is often obliged to use a series of methods based on different separation parameters. If the substance of interest is particularly labile, electrophoresis has several advantages over other methods, especially when applied in free solution. Electrophoresis, particularly in gels, affords a higher resolving power for biopolymers than any other method, and is therefore universally used for the analysis of protein preparations. Electrophoresis for preparative purposes has not enjoyed the same popularity, despite the fact that the resolution is nearly as good as that obtained on the analytical scale. Chromatography is considerably more popular than electrophoresis for

PREFACE

XI11

the isolation o f biological material, although its resolving power is usually lower. This is due to the fact that preparative electrophoresis is much more complicated than chromatography or analytical electrophoresis as regards both the design and construction and the operation of the apparatus. Preparative electrophoresis will not receive the attention that it deserves until easily operated equipment becomes commercially available. Publishers could also help by offering frequently revised bibliographies of electrophoresis literature so that the latest methods could become rapidly known (a loose-leaf file with pages updated every year or two might b e useful). If preparative electrophoresis o n a laboratory scale has received little attention, the situation is even worse regarding the application o f the method on an industrial scale. This is mainly due t o the inherently low capacity of the method. However, displacement electrophoresis might be useful for the preparation of valuable enzymes on a limited scale, because the concentration in a zone can be extremely high, in the range 5-20% (w/v) under ordinary experimental conditions. However, intensive research must be carried out in order t o solve some o f the problems that hinder the practical application of the method. For example, one must find some means o f preventing precipitation of proteins at the high concentrations attained. I believe that displacement electrophoresis would enjoy widespread success if a series o f well defined spacer molecules with small, known differences in mobility became comniercially available, so that one could simply select the appropriate spacer(s) for each particular experiment. Most electrophoresis experiments are carried out in ordinary buffers. There are very few reports on the use o f mixtures o f water and organic solvents, because the substances o f interest are usually water-soIuble and might even be denatured by solvents such as alcohols. However, during recent years there has been a boom in research on the structure and function of cell membranes. As the lipids and many o f the proteins that make up the membranes are not water-soluble, it is obvious that membrane research would be greatly served b y the availability of electrophoresis buffers that are capable of solubilizing these strongly hydrophobic substances. One might even expect that such buffers would be less denaturing than ordinary aqueous buffers. as they more closely resemble the natural hydrophobic environment of the membrane proteins. Of course, one can solubilize niembrane coniponents by the use of deteigents, but tlie resulting micelle formation greatly coniplicatcs the interpretation of the experimental data. Since the time of Tiselius, research at tlie Institutc o f Biochemistry in Uppsala has centred on the development of new separation methods and the improvement of existing methods. Tiselius often pointed o u t that the methodological work usually proceeds best when it is carried out in close collaboration with someone who has a particular separation problem t o solve, which mininiizcs the risk that a poor niethod might be developed for its own sake and ensures that practical ‘‘snags’’ will be detected and hopefully overcome at an early stage. When a new niethod is t o be publishcd, it is important t o aim the paper at investigators who might have little interest in the method as such but who might find it useful in connection with some current separation problem, as well as at specialists in methodology. In the recently founded Jounial of Biochenzical and Biophysical Methods every paper contains a section devoted t o a popular treatment o f the method and some areas of application. In all areas of present-day research we use more or less sophisticatcd equipment,

XIV

PREFACE

including that involved in separation work. Although this is a condition for the solution of various research problems, it also involves considerable risks if we d o not thoroughly understand the principle on which the apparatus is based and its relation t o the data that the apparatus provides. There is also the risk that the accessibility of sophisticated “black boxes” might undermine our faith in manual experimental skill and thus suppress the most straightforward approach to the solution of an experimental problem. One can easily become bound to a particular apparatus even when there are easier and better ways of obtaining the desired information. For example, many electrophoresis experiments can be carried out with very simple apparatus if one has the basic knowledge of electrophoresis contained in this book. In this connection I might remind the reader of E d Fischer’s fundamental contributions to the chemistry of amino acids, peptides and proteins, despite the primitive separation methods available to him. Experimental skill and ingenuity in the planning of experiments are just as necessary now as in Emil Fischer’s day. I have only touched upon possible future developments in electrophoresis. To speculate about the future is easy, but experience shows that such speculations seldom hold up, owing in n o small part to the dynamic character of research. I have therefore carefully avoided great visions, which are better reserved for an afterdinner talk - where they are promptly forgotten.

Institute of Biochemistry, University of Uppsala, Biomedical Center, Uppsala (Sweden)

STELLAN HJERTEN

Introduction

The present volume is the first one of a two-volume project devoted to electromigration techniques and their applications. It was the intention of the Editors to summarize general aspects of these techniques in the first volume in the proportion in which they currently are used in laboratories and at the same time trying to emphasize perspective ones. Applications of these techniques were deferred to the second volume. In order to keep the first volume consistent, it was inevitable to show the capacity of individual types of electromigration techniques by separating various classes of compounds or by using some types of separations as demonstrative examples. Thus some applications are discussed here to a limited extent. On the other hand, we tried to go into more details regarding instrumentation and some prescriptions that are more or less generally applicable. We tried to arrange mathematical and physicochemical background, in such a way so that it also could be easily understood by non-professionals using the electromigration techniques in daily routine work. It is the sincere hope of the Editors that such knowledge could help to abolish trivial problems that we are frequently witnessing in applicatory work, and which are fundamental in unsuccessful separations. Prague, August 1979

ZDEN~KDEYL

This Page Intentionally Left Blank

Chapter 1

Theory of electromigration processes J . VACIK

CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 Equilibria in electrophoretic systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Equilibria in solutions of electrolytes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Weak acids and bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Ampholytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Phase boundary equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Adsorption at the liquid-solid interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Electrical double layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Processes in electrophoretic systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Transport processes in solutions of electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Electroosmotic flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Heat conduction and heat flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Transport phenomena in stabilizing media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Spatial effects of the inner structure of capillary systems . . . . . . . . . . . . . . . . . . . 15 15 Distribution function and the constant R E . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroosmosis in capillary systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Evaporation of the electrolyte and sucking flows . . . . . . . . . . . . . . . . . . . . . . . . 17 Electrode reactions and transport phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Mathematical description of the electrophoretic process . . . . . . . . . . . . . . . . . . . . . . . . 18 20 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

INTRODUCTION Electromigration separation processes are. in general. based on differences in the mobilities of electrically charged particles in an electric field . The term electrophoretic separation processes is commonly used . The system studied consists of (a) liquid phase (the basic electrolyte system including the compounds to be separated). in which separation takes place. (b) solid phase. which is in contact with it (i.e., walls of separation column. stabilizing porous environment. etc.), and (c) gaseous phase. which is in equilibrium with the liquid phase (in methods in which a porous carrier of the liquid phase is placed in the so-called wet chamber. or with the gaseous phase originating during electrode reactions); such a system will be referred to as the electrophoretic system . The term separation medium will be applied to electrophoretic systems that d o not yet contain substances to be separated . It is essential that the separation medium is in thermodynamic

2

THEORY OF ELECTROMIGRATION PROCESSES

equilibrium prior to the experiment. This condition is fulfilled when all components have the same chemical potential at all sites in the system, the same electrical potential or the same temperature; mechanical forces also have to be equilibrated. These conditions are first changed by adding the sample to a certain location in the separation medium. The equilibrium is further impaired by the application of an electric field to the system with the subsequent flow of electric current. During the gradual separation of individual components of the sample, this imbalance further increases. Thus, simultaneously with the separation induced by the outer force (electric field), transport phenomena that tend to equilibrate the subsequent concentration, temperature and other gradients begin to arise. In electrophoretic separations, these spontaneous processes (disturbances) usually occur to a greater extent with increasing time. Time is therefore a very important factor, and the proper choice of the time interval can be used to optimize the results of separations. With some electrophoretic methods a steady state is established that does not introduce any further changes with the advancing electrophoresis. The result of the separation is no longer time dependent after the steady state has been achieved.

EQUILIBRIA IN ELECTROPHORETIC SYSTEMS It is a prerequisite of the electromigration separation principle that the liquid phase of the separation medium is formed by a solution of electrolytes. Substances present in the solution must occur at least partially in the form of charged particles (ions). Ions originate in the solution usually from electroneutral molecules by electrolytic dissociation*. Dissociation of strong electrolytes is complete; all molecules in the solution are dissociated to ions. Weak electrolytes are dissociated only partially; in the solution they are present both in the form of ions and as neutral molecules. From this point of view of transport phenomena, any type of particle should be considered as a separate component. Thus, all types of ions, non-dissociated molecules of the solvent and weak electrolytes and nondissociated molecules of the solute can be considered as separate components of the liquid phase. Dissociated ions and their non-dissociated counterparts of a particular compound are referred to as a substance. A collection of electroneutral molecules of the electrolyte (e.g., component of the substance with zero charge) and ions derived from this electrolyte by electrolytic dissociation (e.g., components of the substance with a charge z where jzI is a number of elementary charges of the component, the sign designating the polarity of the ion) are considered as a single substance. It holds that ci = c i , = ,where ci is the concentration , concentrations ~ of the component with charge z belonging of the ith substance and c ~are to the ith substance. The concentration c of all substances in the solution is then determined by c = 8 ci = ? Z ci,=.Also,the solvent is considered as a separate substance, I

1.

=

and ions originating from its dissociation are not included among components of other substances, in spite of the fact that they contribute to their ionic equilibria. Concentration is defined by the amount of substance of a given component niazin unit volume V.

* The common term dissociation is applied, even if a protolytic reaction of an electrolyte with a solvent is in fact involved.

3

EQUILIBRIA IN ELECTROPHORETIC SYSTEMS

Thus, ci,= = ni,+ V-' . The main unit of concentration is mole * m-3, but concentrations are usually expressed in mole drn-'. Concentrations in these units will be designated cM;c = 10-3cM. Equilibria in solutions of electrolytes Most electromigration separation processes proceed in an aqueous environment. The electrolytic dissociation of a weak electrolyte X" in this solvent can be expressed by the general equation X" H 2 0 + XL-' H 3 0 + and characterized by an apparent* equilibrium constant K = [H30+] * [ X z - ' ] [ X r ] - ' [H20]-', When this constant is multiplied by a practically constant concentration of water, the so-called convection equilibrium constant, K ( X Z - ' ) , is obtained:

+

K(XL-')

=

+

K [HZO] = [Xz-'] * [H30+] * [X"]-'

(1 -1)

A pair of components X L and X'-' differing by a proton is called a conjugated pair. In this pair, XL plays the role of an acid (proton donor) and X'-' is a base conjugated with it** (proton acceptor). Xz = HCl and X"-' = C1- is an example of such a pair. In addition to protonization equilibria in solutions (i.e., equilibria in which protons are involved), equilibria occurring during formation of complexes may also play a role. These include equilibria between the complex, its central particle (usually a metal cation) and a ligand. For the sake of simplicity, equilibria during the formation of complexes will not be discussed in this chapter.

Weak acids and bases

When, for a given conjugated pair A' and A'-', z = 0, the dissociation of a weak monobasic acid is involved, which is characterized by the equilibrium A' + H 2 0 A-' + H30+. The conventional equilibrium constant

*

K(A-') = [A-'1 * [H30']

.[Ao]-'

(1 4

is also traditionally called the dissociation constant of the acid A' and is designated by K A , or simply K A . A weak dibasic acid dissociates in two steps:

+ H 2 0 + A-' + H 3 0 + A-' + H 2 0 * A-* + H 3 0 +

A'

In addition to the dissociation constant for the first step, K A , , the dissociation constant for the second step, K A Z is , also applied: The true or thermodynamic equilibrium constant is defined by means of activities. However, in electrophoresis, solutions with relatively low concentrations are usually used. Hence, the activities can, without substantial error, be identified with concentrations. tt

The designation of equilibrium constants vanes considerably. In this chapter we have characterized the constant by this conjugated base, i.e., K ( X * - ' ) .

4

THEORY OF ELECTROMIGRATION PROCESSES

K(A-2) = KA2 = [A-2] [H30+] * [A-']-'

(1 -3)

To characterize the dissociation the degree of dissociation, alZl,is often used, defined as ([A"'] [AZ])-'. For electrophoresis it is appropriate to use the molar fractionyA,z of the component A' in substance A, defined as

a l r l= [A']

+

(1.4)

YA,z = cA,z

The effective mobility (eqn. 1.28) is defined by eqn. 1.4. The molar fractions defined in this way can be expressed by means of respective dissociation constants and the pH of the solution. For the ith substance we have

Yi,z =

1t For z < 0 (when the molar fraction of anions in a weak acid is calculated), the limits B1 = z, B2 = - 1, B3 = zmin, B4 = - 1, B5 = k and B6 = - 1 are applicable; k is determined by the component present; zmin is the highest number of elementary charges of anions in the ith compound. For monobasic acids (zmin = 1) it holds that yi,-' = ai, This equation holds also for z > 0 (the molar fraction of cations in a weak base is expressed - see below). The following limits are then applicable: R 1= 0, B2 = z - 1, B3 = 1, B4 = z,,, BS = 0 , B6 = k - 1 ;zms is the hughest number of positive elementary charges of cations in the ith substance. Analogous to the dissociation constant of the acid, KA, is the dissociation constant of weak bases, K B :

'.

KB =

[B']

[OH-] [Bo]-'

(1.6)

characterizing the equilibrium Bo + H20 + B+ + OH-. However, instead of the dissociation constant of the base Bo, conventionally the equilibrium constant of the acid B' conjugated with it is often used:

K(Bo) = [BO] [H30+] * [B+]-'

(1.7)

+

characterizing the equilibrium B+ + H 2 0 t Bo H 3 0 + .The relationship between the two (Kw is the ionic product of water). constants is K(Bo) * KB = K w = Ampholytes Ampholytes (zwitterions) are particles which, according to the acidity of their environment, may exist as acids or bases. The properties of ampholytes can best be demonstrated by considering amino acids, the molecules of which are generally designated as NH2RCOOH. However, in solution they are present predominantly in the form of an internally ionized ion NH: RCOO-, which takes part in dissociation equilibria:

5

EQUILIBRIA IN ELECTROPHORETIC SYSTEMS

+ HzO + NH2RCOO- + H30t NHZRCOOH + HZO * NHZRCOO- + H3O'

NHZRCOO-

The constant K1 = [NH2RCOO-] [H30+]* [NHZRCOO-I-' then characterizes the acidity of the group -NHl and the constant K ? = [NHSRCOO-] [H30+] * [NHiRCOOH] -I the acidity of the group -COOH. These constants are usually used in various tables [2] for the characterization of amino acids. When we wish to characterize the basicity of the NHZRCOO- + OH-), we can amino group (i.e., the equilibrium NH2RCOO- + H,O use the constant K3 = [NSRCOO-] [OH-] * [NH2RCOO-]-', which is associated with the constant K 1 by the equation K1 * K3 = Kw . The isoelectric point, pHiso*, is an important characteristic of ampholytes. It is the pH value at which both the acidic and basic dissociation is the same ([NHZRCOOH] = [NH,RCOO-]). The resulting charge of the ampholyte (or a protein) is thus zero, and therefore the ampholyte does not move in the electric field. It follows from the expressions for the various constants that pHi, = s(pK, pK,). It also follows from dissociation of the ampholyte that when (pK, - pK,) < 4, the isoelectric point is determined by a single value of pH. When (pKI - pK2) > 4, the isoelectric condition is fulfilled over a broader pH range. However, also in this instance a single value of pHi, is presented in tables in the literature [2].

*

+

Phase boundary equilibria At the boundary of two phases an increase in concentration of particles with respect to their environment occurs owing to imbalance of various forces (electrostatic, Van der Waals, chemical). This phenomenon is called adsorption. In electrophoretic systems adsorption may play a role mainly at the interface of solid and liquid phases. Adsorption at the liquid-solid interface According to the character of the adsorption forces, adsorption from a solution to a solid phase can be classified as physisorption (adsorption of electrically neutral particles), polar (adsorption of ions) or chemisorption. During physisorption the adsorbed molecules are bound to the surface only by weak Van der Waals forces. Simultaneously with the physisorption of molecules, adsorption of electrically charged particles also takes place and the polar adsorption of ions occurs. Ions are bound by electrostatic forces on the surface. Owing to the nature of the surface, even selective adsorption of certain ion species can occur. Polar adsorption can be considered as a transition between physisorption and chemisorption. During chemisorption the linkage between the surface and the adsorbed particle is the strongest and resembles a chemical bond. Chemisorption can occur only in special electrophoretic systems. This is The adsorption equilibrium is characterized by the adsorption isotherm, ri,z. the amount of the i, zth component adsorbed on the surface of the solid phase of unit The designation pl for the isoelectric point is often used in the literature; it will not be used in Chapters 1-3.

6

THEORY OF ELECTROMIGRATION PROCESSES

weight, which is in equilibrium with a concentration ci,z of this component in the solution. In weak electrolytes, when the adsorption of molecules and ions occurs simultaneously, the isotherm ri reflects the adsorption of the ith substance as a whole. using the equation 0 = Let us introduce the relative coverage of the surface, 0, ri,z[ ( J ? ~ , ~ ) ~where ~ ] - ' (I'i,z)max , is the maximal amount of the i, zth component bound to the surface of liquid phase of unit weight. Table 1 .I shows three adsorption isotherms from a solution, derived under different simplified conditions. In eqns. 1.8-1 .lo, p is the adsorption coefficient, 0'= /3 [(ri,z)max]-', and a is the interaction coefficient. It is evident that when a = 0 the Frumkin equation is converted into the Langmuir equation. With low coverage of the surface (0+= 0), both of these equations coincide with the Henry isotherm.

-

TABLE 1.1 SOME ISOTHERMS FOR ADSORPTION FROM SOLUTIONS Isotherm

Equation

Equation No.

Valid for

Henry

ri,z= p q Z

1.8

Dilute solutions (adsorbed component covers only part o f the surface)

Langmuir

ri,== (qz)- p'

1.9

Dilute and concentrated solutions (limited number of sites for adsorption)

1.10

Interactions among adsorbed particles: for a > 0 - particles are attracted (adsorption is facilitated) a < 0 - particles are repulsed (adsorption is decreased) a = 0 - particles do not interact

-

-

p'

-

C ~ , ~ S(1

or

+OC~,~)-'

p * ci,z = 0 (1 - @)-I Frumkin

p' *

= 0 * (1

-

a)-'

Electrical double layer The electrical double layer is a spatial formation with a non-homogeneous distribution of charges that originates as a result of polar adsorption. Charges of one sign are on the surface of the solid phase, whereas charges of the opposite sign are scattered in the accompanying liquid phase. Hence, a potential difference cp exists between the surface of the liquid phase and the interior of the solution. The electrical double layer can, however, in addition to specific adsorption of ions, also originate in such a way that molecules on the surface of the solid phase dissociate and one species of ions passes into the solution. This type of double layer originates, for instance, on the liquid-solid interface with compounds such as silicic acid, proteins and cellulose. H+ ions resulting from the dissociation pass into the solution while negatively charged anions remain on the surface.

7

EQUILIBRIA IN ELECTROPHORETIC SYSTEMS

Fig. 1 . l . Schematic representation of potential distribution in the electrical double layer.

For a quantitative description of the electrical double layer, it is necessary to construct a model of its structure. The charge on the surface of the solid phase (the surface area of which is designated by cr) is compensated for by an excess of oppositely charged ions in the surrounding liquid. In the solution, the compensating charge is distributed in two layers, in a compact and a diffuse part of the double layer. In the compact part of the double layer a portion of compensating ions with an opposite charge with respect to that on the surface is bound by adsorption forces. The surface density of the charge in this layer is ul and the potential gradient is linear. In the diffusion part of the double layer, the remainder of the compensating ions have a surface density of charge uz; u = u1+ u z . The charge uz can move together with the solution. The potential gradient in the diffusion part oriented towards the solution approaches zero asymptotically. A scheme of the distribution of potential at the boundary of two phases is illustrated in Fig. 1.1. The total potential tp in the double layer is determined by the sum of the potential of the compact part of the double layer cpl and the potential of the diffusion part of the double layer, which is called the electrokinetic potential, [. The value of the electrokinetic potential determines the extent of electrokinetic phenomena. Electrophoresis, electroosmosis, sedimentation potential and flow potential are included among electrokinetic processes. It is therefore of interest to establish the extent to which the electrokinetic potential is involved in the total gradient of the double layer. The ratio of the potentials and [ can be deduced 1. from ratios of the surface density of charge u 0;' = u1 a;' The charge in the compact part of the double layer, ul, depends on surface adsorption forces maintaining an excess of a certain type of ions. By means of the Langmuir adsorption isotherm, an equation for monovalent ions was derived:

-

-

+

(1.11)

where 0- = exp [- (& - F [ ) * (RT)-'],p+ = exp [--.(@+ + F[)(RT)-'],@+ and qL are the specific molar adsorption potentials of cations and anions, R is the universal gas constant, F is the Faraday constant and T is absolute temperature. A simplified equation can be applied to the charge uz of monovalent ions in the diffusion part of double layer:

8

THEORY OF ELECTROMIGRATION PROCESSES

(1.12)

uz = [ ~ ~ R T {exp[Fg-(2RT)-'] -c. -exp [ - F ~ * ( ( ~ R T ) - ' ] ) ] ' ' ~

where e is the permittivity of the medium. It follows from eqn. 1.12 that the charge density uz is proportional to the square root of the concentration of a substance; at low concentrations u1is directly proportional to the concentration of a substance. Therefore, for low concentrations the fraction u (uJ1 + 1. In the liquid phase the diffusion part of the electrical double layer predominates, and [ is approximately equal to p. Electrokinetic phenomena in these solutions are pronounced. In concentrated solutions the proportion of charge in the compact part of the double layer increases, the ratio u (uZ)-l increases and 5 decreases. Thus, in more concentrated solutions, electrokinetic phenomena disappear. The diffusion part of the double layer is sometimes characterized by the so-called effective thickness of the double layer, 6 . This is the distance at which an infinitely thick layer of ions with charge uz would have t o be located in order to produce the same electric force as the real diffusion double layer. It holds that 6 =

[I, ~ ~ ' ( E R T ) - ' ] ' ' ~

(1.13)

where 1, = 0.5Zciz; is the so-called ionic strength of the solution and e is the elementary charge. After the introduction of 6 into the rearranged equation for u, the equation

r;

(1.14)

= u.s.e-1

is obtained. It follows from eqn. 1.14 that the electrokinetic potential can change due to changes in the effective thickness of the double layer, to changes in the density of the electric charge or to a change in the permittivity of the medium. Experiments have shown that alterations in the effective thickness of the double layer are primarily responsible for changes in 5 and that changes in charge density are less significant. The above discussion and the properties of the double layer were based on the assumption that the surface of the solid phase is plane. If the solid phase is considered as a collection of spherical particles of a radius r , the equation [ = u 6r (re tie)-' for the electrokinetic potential is obtained. For large particles (colloidal particles), r %= 6 and the equation can be simplified to

-

g

= 0.6.e-I

+

(1.15)

PROCESSES IN ELECTROPHORETIC SYSTEMS The addition of a sample at a certain location in the separation medium usually alters the equilibrium of the medium. Different values of the chemical potential at different locations in the system are then responsible for the diffusion of particles. However, the electric field applied to the electrophoretic system is the main cause of imbalance and causes the migration of charged particles in the solution. As the electric current passes across the phase boundary (e.g., the surface of electrodes), electrode reactions (i.e., electrolysis) proceed. In addition to the migration of charged particles by diffusion, convection, i.e., transport of components caused by mechanical forces and heat conduction among particles, also occurs.

9

PROCESSES IN ELECTROPHORETIC SYSTEMS

Transport processes in solutions of electrolytes The transport of each component of a solution can be described by using the basic equation of transport phenomena, i.e., the continuity equation. For the i , zth component (1.16) +

where J i . , is the substance flow of the i , zth component, which is determined by the product c ~ , v'~ (. 5is the vector of the rate of movement of a component at a given location in the solution). The substance flow can also be represented by (1.17) where the terms on the right-hand side of the equation express diffusion, migration, convection and thermal components of the compound flow. Dijjusion

The diffusion term of the compound flow (j;,,)&,,is worth consideration if the chemical potential of the i , zth compound is different at different locations in the solution. The diffusion term follows Fick's law**. +

(J.i , z )dlff . =

- - ci,, * Di,,

*

(RT)-'

- grad

(1.18)

where Di,z is the diffusion coefficient of the i, zth component and p i C 2is the chemical potential of this component. In dilute solutions the latter can be expressed as pi,z =

8 , z

(1.19)

+ R T l n ci,z

is the standard chemical potential. When terms from eqn. 1.19 are substituted where in eqn. 1.18, the following equation is obtained:

-

( J i . z ) d i f f = -Di,z

.grad Ci,z

(1.20)

The diffusion coefficient is directly proportional t o temperature and indirectly proportional to the viscosity, v, of the solution. For a spherical particle of radius D,,, = R T * ( 6 7 ~ N ~ ~ r ~ , ~ ) - '

*

4

6 ~ , aJ + -2 6~ av

If J is a vector, div J = -

(1.21)

-

aJ, +. The divergence of the vector (div J ) is a scalar indepen-

az dent of the system of axes. It indicates the flow of the vector field through unit volume. The divergence of the mobility vector expresses a decrease in the substance in unit volume per unit time.

** Under the assumption

that in the scalar function grad

p

ar

= -ax

-

.

I

+ ?!

ay

-

j

ar +- k, where i , j and k az

are unit vectors having directions of positive axes of coordinates and magnitude of 1. The gradient of the scalar function (grad u) is a vector vertical at each point of the region with respect to the surface area of the scalar field. It has a direction of the highest increase and its coordinates are partial derivatives according to x, y and z. It is independent of the system of coordinates.

10

THEORY OF ELECTROMIGRATION PROCESSES

where NA is Avogadro's constant and the radius ri,z is the hydrodynamic radius of the particles; compared with the radius of ions, the latter includes the hydration envelope. Migration -+

The migration term of the substance flow, (Ji,r)mie,characterizes movement induced by an outer force, the intensity of the electric field, E = - grad cp (where cp is the electric potential). When the electric field acts on partiges of the ith species with charge z , the particles move with a velocity proportional to E. Then,

-

(G.z)mig =

SP

2 . ui,z

E

(1.22)

The term sgn z (which is defined in such a way that sgn z = 1 for z > 0, sgn z = - 1 for < 0 and sgn z = 0 for z =,O) respects the fact that a positively charged particle moves in the direction of the vector E and a negatively charged particle in the opposite direction. Ui,z is the mobility. The physical importance of the mobility Ui,, follows from eqn. 1.22. The velocity of a charged particle of the i, zth component in the electric field of unit potential gradient in a given electrolytic system (characterized by the ionic strength) is involved here. The mobility and diffusion coefficients are related by the equation

z

Ui,z = I z ( Di,z F * (RT)-'

(1.23)

Eqn. 1.21 also gives an equation for the mobility of a spherical particle with a radius ri.2

Ui,z = IzI e ( 6 7 r ~ r ~ , ~ ) - '

(1.24)

When the mobility of a colloidal particle is considered, the charge of which is determined by eqn. 1 . 1 5 , the following equation is applicable: Ucol =

E

(6n~)-'

(1.25)

If the colloidal particle is not spherical, a numerical factor is required in the denominator, depending on the shape of the particle, and vanes between 4 and 8 . The dependence of the mobility on the reciprocal of the coefficient of viscosity can be deduced from eqns. 124 and 1.25. This coefficient depends on temperature according to the Arrhenius where B7 and BB are constants, Mobilities equation: ( Q ) - ~= B7 exp (- B8 * T'), depend on temperature in a similar way. However, square root series are usually used for the temperature dependence of mobilities:

U,

= Ula[l

+ k1(7 - 18) + k*(T - 18)2]

(1.26)

where T is temperature ("C) and k l and k2 are constants for various ions, which are given in tables [2]. The mobility, Q Z , of each component (also called the actual mobility) is not a constant, characteristic of a given component, but depends on the concentrations of all kinds of ions present in the solution (on its ionic strength). In order to explain the

11

PROCESSES IN ELECTROPHORETIC SYSTEMS

effect of intra-ionic forces on the mobility of particles, the assumption of an ionic atmosphere, with a decreasing movement of particles in the electric field, was introduced. The effect of intra-ionic forces decreases on dilution of the solution (ions being removed). Tlus effect disappears only in solutions in which the concentration of all dissolved components is approximately zero. In these solutions, the distances among ions are so great that the intra-ionic forces are negligible. The mobility in these solutions [absolute or limited mobility ( U i , z ) O,]independent of concentration, is a characteristic constant of a given component. The Debye-Huckel-Onsager theory is the oldest theory relating mobility with concentration. According t o this theory ui,z

=

(q,zI0

- [ B g ( U i , z ) O - B101(Ic)1’2

[I

+ ri,zWIc)112 I -1

(1 27)

where B , B9 and Blo are terms that can be determined by calculation. Eqn. 1.27 holds for dilute solutions of univalent symmetrical electrolytes. A unified theory characterizing the effect of intra-ionic forces even in more concentrated mixtures of polyvalent nonsymmetrical strong electrolytes formulated by Onsager and Fuoss [3] leads to more coniplex relations. The mobilities Ui,, and (Ui,*)’ defined so far were concerned only with individual components. For a substance composed of several components, it was found useful to introduce a further characteristic, the effective mobility [4] of a substance, (Ui)eff.The following facts substantiate its introduction: (a) the substance is composed of several components that are in equilibrium; (b) the components exhibit different actual mobilities; (c) individual components, however, cannot be separated electrophoretically and the substance moves as a whole in the electric field; and (d) individual components of the substance contribute to its resulting velocity. The contribution of each component is proportional to its relative concentration in the substance. Thus, (Ui)effis defined by (1.28) When the molar fractionjri,z is expressed by means of eqn. 1.5, the following equation is obtained:

C ui,z 1+

c

( 1 29)

To characterize the movement of substances in a porous medium, the macroscopic mobility of the substance, (Ui)mam, and the relative mobility, (Ui)rel,will also be introduced using eqns. 3.4 and 3.9, respectively. Conductivity of electrolytes In addition to the mobility, U i , ,, ionic conductivities Xi,z defined by

12 Xi.2

THEORY OF ELECTROMIGRATION PROCESSES =

I4 O F *ui,t

(1.30)

can be used to characterize individual ions. Ionic conductivities are used particularly for the characterization of the conductivity of the electrolyte as a whole. By this means it is possible to define the molar conductivity of the electrolyte, A, and the specific conductivity, K , using the equations (1.31) and (1.32) The units of molar and specific conductivity are A = S mz mole-' and K = S m-l; for practical purposes they are of the tabulated [2] in S * cmz mole-' and S cm-', respectively.

Convection The convection term of the substance flow characterizes the movement of the solution at a given location: -+ (Ji,z)wnv

-9

=

ci,z * uwnv

(1.33)

In electrophoretic systems fluxes induced by pressure differences at both ends of the separation column, by capillary forces in the porous medium or by the electroosmotic flux are involved. The hydrodynamic flow induced by pressure differences at the ends of the cylindrical column has a constant velocity along the whole column and can be calculated from Poiseuille's equation. However, owing to internal resistance, a velocity gradient originates in the tube, The final profile shown schematically in Fig. 1.2 is the result. The distribution of velocities is different in the case that the hydrodynamic flow is brought about by capillary forces in the stabilizing porous medium. This case will be discussed later.

Electroosmotic flow The electroosmotic flow originates when an electric field is applied to the system. Under the influence of the field the spatial charge of the diffusion part of the electric double layer moves to the oppositely charged electrode, and a one-sided flow of ions originates near to the walls. The whole solution then moves together with these ions. Near to the boundary (walls of the separation columns), a velocity gradient of the liquid originates. The flow velocity is zero at the boundary, increases inside the solution and can reaches a maximum at a certain, very small distance (where = 0). This velocity, GbS, be expressed as +

+

v., = E * E * ~ * Q - '

(1.34)

Also the remainder of the liquid in the column moves with this velocity. In this way the electroosmotic flow differs from the hydrodynamic flow (Fig. 1.2).

13

PROCESSES IN ELECTROPHORETIC SYSTEMS

C

l

Fig. 1.2. Schematic representation of velocity gradient for (a) hydrodynamic flow;(b) electroosmotic flow in open column; (c) electroosmotic flow in closed column.

A different distribution occurs when the electrode spaces are closed. The one-sided flux of ions along the walls of the column then carries along a portion of the attached solution and moves, together with it, to the closed end of the column, where it turns and flows in the opposite direction through the middle of the column.

Heat conduction and heat flow It is a characteristic feature of all electromigration separation methods that the Joule heat originates during the passage of an electric current through the solution. The system is not isothermal; the distribution of the temperature T is a function of spatial coordinates and time. This case is described by the continuity equation:

aT

-t

- -div.F+q*

(1.35)

cs'ps.at -

-+

- -

. F i s the heat flow determined by the equation ,F= k, .grad T + 5. T c, p , , where k,, c, and p s are the heat conduction, specific heat capacity and density of the solution, respectively, and c i s the velocity of the liquid flow at a given location; q* is generally the change of heat density in the solution. In electrophoretic systems this change is determined by both the generated Joule heat, q;, and the dissipation of heat over the surface -t

14

THEORY OF ELECTROMIGRATION PROCESSES

of the separation column, 4: ;q* = q; - 4 ; . The Joule heat*, q j , generated when a current I passes through a resistance R for time t is given by = RPt

qj

(1.36)

When the resistance is expressed by means of specific conductivity, the equation (KS)-' is obtained. For the temperature balance the heat flow generated by unit volume, q; ,is required :

qj = LZ't

-

.

q; = q j (t . S .

q - 1

=

p. (#)-'

(1.37)

Heat is usually dissipated from the liquid phase through the walls of the separation column. The amount of heat, q k , passing during time t through an area P a n d thickness d is proportional to the heat conductivity, k t , of the solid phase and to the difference between the temperatures ( T - To)of the two surfaces of the plate; q k = k, P t(T - To)d-', For the heat flow, q i coming from the unit volume we have

-

~

4; = q k * ( t ' S * L ) - ' = k , ' ( T - T o ) ' ( P * * d ) - '

(1.38)

P* is the fraction of the volume and surface area of the liquid phase. Thus, for a cylindrical separation column with an inner diameter ro and length L , the following approximate equation is obtained (when the inner and outer surface areas of the column are set identical) :

P*

=

L S (P)-' = nriL (27rroL)-' = 0.5 ro

A temperature gradient is responsible for the origination of thermal flow. In the disarranged thermal movement the flow towards a lower temperature prevails over the opposite flow, so that particles are transferred to sites of lower temperature. The velocity of the thermal flow is expressed by

(g,,)therm = -Dtz grad T

(1.39)

where DEZ is the thermal diffusion coefficient of the i , zth component, so that for the thermal term of the substance flow the equation (1.40)

Transport phenomena in stabilizing media

The undesirable effects of certain transport processes can be limited, e.g., by increasing the viscosity of the basic electrolytic system, by forming a density gradient in the electrolyte, by using some dynamic methods or by performing the separation in a suitable stabilizing medium (carrier). Most gels, compact porous carriers with an inner capillary microstructure, columns with suitable loose contents and special capillary columns are examples of stabilizing media.

The unit of energy and of heat is the Joule (J). Heat is also often expressed in calories; 1 cal = 4.18 J.

15

PROCESSES IN ELECTROPHORETIC SYSTEMS

Spatial effects of the inner structure of capillary systems The microstructure of the capillary systems is usually highly complex, the pores being irregular with different curvatures. A number of schemes have been formulated for the theoretical expression. According to the simplest assumption [ 5 ] , pores in a solid carrier are curved and therefore the path followed by molecules during the migration is in fact longer than that which is measured macroscopically. The correction factor, known as the tortuosity factory [ 5 ] ,y, is defined by the ratio of length of the porous medium, L , and the real length of the pores, I:

L.l-1 (1.41) The factor y can be determined experimentally, e.g., on the basis of conductivity measurements. It is also used to correct the free diameter of capillaries in a porous medium, s, with the free diameter of the porous medium, S, and with the total crosssection of the porous medium, S t : y

z

s = ys = y5st

where 5 is the porosity of the medium, defined by S = 5St. Correction factors defined in a different way are described in the literature [&lo], viz., y‘ = (1 ~ - 1 ) zy” ; = (I L-’)* - 1 or y”’ = ( I ~ - ‘ ) 2* 5-1. Correction by means of the tortuosity factor is applicable only when the diameter of all capillaries is larger than the diameter of any of the moving particles. When the hydrodynamic diameters of migrating particles are comparable with those of the pores, or when in the porous medium the pores are distributed according to their radii, more complex schemes have to be applied. These schemes are based on the fact that a migrating particle can travel only through pores with a radius greater than its own radius. Porous materials with particles of different sizes influence the path travelled by the moving particles t o different extents.

Distribution function and the constant RE During their transport across a porous medium, a distribution of separated substances between the solid (stationary) and liquid phases occurs. As a result, the velocity of movement of individual substances in the porous medium decreases in comparison with that in the electrolytic system. This deceleration can be characterized by R E ,defined as the fraction of the velocity of migration of the ith substance in the porous medium relative to the velocity of migration ($)mig of the same substance in the electrolytic system. The velocity ( K [usually if (UJefr > (Uel)eft] or

* The “front” refers to the concentration profile of the ith substance for which ( a q / a t ) , > 0; the “back” boundary is characterized by the inequality (aci/at), < 0.

25

ZONE ELECTROPHORESIS

I

I

x-

-

x-

-

Fig. 2.1. Schematic representation of the course of separation of three substances, A , B and C, by zone electrophoresis for (Ug)eff > (LIAIeff > (Uc),ff:(1) at the beginning of the experiment at zero time (i = t o ) ;(2) at time t l ( t l > t o ) ;(3) at time fZ(f1 > t , ) . Substance A ( H ) forms a symmetricalzone, forms a zone with a focused front and C(H) forms a zone with a focused end. The solid line represents the total concentration.

B(a)

K, < K [usually if (UJeir < (LleJeff]. In the former instance the potential gradient is higher at the site of the zone than outside it. This results in a situation where the back boundary of the zone becomes sharpened (ions that had been delayed enter a region of a higher potential gradient and catch up the zone), whereas the front boundary becomes extended (ions that had outrun the zone move to a region of a higher potential gradient and become even further removed). When K, < K the opposite situation occurs (the front boundary becomes sharpened and the back boundary becomes tailed). In a similar manner t o the distribution of the potential, the shape of the zones is also influenced by the participation of the ith substance in adsorption equilibria*. A linear

A detailed theoretical description of the effect of adsorption isotherms on the shape of zones was published for chromatographic methods, in which the difference in the adsorption properties of different substances is one of the basic separation principles.

26

CLASSIFICATION OF ELECTROMIGRATION METHODS

adsorption isotherm (a2ri/ac: = 0), in a similar manner to the constant potentialgradient, does not deform the zone, which remains symmetrical during the whole separation. The same “spreading” of both boundaries of symmetrical zones is caused both by participation of all z,i components of the ith substance in ionic equilibria and by diffusion. An adsorption isotherm concave with respect to the concentration axis (azri/ac;< 0, e.g., Langmuir’s isotherm), causes a sharpening of the front boundary; an adsorption isotherm convex to the concentration axis (a2Pi/ac? > 0) causes a sharpening of the back boundary of the zone. The time course of separation is also of importance. It can be seen that in symmetrical zones, the interzonal distances (determined according to the positions of maximal concentrations in the zones) increase. However, the width of the zone increases with increasing time (the zone spreads) and the maximal concentration of a substance in the zone decreases. With dilute samples a large amount of the sample becomes localized in places, where the concentration decreases below the limit of determination and the amount of sample apparently decreases. The widening of asymmetrical zones is even more pronounced, with the result that certain zones are incompletely separated. The characteristics of individual zones are presented in Table 2.2. Zone electrophoresis is analogous to elution chromatography.

TABLE 2.2 CHARACTERISTICS OF ZONE ELECTROPHORESIS ~

Shape of zone Symmetrical Cause of zone shape

KC

=K

and

Sharp front boundary KC

K

01

01

a2ri o ad

ac?

Maximal concentration

ionic equilibria and diffusion decreases permanently during separation

in the zone Zone width

increases permanently during separation

THE MOVING BOUNDARY METHOD

This method resembles closely the zone method, and can be characterized by the following parameters. The separation column is separated into three parts. Both side parts are filled with the same basic electrolytic system and the middle part contains, in addition to the same electrolytic system, a mixture of substances to be separated. The length of the middle part is usually comparable to the lengths of the side parts and, as a result, the

ISOTACHOPHORESIS

21

mixture cannot be separated into independent zones in a separation column of a given length. When the whole mixed zone is moved, only a gradual separation of substances from the front boundary occurs: the most mobile substance moves first, followed by a mixture of two most mobile substances, then a mixture of three most mobile substances, etc. The slowest substances behind the boundary of the mixed zone are delayed: the slowest substance comes last, preceded by a mixture of two slowest substances, etc. The solution of the equations for electrophoretic transport in the moving boundary method leads to a general equation for the regulating function (also known as the Kohlrausch function) for each location in the separation column:

which makes it possible to determine concentration profiles of individual substances related t o time and positional coordinates. The course of separation when adsorption is not involved and when all substances of the separated mixture migrate in the same direction is illustrated schematically in Fig. 2.2. In addition to the concentration profiles of individual substances and derivative curves, Fig. 2.2 also shows the distribution of the potential, which at each instant is proportional to the concentration distribution of the substances t o be separated. Also in this instance (if K , > K) a non-uniform distribution of the potential gradient causes desharpening and sharpening of the front and back boundaries respectively (when K, < K , the front and back boundaries would be sharpened and desharpened, respectively). The shape of the boundary is also influenced by diffusion and convection flows. In addition, flows brought about by the radial temperature gradient occur when using electrolytes with a higher ionic strength. The moving boundary method is analogous to the frontal chromatographic method. This analogy is particularly pronounced in arrangements of the electrophoretic experiment such that only “front” boundaries are collected (substances pass to the separation column from the reserve space) or only “back” boundaries are collected (substances from the separation column present there at the beginning are obtained). In this arrangement, the concentration profiles and distribution of the potential can differ from those shown in Fig. 2.2, as in this instance the basic electrolytic system need not be in the space in which the mixture of substances to be separated is localized.

ISOTACHOPHORESIS It is a characteristic feature of this method that it is not possible to separate simultaneously substances that carry both positive and negative charges. The separation of anions will be described here (analogous conditions are used for the separation of cations). The whole separation compartment is divided into three unequal parts. One part of the anodic compartment and the separation column is filled with the leading electrolyte. A second part is formed by a compartment into which a mixture of substances to be separated is introduced. The third part represents the compartment filled with a terminating electrolyte (the cathode compartment in this instance). The leading electrolyte contains anions with an effective mobility higher than that of any of the anions in the

28

CLASSIFICATION OF ELECTROMIGRATION METHODS

I

4 I

I

c

A+B+C

X -

I

2

-x

3

A+0

I

A

x-

dc

dx 4

I

I

I

Fig. 2.2. Schematic representation of the course of separation of three substances by the moving boundary method: (1) at the beginning of the experiment ( t o = 0); (2) at time t , ( t , > t o ) ;( 3 , 4 and 5 ) at time t , ( t , > t i ) . The solid tine designates the total concentration. Substances: R,A ; B; C; basic electrolyte.

m,

m, m,

mixture and the cations, the buffering capacity of which can be utilized. The terminating electrolyte contains an anion with an effective mobility lower than that of any anion in the mixture to be separated. Cations of the terminating electrolyte are not important for the separation. On applying an electric field, separation proceeds until a steady state is established. This steady state is characterized by the fact that individual substances are separated, according to their effective mobilities, into independent sharp, yet close zones. The so-called separated boundaries are localized among the zones. The separated substance is localized only at one side of the boundary. This steady state, during which

FOCUSING METHODS

29

all substances (zones, zone boundaries) move with the same velocity (hence the name of the method) is characteristic of isotachophoresis. Prior to establishment of the steady state, individual zones are also mutually separated; however, in addition t o zones that contain only one of the separated substances (pure zones), zones with more separated substances (mixed zones) are present. The solution of the equations of the electrophoretic separation for the steady isotachophoretic state leads t o certain important conclusions: (a) The potential gradient in the ith zone is determined by the effective mobility of anions in this zone, as

v

=

(Ul)eff- E l = (U2)eff*E2= (Ui)eff.Ei = constant

(2.3)

where ZI is the velocity of movement of zone boundaries. (b) The concentration of any separated anion B with charge zB is determined by the concentration of the leading anion A (with charge zd,by the mobilities of both anions and by the mobility of a common counter ion C. This relationship is expressed by means of the Kohlrausch regulating function in the form

This feature of isotachophoresis is particularly important for dilute solutions, which are “concentrated” to a concentration corresponding t o eqn. 2.4, which represents a considerable difference to zone electrophoresis. The separation of three substances is illustrated schematically in Fig. 2.3. In addition to the concentration profiles of individual substances, the distribution of the potential gradient is also presented. After establishment of the steady state, the distribution has a characteristic stepwise course. In isotachophoresis the self-sharpening effect of the electric field on all zone boundaries is of considerable importance. (When an ion is delayed and therefore enters the following zone with a higher potential gradient, its velocity increases, the ion catches up “its” zone and enters again. On the other hand, when an ion outruns “its” zone, it moves to a zone with a lower potential gradient and its velocity therefore decreases and the ion is overrun by “its” zone.) Diffusion flows and flows brought about by pressure or temperature gradients act against the self-sharpening effect. The shape of zone boundaries can also be influenced by phase equilibria. Isotachophoresis is analogous to displacement chromatography.

FOCUSING METHODS The methods mentioned above are based on the fact that in the substance flow every substance is characterized by a non-zero migration term, oriented in one direction during the whole separation. Also, the substance flow of all components is non-zero at any location in the separation column. In methods in which the substance flow of each substance decreases from a maximal positive value at one end of the separation column, through a zero value to a maximal value at the other end of the column, the course and particularly the result of the separation are completely different. A number of procedures that make it possible to demonstrate in practice this

CLASSIFICATION OF ELECTROMIGRATION METHODS

L ct

-X

I L

-x

Fig. 2.3. Schematic representation of the separation of three substances by isotachophoresis: (1) at the beginning of the experiment; (2) beginning of separation (mixed zone of B and C still exists); (3) steady state; (4) distribution of the potential gradient. leading electrolyte (L); 81, terminal electrolyte (T); H,substance A; substance B; substance C. The solid line represents the total concentration.

m,

a,

m,

relationship between the substance flow and the positional coordinate have been described. Two methods that differ in the way in which they bring about the required course of the substance flow, isoelectric focusing and electrorheophoresis, will be described here as examples. The former method utilizes the relationship between the effective mobility and the

0

-

.(pH)A

B

Fig. 2.4. Dependence of the velocity, i?, on thesstance x in the column. (A) g i s determined by the U e f f * E )a; linear gradient of pH occurs in the column. effective velocity of the ampholyte=;( the migration velocity is constant. (B) G'= +

zmie

Electrorheophoresis utilizes the dependence of the convection flow on the spatial coordinate and can be used for any substance. When the convection flow is caused by a sucking flow, the velocity of this flow is characterized by eqn. 1.48 and the relationship between the substance flow and distance has a course as shown schematically in Fig. 2.4B, provided that the value of the migration term is constant. The course of separation of three substances in an experimental arrangement with the above dependence of the substance flow on distance is shown schematically in Fig. 2.5. It can be seen that the established dynamic equilibrium rather than the course of separation plays an important role in this experimental arrangement. Individual substances are focused at locations at which values of the substance flow of any given substance are zero, irrespective of the starting point and width of the initial mixed zone of the sample. Isoelectric focusing This technique is usually used for the separation of ampholytes, mostly macromolecular. The separation compartment is formed by three parts: two electrode compartments and the separation column. A stable gradient of pH values, limited by minimal and maximal pH values in the anode and cathode spaces, respectively, is formed in the column. In this gradient ampholytes, the pHis,, of which is within the range pHmi, < < pHiso < pH,, can be focused. When a mixture of ampholytes is applied at an arbitrary location in the column and an electric field is applied to the system, the ampholytes

32

CLASSIFICATION OF ELECTROMIGRATION METHODS

X-

X-

X-

A

X-

B

Fig. 2.5. Schematic representation of the separation of three components by the focusing method: (A) mixture applied to the middle part of the column; (B) mixture applied to the whole column. ( 1 ) Beginning of the experiment; (2) focusing occurs; (3) end of focusing. The solid line represents the and H, concentrations of individual substances. total concentration.

a,

begin to migrate either towards the cathode or the anode, according to their effective mobilities. The velocity of movement of all substances decreases during the separation. After a certain time, a dynamic equilibrium is established, which is characterized by the fact that each ampholyte has already moved to a position at which its effective mobility is zero, i.e., to a position where the pH is identical with the pHis,, of a given ampholyte. A similar dependence of Ueffon pH also occurs with certain complex substances. Hence these substances can also be separated as mentioned above. A stable pH gradient in the column is essential for isoelectric focusing and is usually produced as described below. The column is filled with a mixture of special ampholytes, the so-called ampholyte carrier. Each of the ampholytes in this mixture must have a certain buffering capacity near to its isoelectric point, and adequate conductivity and solubility. The mixture should contain such ampholytes in relative proportions such as to cover the required pH range by its pHiso values. The pH gradient is first formed as a result of different pH values in the anode compartment, the column and the cathode compartment. As the electric current passes through the column, the substances of the

33

COMBINED METHODS

ampholyte carrier, originally homogeneously distributed along the whole column, begin to move according t o their pHis,, values as a result of isoelectric focusing. The distribution originating in this way, in addition t o the buffering capacity of individual components of the ampholyte carrier, secure and stabilize the required pH gradient.

Electrorheophoresis A gradient of hydrodynamic flow of the electrolyte occurs as a result of sucking flows caused by evaporation of the electrolyte from the surface of the stabilizing porous medium. The velocity of a substance at any location X in the column (usually a paper strip) is determined by the sum of the migration velocity (eqns. 1.22 and 1.28) and the velocity of the electrolyte flow (eqn. 1.48). Thus, ( v , ) ~ = sgnz*(UJeti-E+ rn*(X-OSL).(pSd)-'

(2.5)

Focusing of the i t h substance occurs at a site Xi:

It is apparent that in a given gradient of hydrodynamic flow it is possible to focus substances the migration velocity of which is lower than velocity of the electrolyte flow. Also in this method the formation of a stable velocity gradient is a necessary condition, and can be achieved by using a suitable composition of the basic electrolyte system, in which all substances have approximately the same volatility. In the opposite case, volatile substances (solvent) are preferentially evaporated, so that the electrolyte concentrates in the porous carrier. The evaporation of the electrolyte from the whole surface and its supplementation from the electrode spaces leads to concentration of the electrolyte in the direction of the point of zero hydrodynamic flow. The formation of the concentration gradient is also reflected in a non-uniform distribution of the potential gradient (a decrease in the middle and an increase at the sides of the porous carrier). The increase in the potential gradient at both ends of the carrier can cause an increase in migration velocity such that a substance focused at the beginning, near one of the ends, starts to move towards it and finally leaves the carrier completely. Characteristic features of the above electrophoretic methods are compared in Table 2.3.

COMBINED METHODS Of combined methods, which have increased in importance in recent years, we can consider as a representative example of the combination of two different methods the so-called disc electrophoresis and immunoelectrophoresis.

Disc electrophoresis In this method the experimental arrangement (which also gave the name to the discontinuous method) allows the consective application of isotachophoretic and zone principles .

h charges s only a zone of the fastest substance in the basic electrolyte; other zones are mixed

basically particles with charges of both signs, practically particles with charges of one sign

frontal

Moving boundary

particles with charges of only one sign

displacement

Isotachophoresis

IC ELECTROPHORETIC METHODS

ividual n the lyte

Focusing methods

particles with charges of both signs each zone contains only a given substance

both boundaries sharp

remains constant during separation

a l l boundaries sharp

decreases during focusing; remains constant after establishment of dynamic equilibrium

increases during focusing, remains constant after establishment of dynamic equilibrium

increases or decreases during separation; remains constant after establishment of the steady state

rmanently ation

one sharp, the other spread (second boundary need not occur) increases permanently during separation

different for different substances; remains constant during separation

after establishment of the steady state identical for all substances and constant with time

after establishment of the steady state each zone contains only one separated substance can both increase and decrease during separation; after establishment of the steady state it is constant and is determined by a regulating function

he other th spread

rmanently ation

different can change ation

during focusing it changes with distance (to a different extent for different substances); after establishment of dynamic equilibrium the velocity of all separated substances is zero

COMBINED METHODS

35

The separation column, packed with a suitable (usually polyacrylamide) gel, is divided into two parts: the focusing part and the separation part. In each of these parts there is a different separation medium and different electrophoretic arrangements apply. The focusing part of the column is shorter, and is packed with dilute gel, the internal structure of which has virtually no effect on the mobility of individual substances, which is thus determined solely by the effective mobilities of the substances in question. The separation part of the column is considerably longer, the gel density is higher and, with the use of a higher content of cross-linking agent, its internal structure must be considered during the separation. Equally, the composition of the electrolyte system and its pH are different in the separation and focusing parts. An appropriate choice of the electrolyte systems in the cathodic and anodic compartment and in the focusing and separation parts of the column can result into concentration and ordering of the individual substances into narrow, closely contacted zones on the border between the dilute and solid gel by an isotachophoretic mechanism. In the separation part the substances are separated into individual zones according to the laws of zone electrophoresis in solid carriers (the sieving effect of the polyacrylamide gel occurs in this stage). Substances are detected by staining after the separation is completed. The advantage of this method is that optimal starting conditions for the zone electrophoretic separation are produced by the preceding isotachophoresis resulting in a concentrated narrow band. Immunoelectrophoresis This is an electrophoretic procedure that exploits the highly sensitive and highly specific immunochemical precipitation of the antigen-antibody system for detection. If both parts of the complex, e.g., antigen and antibody, come into contact at optimal concentrations, characteristic precipitation lines are formed (the complex, however, is soluble in the excess of both the antigen and antibody). The shape of the precipitation lines is also determined by the way in which both antigen and antibody come into contact. For this purpose differently oriented diffusion and migration flows are made use of. In the classical version of the method, the mixture of antigens is first separated in agarose gel by the zone electrophoresis. Serum antigens can be considered as an example. Then a suitable antibody (e.g., rabbit or horse polyspecific antiserum) is applied to a channel located parallel to the direction of the separation in the gel plate, so that the conditions for the occurrence of diffusion flows are thus fulfilled. Antigens diffuse almost radially around the locations that they have reached during the zone electrophoretic separation. Antibodies migrate laterally with respect to the channel, e.g., to meet the antigens. After a period of time (which is usually long, of the order of days, because of the slowness of the diffusion flows), a series of radial precipitation lines are formed in places where both antibodies and antigens have met (Fig. 2.6A). In other modifications of immunoelectrophoresis, the slow diffusion flow of antigens against antibodies is replaced with a much faster migrational flow. A method analogous to the classical version is the so-called “crossed electrophoresis”. In the first stage (separation of antigens by zone electrophoresis on agarose gel) the method is coincident with the classical version. Then a new separation medium is formed in such a way that another

36

CLASSIFICATION OF ELECTROMIGRATION METHODS

A

Fig.2.6. Schematic representation of the immunoa ctropherogram.( Classical arrangement: 1location of antigens after electrophoretic separation;6 , antibody; n, precipitation lines. (B) “Laurell rocket” method; 1-3, comparative antigens of known concentration; 4-8, unknown samples; n, precipitation lines; h , height of precipitation peak.

layer of agarose gel is added t o the strip of agarose gel with separated antigens. This part, in addition to the electrolyte system, also contains the polyspecific antiserum. The composition and pH of the electrolyte are such as to make the effective mobility of antibodies zero. If an electric field is imposed on the system in a direction perpendicular to the direction of separation, antigens migrate into the antibody-containing media. Equally in this case in areas where antigens meet antibodies at optimal concentrations characteristic precipitation lines are formed. The migration of antigens against non-migrating antibody is also made use of in the “Laurell rocket” method, which is used in clinical biochemistry for the determination of a single antigen. In this instance agarose gel (with the exception of a narrow strip at the edge of the plate where the samples are applied) is saturated with the basic electrolyte together with the monospecific antibody which has a zero effective mobility. After a series of samples has been applied (at least three are usually employed as standards of known concentration and are used for designing the calibration line), the electric field is applied and most substances from individual samples pass through the agarose carrier. The estimated antigens, however, form characteristic precipitation lines (Fig. 2.6B). The size (height) of the precipitation arcs is proportional to the antigen concentration in the sample. A more generally applicable method is the “fused rocket” method. The basic

REFERENCES

37

electrolyte system (which penetrates the agarose gel plate with the exception of a narrow strip at the ed@ where the samples are applied) also contains the polyspecific antiserum with a zero effective mobility of antibodies. To this plate are applied a series of samples that have been obtained as fractions in a prior separation of the original mixture (e.g., by chromatography, preparative electrophoresis, centrifugation or isotachophoresis). The precipitation lines obtained after an immunoelectrophoretic separation can be intensified by staining, using dyes suitable for the particular type of substances separated. Sometimes two or more dyes are used for different species present in the sample.

REFERENCES More detailed information regarding the topics discussed throughout this chapter can be found in more specialized papers or books devoted to electrophoretic and related techniques. Some of these are Listed below. 1

M. Bier (Editor), Electrophoresis. Theory, Methods and Applications, Academic Press, New York,

2

R. J. Block, E. L. Durrum and G. Zweig, A Manual of Paper Chromatography and Paper Electrophoresis, New York, 1955. H. Bloemendal, Zone Electrophoresis in Blocks and Columns, Elsevier, Amsterdam, 1963. H. G. Cassidy, Adsorption and Chromatography,Interscience, New York, 1951. H. G. Cassidy, Fundamentals of Chromatography,Interscience, New York, 1957. L. P. Cawley, Electrophoresis and Immunoelectrophoresis, Little and Brown, Boston, Mass., 1969. H. J . McDonald, R. J. Lappe, E. Marbach, R. H. Spitzer and M. C. Urbin,lonography, Electrophoresis in Stabilized Media, Chicago, Ill:, 1955. F. M. Everaerts, J . L. Beckers and T. P. E. M. Verheggen, Isotachophoresis;Theory, Instrumentation and Applications, Elsevier, Amsterdam, 1976. A. H. Gordon, Electrophoresis of Proteins in Polyacrylamide and Starch Gels, North-Holland, Amsterdam, 1969. C. J . Gidding, Dynamic of Chromatography.Part 1 . Principlesand Theog', Marcel Dekker, New York, 1965. S. Hjerten, Free Zone Electrophoresis,Almquist and Wiksell, Uppsala, 1967. M. Lederer, Introduction to Paper Electrophoresis, Elsevier, Amsterdam, 1955. M. Lederer (Editor), ChromatographicReviews (Progress in Chromatography.Electrophoresis and Related Methods), Vol. 1-3, Elsevier, Amsterdam, 1960. H. R. Maurer, Disc-Elektrophorese,Theorie und Praxis der diskontinuierlichen Polyakrylamidgel Elektrophorese, De Gruyter, Berlin, 1968. G. Sohay, Theoretische Grundlagender Gaschromatographie,VEB Deutscher Verlag der Wissenschaften, Berlin, 1960. D. J. Shaw, Electrophoresis, Academic Press, New York, 1969. J . Smith (Editor), Chromatographicand Electrophoretic Techniques. Vol. 2. Zone Electrophoresis, Interscience, New York, 1968. K. E. Stensio and G. Ekedahl, Chromatography and Electrophoresis, Norstedt, Stockholm, 1969. H. Svensson, Electrophoresis by the Moving Boundary Methods. A Theoretical and Experimental Study, Almquist and Wiksells, Stockholm, 1946. Ch. Wunderly, PIinciples and Applications of Paper Electrophoresis, Elsevier, Amsterdam, 1961. G. Zweig and J. R. Whitaker, Paper Chromatographyand Electrophoresis, Academic Press, New York, 1967.

1959.

3 4 5 6 7 8.

9 10 11 12 13. 14 15 16 17 18 19 20 21

This Page Intentionally Left Blank

Chapter 3

Evaluation of the results of electrophoretic separations J. V A C ~ K

CONTENTS Introduction .................................................................. Determinationofmobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calculation for zone electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calculation for the moving boundary method. ..................................... Calculation for isotachophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calculation for focusing methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principles of quantitative evaluation. ...............................................

39 40 40 41 41 42 42

INTRODUCTION The objectives of electrophoretic separations may differ. Preparative electrophoresis is used for separation of individual substances from a mixture, and these substances are further identified and determined. When using electromigration methods for analytical purposes, the purpose is to characterize qualitatively and quantitatively the components of a separated mixture. In this instance even electrophoretic methods that do not give a complete separation of individual substances (moving boundary electrophoresis) can be applied. The evaluation of the results of the electrophoretic separation depen:!s coiisiderably on the separation method used and on the method of detection of individual substances. The substances can be detected as follows: (a) the positions of the individual substances (concentration profile) in the electrophoretic system are determined at a given time interval; (b) changes in the concentrations of individual substances at a given location with time (e.g., at the end of the column) are determined according t o changes of a measurable quantity. Mobilities determined from positional and time coordinates of individual substances during their detection are usually used for qualitative examination. Further possibilities for the detection and identification of different substances using various special methods will be described in later chapters. Specific reactions with suitable agents, formation of precipitates (e.g., immunoelectrophoresis, characteristic interactions with the solid phase, characteristic interactions with radiation of different wavelengths and characteristic Properties of certain substances ( e g , radioactivity, enzymatic activity)) can be utilized.

40

EVALUATION OF THE RESULTS OF ELECTROPHORETIC SEPARATIONS

DETERMINATION OF M O B I W The effective mobilities of individual substances are usually obtained by evaluating their positional and time coordinates. The determination of the effective mobility is different for different electrophoretic methods. Calculation for zone electrophoresis The zone method is usually used for separations on carriers. In addition to the migration term of the substance flow, interactions of substances with the carrier and electroosmotic flow are also involved. Individual substances on the electropherogram are characterized by the distance Xi travelled by the substances during time t. For the velocity + oci of the ith substance, eqn. 2.1 applies. It is adjusted (osmotic flow is introduced) to

zOsm+

The term sgn z.(U&ff *Ei t; & = vmi, + = v characterizes the velocity of the substance flow and the term (1 aI':/aci)-' interaction with the carrier. When the quantity ( R E ) l , defined by eqn. 1.44, is used, eqn. 3.1 can be transformed into -t

+

-

-+

(&)eft a Ei @Eli -k %sm * (RE)i

vq =

(3.2)

When an extension of the path of a substance in a porous carrier, characterized by the tortuosity factor 7 , is considered, together with the fact that the intensity of the electric field for zone electrophoresis (p. 24) usually does not depend on the positional coordinates, eqn.l.45, adjusted to = X I - L [7'

sgn z

RE)^ II. t]-'

(3.3)

can be used to express the effective mobility. When values of (R& and y are not known can be calcufrom other measurements, the so-called macroscopic mobility, lated. This parameter can be obtained from the macroscopic velocity in a porous medium, i.e., from the velocity defined by the fraction of the macroscopically measured distance Xi that the ith substances had travelled during time t, with which the substance moves in unit effective potential gradient, i.e., gradient related with the macroscopically measured length of a porous medium L. Thus Edt = U-L-'. It holds that

((li)macn,and the effective mobility, (qleff, are related according to the equation (UiIrn-

=

7'(Y)e&~)i

(3.5)

When the electroosmotic flow is also involved, a corresponding correction has to be made. Osmotic flows are usually determined in such a way that a reference substance that does not migrate in the electric field ( a non-electrolyte) is added to the mixture.

41

DETERMINATION OF MOBlLITY

The velocity of the osmotic flow can be calculated from its position X, at time t :

- -’

X, t

=

Gosm -

(3.6)

(RE)scharacterizes the interaction of the reference substance with the carrier. By combining eqns. 3.2-3.6, the equation (y),,

= [4 -

xs’ (RE)s

*

(RE);‘1

a!d

’ (U * t)-’

(3.7)

is obtained. A substance with an ( R E ) s value close to (RE)i should be chosen as the indicator of the osmotic flow. Eqn. 3.7 is then simplified to the approximate equation

( ~ 1 ~ -(X~-X,).L-(I/-~)-’

(3.8)

Mobility is often expressed with respect to the standard substance selected. It holds that (Cli,s)rel = (I/i)maeK, * [(UstmAmcmI-’

(3.9)

Depending on the standard substance selected, the relative mobility, (Uilrel,can be either higher or lower than unity.

Calculation for the moving boundary method This method is usually applied in free solutions and therefore interactions with the carrier need not be considered. However, in comparison with the previous method, an unequal distribution of the potential gradient must be taken into account. When using the position x i , to which the boundary of the measured i t h substance had travelled during time t , the equation

xi-?-’= (Uj)eff.Ei= (Ui)eff*l*(Ki*S)-l

(3.10)

where K~ is specific conductivity in the measured zone, is obtained. It follows that

(Q)eff =

Xi

-

Ki

‘s

(1-r)-’

(3.1 1)

Calculation for isotachophoresis The calculation for isotachophoresis is performed for the steady state characterized by eqn. 2.4. We have (&),if

=

(UL)eff-EL*E:l = (I/L)eff-~i* =~(UL),ff-hL-h;’ i’

(3.12)

It can be seen that for the calculation one has to know the effective mobility of the and the potential gradients, EL and Ei(and occasionally specific leading ion, (UL)eff, conductivties, K~ and K ~ ) ,of the leading and the ith zone. An isotachopherogram, which is a graphical record of the time dependence of the response of the detector, through whch zones of individual substances pass, is usually examined (Fig. 3.1). EL and Ei(or KL and K ~ are ) proportional to the response of the detector at the time of passage of the zone, i.e., to the “heights” of the zones, hL and hi.The “relative height”, hrel,defined by (hi)rel =

(hi - hL). (hs - hL1-l

(3.13)

42

EVALUATION OF THE RESULTS OF ELECTROPHORETIC SEPARATIONS

i

il

0

I

t

I

1

I

I

I

t

----c

Fig. 3.1. Schematic representation of an isotachopherogram.

where the subscript i indicates the substance being followed, L the leading electrolyte and S a suitable (strong) electrolyte, is also frequently used. When the terminating electrolyte is chosen as the standard, (hi)relis within the range 0-1. The effective mobilities characterize the substances only under given experimental conditions. As the experimental conditions differ in different applications, these data are usually not comparable. It is therefore advantageous also to determine absolute characteristics, i.e., the limiting mobilities of individual components, (Uo,,)o. This can be performed, for instance, by evaluation of isotachophoretic experiments by means of a programme available from the author upon request. Calculation for focusing methods At the end of the separation, the velocity of all substances reaches zero (vcj = 0) and each substance is determined by a characteristic distance Xi. When using isoelectric focusing, it is possible to determine the isoelectric point of any given substance on the basis of this value (provided that the course of the pH gradient is known). If electrorheophoretic experiments are carried out, corresponding effective mobilities are determined. This calculation utilizes eqn. 3.1, supplemented by the velocity of the sucking flow (eqn. 1.48), so that the equation +

vci = [sgn z.(Ui)eff *Ei+ v,,

-+

-n2*/Xi - O.SL/pdV]

(3.14)

is obtained. The rearrangement of this equation leads to that used to calculate (Ui)eff: sgnz-(LQeff =

posm -m*(Xi-0.5L)pd~]*L.(U,)-'

(3.15)

PRINCIPLESOF QUANTITATIVE EVALUATION The quantitative determination of substances also depends on the separation method used and the method of detection. Methods used also in chromatography, i.e., direct

43

PRINCIPLES OF QUANTITATIVE EVALUATION

examination of electropherograms or examination of zones after their elution, can be utilized for the evaluation of results obtained by means of, e.g., zone electrophoresis and isoelectric focusing. Direct evaluation is usually possible by photometry, sometimes after a suitable treatment (e.g., staining, treatment of the carrier to make it transparent). Light passing through zones or reflected by them is measured. The relative proportions of components in the mixture can be determined from the curve of the dependence of the light intensity on the positional coordinates (Fig. 3.2), as areas below the curve are proportional to the concentrations of components, When a concentration of one component is known, e.g., the internal standard, absolute concentrations can be determined.

Fig. 3.2. Schematic representation of the dependence of light intensity (I,) on distance (x): evaluation of the separation of substances A, B and C when B and C are incompletely separated. Areas below the curve (B,A; B; C) represent the rehtive concentrations of the substances. The solid line represents the total concentration.

m, m,

The determination of areas is less accurate, particularly when the zones are not perfectly separated. The derivative curve dc/dx versus x (Fig. 2.2), obtained after separation by means of the moving boundary method, is similar and is evaluated in an anologous manner. Also in this instance the areas below this derivative curve are proportional to the concentrations of individual substances. Much simpler is the evaluation of results obtained by the immunoelectrophoretic “Laurel1 rocket” method, although even here we are considering the evaluation of zone electrophoresis. The height of the precipitation curve, hi(Fig. 2.6B), is the concentration of the estimated antigen in a sample. The proportionality constant, K , can be determined for a given arrangement according to the height of the precipitation curve of standards. The quantitative evaluation of isotachopherograms, particularly after separation with the use of constant current, is much simpler. The fact that any zone length is proportional to the total amount of a substance contained in it is utilized. The determination of the ith substance can be performed most simply by comparing its zone length, Li, with the zone length of an internal standard, L,, a known concentration of which, ( c , ) ~ is, added to the sample. For a concentration (& of the ith substance in the sample we have (ci)o = (ci)s * ~i

* ( ~ *Di, s s1-I

(3.16)

where Digsis an experimentally determined relative correction factor. The concentration (cJ0 can also be determined according to the zone length Liby calculation using values of quantities that characterize a zone of the leading electrolyte. These problems of quantitative evaluation will be dealt with in detail in later chapters concerned with special electrophoretic techniques.

This Page Intentionally Left Blank

Chapter 4

Molecular size and shape in electrophoresis Z. DEYL

CONTENTS Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elimination of charge differences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two gels of different concentration (Zwaan's procedure) . . . . . . . . . . . . . . . . . . . . . . The SDS procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conditions for binding SDS to proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations of the SDS procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anomalous behaviour of SDS complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The N-laurylsarcosineprocedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other masking proceduresand urea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Choice of standard series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular size determination by pore limit electrophoresis . . . . . . . . . . . . . . . . . . . . . . . Determination of Stokes radii byrheophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45 49 49 50 52 53 55 58 59 61 62 63 66

INTRODUCTION

In addition to their separation properties, many gel-like supports can also be used for the determination of the molecular weight of an unknown macromolecular solute. Both electrophoretic and chromatographic principles can be made use of. As has been demonstrated in gel chromatography, the separation of a protein depends more on the size and shape of the molecule than on its weight. A close relationship exists between the distribution coefficient and the Einstein-Stokes radius. Applying the equation of restricted diffusion, the distribution coefficient can be defined as a function of the ratio of the pore radius to the molecular radius [ 1 , 2 ] . The molecular charge has no influence on net migration in electrophoresis if the pore and molecular diameters are identical. Then, the molecule is excluded from the gel, i.e., the exclusion limit for a particular protein has been reached. If within a series of diverse proteins no attempt is made to equilibrate the charge differences, then a semi-logarathmic plot of the relative mobility ( R F )versus acrylamide concentration (in polyacrylamide gel) will give a series of almost parallel lines, as shown in Fig. 4.1. If the exclusion limit is defined as the intersection of the retardation line [log (% acrylamide) versus RF] with the ordinate (for RF = 0 the pore radius and molecular radius should be identical) the relationships between exclusion limit and molecular weight and between exclusion limit and EinsteinStokes radius can be drawn (Figs. 4.2 and 4.3). With proteins that have Stokes radii less than 3 nm no retardation occurs below a certain acrylamide concentration (retardation limit). Beyond this limitation the migration velocity depends on charge and the ratio of pore diameter to molecule diameter (retardation).

46

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

I

\

\

Albumin

\ 10

20

40

30

50

60

70

RF

Fig. 4.1. Retardation lines of the four test proteins used.

Fig. 4.2. Relationship between exclusion limit and Einstein-Stokes radius. 1, Insulin; 2, myoglobin; 3, ovalbumin;4, haemoglobin; 5, albumin; 6, malate dehydrogenase; 7, peroxidase; 8, lactate dehydrogenase;9, haptoglobin 1-1 ; 10, coeruloplasmin; 1 1 , catalase; 12, fumarase;

13, a2-macroglobulin.

INTRODUCTION

1000 000

47

I -

01

500000

-

100000

-

50 000

-

im .9 3

-s &

9

010

10

0 14

20 Exclusion limit o/o Acrylamida

30

40

Fig. 4.3. Relationship between exclusion limit and molecular weight. 1, cyI -Macroglobulin; 2, xanthineoxidase; 3, glycerokinase; 4, fumarase; 5, catalase; 6, leucyl-pnaphthylamidase; 7, coeruloplasmin; 8, lactate dehydrogenase; 9,dimer albumin; 10, haptoglobin 1-1 : 12, haemoglohin; 13, albumin; 14, praealbumin; 15, yglohulin; 16, ovalbumin; 17, malate dehydrogenase; 18, peroxidase; 19, a1-antitrypsin; 20, a1-acid glycoprotein; 21, myoglohin.

If the medium in which migration occurs exhibits a sieving effect, then the retardation coefficient, K R ,can be measured. The most direct approach is to observe migration rates in gels with different acrylamide concentrations, as formulated by Ferguson and Wallace [3]: log (ui)macro = log kacro- K R Y P

(4.1)

where (UJmaerois the electrophoretic mobility in a gel with an acrylamide concentration ofy,. According to Hjerten et al. [ 4 ] , y pis the sum of the amounts of acrylamide and bisacrylamide in grams per 100ml of solution. Frequently, relative electrophoretic mobilities ( R F )are measured and plots of R F versus y p (Ferguson plots) are constructed. The slope of these linear relationships is equal to K R and the intercept is (Ui)&acTo(for zero gel concentration). This relationship was derived from theoretical considerations for a particle moving through a gel of linear polymers and has been validated for a large number of proteins [which are used preferentially in the form of sodium dodecyl sulphate (SDS) complexes] (Fig. 4.4).

48

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS I

0.1

I 0

I

2

4

6

8

YP

Fig. 4.4. Typical Ferguson plot of several soluble proteins. A linear relationship is seen except for phosphorylase A ( l l ) , which shows a slight decrease in slope at y p > 6. Each point is the mean of two or three RF measurements. 3, Myoglobin;4, y-globulin, light-chain; 5 , chymotrypsinogen; 6, pepsin; 7, lactic dehydrogenase;8, ovalbumin; 9, glutamic dehydrogenase; 10, bovine serum albumin.

For gels in which the gel fibres are much longer than the separated molecule, the probability of terminal contacts is infinitesimal and only the probability of tangential contact has to be considered [5] , For such gels,

Kg2 = kl(r+k2) (4.2) where K , is the retardation coefficient, I is the geometric mean radius of a macromolecule based on the assumption of zero hydration and a partial specific volume of 0.74 for each protein. Variable hydration will decrease the slope and change the scatter around the line but will not affect the linearity. k l and k2 are constants characteristic for any specified electrophoretic system (k, being related to the thickness of the gel fibres). Eqn. 4.2 predicts that there will be a good'correlation between K R and molecular weight and also predicts that this linearity will be applicable for a wide range of molecular weights*. Comparison of actual plots with varying k2 (gel porosity) values indicates that in gels of larger porosity (4-5 nm) the selectivity would decrease. For this reason, starch gels are not suitable for molecular weight measurements. Eqn. 4.2 also implies that even very small molecules would be distributed according to their molecular weights (sieved) provided that for a zero mean molecular radius the retardation coefficient is still positive (Fig. 4.5).

* In many papers and books molecular weights are expressed in daltons. As molecular weights are relative values, there is no reason to introduce such a unit here and the use of daltons has therefore been avoided throughout this chapter.

ELIMINATION OF CHARGE DIFFERENCES 0.5

-

0.4

-

0.3

-

0.2

-

49

o”!ll 0-

1

0

1

I 2

’ ’ ’ 3

4

5

6 I

’

7

r (nm)

Fig. 4.5. Relationship between the square root of the retardation coefficient (KR)and the molecular radius; r = geometrical mean radius.

ELIMINATION OF CHARGE DIFFERENCES Two gels of different concentration (Zwaan’s procedure) In order to decrease the scatter of data it is necessary to find a suitable means of eliminating charge differences between different separated molecular species. One of the first proposals for eliminating the charge differences was that suggested by Zwaan [ 6 ] ,in which the retardation coefficient is defined as

where (Ui)y and (Ui); are the absolute mobilities in gels of higher (subscript 1) and lower (subscript 2) concentrations. As the effect of the electric charge of the protein is eliminated by the quotient, K L should be dependent on molecular size only relative to the pore sizes of the gels. Thus, 1ogM = k(Kk)+k‘,

(4.3)

where is molecular weight and k and k’,are experimentally determined coefficients. The value of K k will vary from zero to unity as migration in the gel of higher concentration can never be negative or larger than the mobility in the gel of lower concentration.

50

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS protains u n d

log

molecular wohht

Fig. 4.6. Relationship between the logarithm of the molecular weight of some standard proteins and retardation quotients based on single runs in 10 and 4%( X I , 10 and 5% (a), 10 and 6%( o ) , and 10 and 8%(m) gels. E-1, Chicken ovalbumin; H, human haemoglobin A-1 ;B-1, bovine serum albumin, monomer; T, human transferrin; Hp, human haptoglobin 1-1 ; B-2, bovine serum albumin, dimer; L, human lactate dehydrogenasc; B-3, bovine serum albumin, trimer; B-4, bovine serum albumin, tetramer.

It is assumed that polyacrylamide gel is neutral and shows no endoosmosis and absorptive or repulsive effects. The magnitude of k', depends on the size of the protein that cannot penetrate into the gel of highest concentration. If one draws the lines according to eqn. 4.3 for combinations that share the same concentration of the more concentrated gel, these will indeed intersect the abscissa at the same point, as shown in Fig. 4.6. The SDS procedure The requirement of identical mobilities of different proteins and their fragments is nowadays mostly ensured by using surface-active agents such as sodium lauryl sulphate (sodium dodecyl sulphate, SDS). N-Laurylsarcosine or Triton X-100are less popular but can be used with good results. In fact, the presence of these agents results in the formation of strongly negatively charged protein-detergent complexes with almost identical mobilities in free solution, whatever their original charge might be. In addition, the predominating conformation of the protein is a statistical coil, although a choice of conditions can be made so as to preserve the native state of the protein t o a considerable extent. As indicated by Pitt-Rivers and Impiombato [7],the amount of SDS bound to a protein is almost constant, ranging between 90 and 100%(w/w) in proteins containing S-S bonds. In those proteins which do not possess these bonds, the amount of bound

ELIMINATION OF CHARGE DIFFERENCES

51

SDS increases up to 140%.There are some exceptions (cf., glycosylated proteins), which will be discussed later. If an electrophoretic separation is carried out in a gel that has a pore size that is small enough to restrict mobility, the relationship between the distance travelled by a particular zone and its molecular weight is nearly linear. Proteins with higher molecular weights are more retarded on the gel matrix compared with those with lower molecular weights. Good results can be obtained within average values of molecular weight. Compounds with very high molecular weights hardly penetrate the gel and the method cannot be applied to them. Compounds with very low molecular weights, on the other hand, travel with identical speeds, which also prevents the precise determination of their molecular weights (see below and pp. 53 and 61). This method was first applied to the determination of the molecular weights of proteins by Shapiro [8].Obviously, the precision of the method depends on the accuracy of the determination of the molecular weights of the proteins used to construct the calibration graph. One can use some of the recommended series or a standard calibration kit as supplied by various manufacturers. For special purposes there are n o objections to using special series when needed. This is particularly true for rod-like molecules that do not result in a random-coil conformation in the presence of SDS or with proteins that bind different amounts of SDS [9]. The conditions for molecular weight determinations were discussed in detail by Weber and Osborn [lo] ,Neville [ 1 1, 121 ,Banker and Cotman [13] and Williams and Gratzer [14]. An excellent review was published by Gordon [ I S ] . As mentioned, the linearity between the relative mobility of the zone and the logarithm of the molecular weight generally extends between molecular weights of 15,000 and 70,000 and depends substantially on the gel concentration. Outside this range, deviations occur that make the relationship a sigmoid curve (Fig. 4.7). It has been shown by Williams and Gratzer [14] that proteins with molecular weights less than 6000 form complexes that migrate with identical mobilities. This is explained by the fact that at these low molecular weights there is no longer any dependence of the friction coefficients of SDS complexes on molecular weight. The presence of 6 M urea or another hydrophobic bond-breaking agent re-establishes this relationship and gel electrophoresis in SDS-urea permits the determination of molecular weights below the limit of 6000. With an unknown sample it is required that the molecular weight be determined by interpolation only, and it is advisable to avoid extrapolations. A check on the mobility at zero gel concentration is recommended in order to see whether the value estimated for the sample coincides reasonably with those of the calibration series. When there is insufficient material for this purpose, at least runs at two different gel concentrations should be performed (cf., Zwaan’s method). It has been mentioned already that the presence of S-S bonds influences substantially the amount of SDS bound. Conversion into -SH groups changes the amount of detergent bound but the effect on mobility can be neglected. It has been shown by Pitt-Rivers and Impiombato [7] that with bovine serum albumin (a frequently used standard for molecular weight determinations) the difference between the reduced and non-reduced form is only 10%. It is believed that the changes in the friction coefficient due to the change in the molecular shape compensate for most of the charge difference.

52

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

Conditionsfor binding SDS to proteins It was noticed at the beginning of the development of the SDS procedure that on the addition of SDS to proteins under conditions under which denaturation does not occur, the native conformation of the protein remains preserved even after the detergent has been added. The optical rotatory dispersion (ORD) spectra revealed the preservation of a considerable amount of structures such as the a-helix on the addition of SDS. For molecular weight determinations, however, conditions are frequently preferred in which the sample consists of wholly or mainly denatured proteins binding the same amount of SDS. It was noticed, however, that certain protein species, e.g., viral proteins, require pre-treatment at 100°C in order to bind the expected proportion of SDS.Without this treatment protein aggregates are formed that are responsible for the poorly reproducible results in molecular weight determinations. It has been also stressed by Gordon [ 151 that it is advisable to use high temperatures for the preparation of SDS complexes in order to destroy all enzymic (proteolytic) activity that might be present and that would consequently alter the results. In general, it is recommended to allow the proteins to react at 100°C for about 2 min in addition to reduction of S-S bonds. Some proteins withstand this treatment and do not separate into subunits. On most occasions, however, less drastic treatments are used, as exemplified in Table 4.1. TABLE 4.1 CONDITIONS FOR BINDING SDS TO PROTEINS Protein

Treatment

Reference

Liver membrane proteins, erythrocyte ghost membrane proteins

Heating with 2 parts of Na,CO, and 8 parts (w/w)of SDS in 10%mercaptoethanol, followed by dialysis against 0.05% mercaptoethanol in 2 M urea with 0.01%SDS

Glossman and Neville [ 161

Glomerular basement membrane proteins

37"C, 24 h, 1% mercaptoethanol and 1% SDS in sodium phosphate buffer at pH 7.0

Hudson and Spiro [ 171

On the other hand, most and perhaps all SDS can be removed from proteins, e.g., by dialysis, but only over a very long period of time. It is worth mentioning that a complete renaturation does not occur and in most instances parts of the protein molecules remain folded and do not attain the original structure. Keeping the random coil configuration by preliminary addition of 6M urea helps to remove SDS more rapidly. Active enzymes can be recovered after SDS-gel electrophoresis by soaking the gel segments in 1.O% SDS followed by dialysis against 6M urea [ 181 . SDS is then removed by chromatography on Dowex 1-X2and urea is removed by dialysis. Activity recoveries up to 60%have been reported,

ELIMINATION OF CHARGE DIFFERENCES

53

Limitations of the SDS procedure Regardless of the polyacrylamide gel concentration used, the plot of the logarithm of molecular weight versus relative mobility has two limbs, one at high and the other at low molecular weights. In the latter range, the mobility is constant and independent of molecular weight (Fig. 4.7). The two lines intersect at a molecular weight of lo4,and this value is not grossly influenced by the gel concentration used. The theoretical background of this observation can be appreciated from the following considerations. According to Reynolds and Tanford [ 191, the exponent in the MarkHouwink equation for protein-SDS complexes is 1.3, indicating the high asymmetry of these complexes. Assuming a prolate ellipsoid, they derived the major and minor semiaxial lengths for a series of protein-SDS complexes. It was shown that the minor semiaxis for all complexes is about linearly proportional to molecular weight. Reynolds and Tanford [19,20] indicated that two conditions have to be fulfilled in order to obtain a simple relationship between mobility and molecular weight: (1) the charge to mass ratio must be constant, and (2) the particles must be hydrodynamically homologous. The first requirement is embodied in the fact that the electrophoretic charge of separated particles is determined by the coating of SDS that forms the double layer on the surface of every particle. The other condition is not fulfilled if instead of an ellipsoid the shape of the particle approaches a globule with identical axes. As shown by Reynolds and Tanford [ 1 9 , 2 0 ] , if the axes are identical, the protein should have a molecular weight of 5000. At this point the frictional coefficient will become essentially independent of molecular weight. Beyond this point for some range of molecular weight (up to about 15,000) the relationship between the logarithm of molecular weight and mobility is expected to deviate from linearity, The behaviour of proteins was studied by Dunker and Rueckert [21] on gels of different concentration (5, 10 and 15% gels). The plots of the logarithm of molecular weight versus mobility were linear within a molecular weight range that was characteristic of the gel composition. As indicated in Fig. 4.8, the plots of molecular weight versus mobility show an inflection when the molecular weight falls below a critical size, viz., about 20,000 for 5% gels and 10,000 for 15 and 10%gels. From the data in Fig. 4.8, it is possible to conclude that if an appropriate calibration is carried out it is possible to extend molecular weight determinations on 5% gels down to the 20,000-5000 range. However, the steepness of the slope requires a rather precise determination of mobility. A survey of results obtained from gels of different concentration is presented in Table 4.2. This table can also serve as a guide for choosing appropriate standards. In general, the precision of this procedure is about 2% of the molecular weight determined, but clearly with several proteins larger discrepancies occur, which can be attributed to small but significant variations in mobility versus molecular weight in the different proteins. It is likely that such variations result from differences in conformation or intrinsic charge and from differences in the proportion of detergent anions bound to the structure. With lysozyme and pepsin the anomalies are attributed to differences in the detergent bound, but no such explanation is feasible for ribonuclease A.

54

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

1.0

400 X

X

200

. x X

100 )

X

80

X X

"60

's

X

I#

40

x

x X

XX

20 S

z

10

4

I

I

I

I

a2

OA

06

0.8

RF

X

ELIMINATION OF CHARGE DIFFERENCES

55

It has been also shown that the state of folding can also have a considerable effect on mobility in non-SDS gels; in the presence of the detergent such effects are small, As noted by Williams and Gratzer [14] , the SDS method is likely to fail for strongly acidic proteins. Here cationic rather than anionic detergents should be used, e.g., trimethylammonium bromide. In these detergents the plot of the logarithm of molecular weight versus relative mobility is similar to that obtained with anionic detergents, except for some curvature at higher molecular weights (Fig. 4.9). Equally, at a molecular weight of lo4 the relationship becomes parallel to the molecular weight axis, i.e., all particles move with the same mobility regardless of their molecular weight and no estimates of the latter can be made. Highly charged proteins (e.g., ferredoxins) with too many negative side-chain groups have a lower mobility than those of other proteins of similar size, indicating a low binding of the detergent, while those with too many positive charges bind too much SDS, resulting in a higher relative mobility. Reducing the number of charges per molecule, e.g., by esterification, makes the mobilities of the proteins in question fit the linear plots of logarithm of molecular weight versus mobility. Papain is a good example of proteins that bind an unusual amount of SDS: for no obvious reasons it binds only about 20% (w/w) of the expected amount of SDS.

Anomalous behaviour of SDS complexes Both unusual conformations and the binding of unusual amounts of SDS can be revealed by using Ferguson plots (p. 47). Determining the relationship between (Uj),.,,-(electrophoretic mobility at zero gel concentration) and K R can establish whether or not a particular determination of molecular weight is correct in the following way. For a given standard series of proteins, this relationship is normally linear and its determination requires the construction of Ferguson plots for each member of the series tested. As the effective size of protein-SDS complexes increases linearly with increasing molecular weight over a wide range, it is reasonable to expect that proteins migrating anomalously in SDS gels must differ from the normal counterparts in their (Ui)ma- versus K R relationship. This has been demonstrated on maleylated proteins, which are known to migrate more slowly than the corresponding parent proteins, thus leading to overestimates in molecular weight determinations (Table 4.3). The discrepancies with proteins of low binding capacity for SDS can be lowered in two ways: either by using small pore gels as described by Segrest and Jackson [22] or by using an appropriate standard series, composed of proteins (glycoproteins) of equal binding capacity with regard to SDS. In the Segrest and Jackson [22] procedure gels of

Fig. 4.7. (a) Relationship between molecular weight and electrophoretic mobility of proteins and peptides in SDS-polyacrylamidegels. Mobilities are expressed relative to the Bromphenol Blue markers. 0,5% Acrylamide; =, 10%acrylamide; 0,15% acrylamide. (b) Molecular weight @) on a semilogarithmic scale versus relative mobility ( R J for ~ a variety of protqin-SDS complexes subjected to electrophoresis at pH 9.5. In the upper molecular weight region the curve is hyperbolic while in the lower regions the scatter obscures the nature of the relationship.

305,000 260,000 210,000 164,000 134,000 6 2,000 46,500 37,500 3 1,000 26,300 24,500 N.D. 17,000 17,500 N.D. 13,800 N.D. 18,500 N.D. 14,000 N.D. 12,000 9,500 6,000

Apparent molecular weight

3.6 6.0 5.3

1.4

36.0

4.2

5.5 0.5

7.6 1.5 6.0 2.5 1.5 6.5 1.1 5.6 9.9 2.0 2.9

(%I

Deviation

Data from 5% gels

14,000 13,500 13,600 10,500

15,500

62,000 46,400 37,000 31,300 26,300 24,100 20,800 16,000 17,500 18,600 13,900 13,700 16,500

Apparent molecular weight

6.0 3.5 8.7 21.0 6.0 0.7 1.6 9.1 3.9

0.5

6.5 1.0 4.2 9.0 2.1 1.3 2.4 11.0

(%I

Deviation

Data from 10%gels

STANDARD SERIES OF PROTEINS ON 5,lO AND 15%GEL Molecular weight

330,000 264,000 198,000 160,000 132,000 66,000 46,000 35,500 34,400 25,741 23,800 20,300 18,000 17,600 17,400 14,400 15,000 13,680 14,620 13,927 13,729 12,400 10,157 5,700

7.3 4.8

15.0 2.6 1.44

5.5

10.0 0.5

8.2 1.7 7.0 7.4 1.9 1.7

Deviation (%)

Data from 15%gels Apparent molecular weight

61,000 46,800 38,000 32,000 26,200 24,200

N.D. 16,200 17,500 N.D. 13,600 N.D. 15,800 15,000 14,100 N.D. 13,300 10,600

wl

m

5

r m

5 m

k >

3 = v1

5

2

m p

3 0

ga

z

m

P

57

ELIMINATION O F CHARGE DIFFERENCES

L 4

relative mobility

8

12

16

20

Fig. 4.8. Plot of the logarithm of the molecular weights of a series of proteins versus their relative mobilities on (A) 5%, (B) 10% and (C) 15% polyacrylamide gels containing 0.1% SDS. Each protein is identified by its number in Table 4.2. The ordinate intercepts obtained by extrapolating to zero mobility u e characteristic of gel composition and therefore may be useful for comparing results from different laboratories. Such extrapolations, however, probably do not accurately describe the behaviour of polypeptides larger than the ones presented, especially since the slow-moving large polypeptides spend an appreciable portion of their time passing through an atypically porous zone, which is formed at the water-polymer interface during polymerization. Mobilities are normalized with respect to chymotrypsinogen A (RF = 1).

Fig. 4.9. Relationship between molecular weight and electrophoretic mobility of proteins and peptides in polyacrylamide gel containing the cationic detergent cetyltrimethylammonium bromide. Acrylamide concentration, 10%.

58

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

ACRYLAYIDE CONCENTRATION (percent 1

Fig. 4.10. Observed molecular weights of four glycoproteins. 8-0, Human erythrocyte membrane glycoprotein;s o ,human erythrocyte membrane tryptic g1ycopeptide;-, fragmented human erythrocyte membrane glycopeptide;w, porcine ribonuclease,higher molecular weight species;-, porcine ribonclease,lower molecular weight species.

of different concentration can be used with the result of obtaining curves approaching asymptotically the real value of molecular weight (Fig. 4.10). The N-laurylsarcosineprocedure When the problem of retaining enzymic activity is foreseen it can be recommended to use sarkosyl (N-laurylsarcosine) instead of SDS. A linear relationship between the logarithm of the molecular weight and the mobility of these complexes has been demonstrated by Morgan [23] (Fig. 4.1 1). This complexing agent would be removed at a higher pH than with SDS because of the higher pH of the carboxyl group of N-laurylsarcosine. The lower inhibitory effect of this compound on a number of enzymes is also worth mentioning. A survey of the applications of this compound is given in Table 4.3. The application of sarkosyl is not without problems, however. Thus, ribosomal proteins or membrane proteins do not form sharp bands in this type of separation. In this case, however, the results with SDS are very good. If the membrane proteins are first purified and then applied in electrophoresis as fairly pure samples, they do form sharp

59

ELIMINATION OF CHARGE DIFFERENCES TABLE 4.3 APPLICATIONS OF SARKOSYL IN MOLECULAR WEIGHT DETERMINATIONS Protein

Reference

Mouse liver plasma membranes 5'-Nucleotidase Nucleotide pyrophosphatase Ribosomal proteins

Evans and Gurd [ 24 1 Evans and Gurd [ 241 Evans and Gurd [ 251 Brimacombe et al. [ 261

1.0

-

0.6

-

0.2

-

r n

t

al

._

-?

I

I

I

I

4.0 log

I

I

4.4 molecular weight

I

I

I

I 4.8

Fig. 4.1 1. Plot of relative mobilities of proteins against the logarithm of molecular weight using N-laurylsarcosine. Cyt c, cytochrome c from horse heart; RNAase, ribonuclease from ox pancreas; Lys, lysozyme; Myg, myogIobin from sperm whale; Hb, haernoglobin monomer (bovine); SBT, soya bean trypsin inhibitor; Chy, chymotrypsinogen; LDH, lactic dehydrogenase, rabbit muscle; Perox, peroxidase; BSA, bovine serum albumin; Cat, catalase.

bands, This fact was used by Evans et al. [27] to demonstrate phosphodiesterase activity, the overall protein pattern and the presence of carbohydrates in the same gel layer.

Other masking procedures and urea In addition to sarkosyl, Triton X-100 has also been used [28] (0.25% Triton X-100 and 5M urea), but this procedure has not been widely adopted. Ballou et al. [29] introduced chloral hydrate; in contrast to the masking effect of SDS, protein charges in chloral hydrate treatment remain preserved. A useful property of this reagent is that it can easily be removed and the activity of the original protein can be reconstituted. The use of urea in molecular weight determinations has been superseded by SDS electrophoresis. However, in many instances SDS-urea gives very sharp bands and, in particular, molecular weights of low-molecular-weight peptides can be determined in this

60

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

60

1

0

20

40

60

Distance

80 100 migrmted tmm)

120

Fig. 4.12. Effect of 8Murea on relative electrophoretic mobility of oligopeptides in SDSpolyacrylamide gel: 0,No urea; 0 , 8Murea, 12.5% acrylamide, 1 : 15 cross-linkage. Numbered peptides are encoded as follows:

No. 1 2 3 4 5 6 7 8

15 16 17

Protein

Molecular weight

Ovalbumin Carboxypeptidase A Myoglobin Myoglobin I+II Cytochrome c Myoglobin I Cytochrome c I Myoglobin I1 Bovine trypsin inhibitor Adrenocorticotropic hormone Insulin Insulin B Insulin A Glucagon Cytochrome c I1 Myoglobin I11 Cytochrome c 111 Bacitracin Polymyxin B

46,000 34,500 17,200 14,900 12,300 8,270 7,760 6,420 6,160 4,550 5,700 3,400 2,300 3,460 2,780 2,550 1,810 1,400 1,225

Roman numbers designate cyanogen bromide cleavage products arising from a particular protein. Myoglobin I+II refers to an incompletely cleaved peptide.

CHOICE OF STANDARD SERIES

61

way. As mentioned above, when the molecular weight is less than about 15,000, the linear relationship between the logarithm of molecular weight and mobility in SDS gels alone n o longer exists. However, in 8M urea, as shown by Swank and Munkres [30] ,the relationship remains reasonably linear and molecular weights of peptides down to 2000 can be estimated (Fig. 4.1 2). It is also worth mentioning that in many separations the presence of urea improves the sharpness of the bands, although in other separations this is not the case; a detailed discussion supported by evidence obtained from isoelectric focusing was published by Salaman and Williamson [31]. For molecules too large to be separated on polyacrylamide gels, combined acrylamideagarose gels can be used; 0.5% of agarose makes dilute gels much stronger [32] and very large molecules of nucleic acids or aggregates such as ribosomes and ribosome clusters can be separated in these media. Mobilities in large-pore acrylamide-agarose gels are related to the logarithm of the particle weight in an analogous manner to SDS-protein complexes.

CHOICE OF STANDARD SERIES In addition to standard protein mixtures (cf., Table 4.3), polymers of a single well defined protein can also be used. Bovine serum albumin (monomer up to pentamer) and lysozyme (monomer up to pentamer) can be recommended, the former covering the molecular weight range of ca. 60,000-300,000 and the latter 20,000-80,000. These can, of course, be combined either between or with other standards; insulin is frequently used in order to extend the molecular weight range towards lower values. The procedure for the preparation of lysozyme polymers is as follows. A 200-mg amount of egg-white lysozyme is reduced by incubation with mercaptoethanol in 10 ml of 0.2MTris-HC1 buffer, pH 8.1 (containing 10%of the reducing agent). The reducing solution is 8M with respect to urea and contains 0.01% of ethylenediamine tetraacetate. The product is dialysed against water and lyophilized. Material insoluble in 0.005M acetic acid is removed by centrifugation prior to use [33]. Bovine serum albumin polymers are even simpler to obtain as most commercial preparations already contain the polymers. The use of polymerized urease was suggested by Blattler and Reithel [34] : if the method of Gorin et al. [35] is used for the extraction of urease and the extraction procedure itself is carried out with 160 ml of acetone, 0.01 Mp-mercaptoethanol, 0.001 M EDTA and deionized water in a volume of 500 ml, and if the product is stored as crystals in acetone-citrate solution, then the protein slowly polymerizes. Solutions of urease contain the monomer up to the pentamer; higher polymers are believed to be insoluble. Direct detection of the calibration series without additional staining has been proposed by Inoue [33] ,by labelling the calibration series with 1-dimethylaminonaphthalene5-sulphonyl chloride (Dns-Cl). In the recommended procedure, 5 ml of the protein sample (4 mg/ml) in 2%NaHC03 are mixed with 5 ml of 2 rng/ml Dns-C1; the reaction is allowed to proceed for 2 h at 37°C with shaking. Then 20 ml of acetone are added, the precipitate is centrifuged off, washed with 10 ml of acetone and dissolved to a final concentration of

62

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

2 mg/ml in 0.01M sodium phosphate buffer (pH 7.1), containing 1% of SDS and 10%of glycerol. Occasionally heating to 70°C helps to dissolve the precipitate. In order to avoid problems arising from different mobilities caused by a number of external factors (gel reproducibility), a simple technique of internal calibration was introduced by Dunker and Rueckert [21]. The procedure is referred to as the split gel technique; polyacrylamide gels that have been prepared in the usual manner were divided in their sample compartment by inserting a Styrofoam partition (see also Traut et al. [36]), and mixtures of proteins to be compared were applied on opposite sides of the sample compartment.

MOLECULAR SIZE DETERMINATION BY PORE LIMIT ELECTROPHORESIS In gradient polyacrylamide gels, narrowing of protein bands occurs concomitantly with a decreasing migration rate [37]. The proteins move more and more slowly until they reach a certain relationship to the polymerized chains of the gel such that their further movement is linear with time at constant voltage. In other words, the principle of the separation can be such that proteins migrate through progressively smaller pores, the sizes of which are regulated by the gel concentration and finally tend to stop and concentrate where the pore size is too small to allow further migration. Although in the early stage the separation occurs both through charge and size differences, towards the end of the separation the influence of charge is lower and the resulting separation is based on size differences only. Polyacrylamide gradient slab gels (4-26% of polyacrylamide) for this purpose are commercially available (Universal Scientific Ltd., London, Great Britain). Based on this observation, ‘pore limit’ can be defined as the distance migrated from the start in a specific gradient after which further migration is directly proportional to time at any given voltage. After all proteins in the mixture have reached their pore limits, there is no reason for further electrophoresis as all the components migrate at the same rate. Further, if at least two proteins were included in the mixture, it is possible to determine the molecular weights of unknown proteins provided that a simple log-log relationship between migration distance and molecular weight exists. This is demonstrated in Fig. 4.13 (Slater [38]). The limitations of this procedure are related to the charge of the protein molecule and its shape. Obviously, the protein must carry enough net charge to be able to move at an initial speed higher than that which it would have after reaching the pore limit. It is also to be expected that rod-like molecules will have an anomalous behaviour here because of their molecular shape. Association-dissociation equilibria must also be considered as the procedure requires sufficient time to allow these phenomena to occur. Most of the experiments were carried out in alkaline buffers but acidic buffers will also suffice. Gradients between 5% and 30%polyacrylamide will cover the molecular weight range between 32,000 and 48,000. The standards that can be used are summarized in Fig. 4.13. If 8M urea is added to the buffer used, then studies can be carried out with molecules of molecular weight down to 20,000.

63

DETERMINATION OF STOKES RADII BY RHEOPHORESIS 20

---

-

-

2

I

I

I

I

20 000 30 000

I

I

I

l

l

100000

I

I

200000

1

I

I

l

l

1

500 000

Fig. 4.13. Migration distance versus protein molecular weight in pore limit electrophoresis.

No.

Protein

1 2 3 4 4 4 4 4 5

Prolactin Pepsin Ovalbumin Human albumin Dog albumin Rat albumin Rabbit albumin Bovine albumin Human transferrin Bovine dialbumin Human y-globulins Catalase Horse apoferritin Ferritin, CdSO,

6

I 8 9 9

Molecular weirrht 32,000 35,000 44,500

69,000 65,000 63,000-69,000 68,000 67,000 88,000 134,000

160,000 24 8,000

480,000 480,000

DETERMINATION OF STOKES RADII BY RHEOPHORESIS As demonstrated by Waldman-Meyer [39], rheophoresis appears to be a unique method with respect to offering simultaneous information about electrophoretic, chargebased properties of a substance and about retention properties based on the hydrodynamic radii of the migrants. The principle and the theoretical background of this technique are described on p. 33. The migration distance of any unretained solute in a two-phase gel (Sephadex thin-layer electrophoresis) can be expressed as

so+= s, + s, + s,

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

64

where S,, S, and S, are the rheophoretic, endoosmotic and electrophoretic contributions, respectively. It is assumed that the substance applied has a positive charge and has been applied on the anodic side, endoosmotic flow is directed towards the cathode, and the direction of flow towards the cathode is designated as positive. If the same substance is applied at the same distance from the centre but on the cathodic side, then the distance migrated will be

s,-

=

-s, +s, +s,

Hence the rheophoretic migration is

s,

=

s; -s,

2

For a retained molecule, the corresponding distances will be

s: = P ( s , +s, +S,) s;

= p(-

s, + s, + S,)

where p is the retention coefficient, defined as p = - Vret Vunret

In eqn. 4.9, v,,, and vUmetrepresent the velocity of a retained and an unretained molecule, respectively. Obviously, there is an analogy with the retention index defined in , wo is the retention volume of the totally excluded chromatography as p = w O / w e where molecule and w e is that of the test substance. Combining eqns. 4.7 and 4.8, we obtain p = -

s: - s;

(4.10)

2sx

and substitution of eqn. 4.6 for S, gives p=-

s,' - s; s,'- s,-

(4.1 1)

According to eqn. 4.1 1, it is possible to obtain distribution coefficients by measuring the migration distances of retained substances and an unretained migrant. In practice, several samples of the same substance can be applied at different distances from the anode and an electric current induced in the system. Then the migration distances are plotted against the points of application. In the absence of temperature gradients, straight lines are obtained, as indicated in Fig. 4.14. The slope of these straight lines can be expressed as Pret

=

-Asr - - A s , - p --

4

for a retained molecule and

(4.12)

65

DETERMINATION OF STOKES RADII BY RHEOPHORESIS

Fig. 4.14. Migration distances as a function of application position L . Pure rheophoretic migration of any unretained molecule in the given experiment (only directly measurable in thermal gradient rheophoresis). (So), Measured migration of unretained, negatively charged molecule (e.g. Blue Dextran); - - -, measured migration of unretained, positively charged molecule; -- (S), measured migration of a molecule of same charge but a retention coefficient of p = 0.50. The first three slopes are equal to pa = Punret = AS,/AI,,, while the last has a value of p = p, p = Pret = (AS,/AI,,) p . In the lower part, the same distances So and S are depicted as they appear in the experiment. Migration towards the anode is denoted as negative. Electroosmosis produces a displacement in the opposite direction. The minor axis of symmetry is indicated by M. - e m ,

-- -.-

(4.13) for an unretained species. Combining eqns. 4.12 and 4.13, we have P=-

Pret

(4.14)

Punret

Therefore, for a given series of macromolecules (cf., proteins) if the Pret values are estimated and flu, is determined with Blue Dextran on a Sephadex thin layer, then the resulting p values can be plotted against the effective hydrodynamic radii determined by other methods and a calibration line similar to that depicted in Fig. 4.1 5 is obtained. Then, by measurement of the migration distances and determination of the Pret and p values of an unknown substance one can read off its effective radius from this calibration graph. The practical procedure as proposed by Waldman-Meyer [39] is as follows. A 7-g amount of Sephadex (2-100Superfine is dissolved in 133 g of 0.12M sodium barbital buffer (pH 8.6) in a flask. The stoppered flask is shaken at 37°C for 3 days, the supernatant is sucked off and 20 x 20 em (0.75 mm thick) layers are prepared using a thin-layer chromatographic applicator. The plate for analysis is then placed on a thermostated

66

MOLECULAR SIZE AND SHAPE IN ELECTROPHORESIS

Fig. 4.1 5. Correlation between retention coefficient and Stokes radii; r; denotes the exclusion radius. For all molecules of re > rk, P = 1.00. The intercept on the abscissa, ro (< 0) shown in the inset together with a p versus re plot chosen at random. Evidently, Ap/Ar, = 0 for re 3 r,. Proteins: a, cytochrome c , heart, equine; b, ribonuclease, pancreas, bovine; c, lysozyme; d, myoglobin, sperm whale; e, carbonic anhydrase, bovine; f, plactoglobulin; g, ovalbumin; h, serum albumin, human; i, transferrin, human; j, alcohol dehydrogenase, yeast; k, Blue Dextran.

horizontal support provided with levelling screws, corrections are made with the electrode buffer and, during the equilibration period, excess of buffer is allowed to flow into the electrode compartments. The plate is covered with a frame and a plate with holes for sample application and a cover. The time required for equilibration is followed with an ohmmeter connected to measuring electrodes. When there is no further increase in resistance, samples are applied. The distances between samples are arranged so as not to allow overlapping during the subsequent overnight run. The voltage applied is 1.52.0V/cm. The voltage over the measuring electrodes is recorded and a decrease begins 30-60 min after the run has been started. Finally the migration distances are measured by covering the gel with a sheet of Whatman paper and illuminating the plate from beneath in order to detect the positions of sample application. The paper is then removed and stained in the usual manner.

REFERENCES 1 C. C. Ackers, Biochemistry, 3 (1964) 723. 2 K. Felgenhauer, in H. Peeters (Editor), Prorides o f t h e Biological Fluids, Pergamon Press, Oxford, 1970, p. 505. 3 K. A. Ferguson and A. L. C. Wallace, Nature (London), 190 (1961) 629. 4 S. Hjerten, S. Jerstedt and A. Tiselius,AnaZ. Biochem., 11 (1965) 211 and 219. 5 D. Rodbard and A. Chrambach, Proc. Nut. Acad. Sci. U.S.,65 (1970) 970. 6 J. Zwaan, Anal. Biochern., 21 (1967) 155.

REFERENCES 7 8 9 10 11 12 13 14

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

67

R. Pitt-Rivers and F. S . Impiombato, Biochem. J . , 109 (1968) 825. A. L. Shapiro, Biochem. Biophys. Res. Commun., 28 (1967) 815. R. Helm, 2. Deyl and 0. Vancikova, Exp. Cerontol., 12 (1977) 245. K. Weber and M. Osborn, J. Biol. Chem., 244 (1969) 4406. D. M. Neville,J. Biol. Chem., 246 (1971) 6328. D. M. Neville, Biochim. Biophys. Acru, 154 (1968) 540. G. A. Banker and C. W. Cotman, J. Biol. Chem., 247 (1972) 5856. J. G. Williams and W. B. Gratzer,J. Chromatogr., 57 (1971) 121. A. H. Gordon, Electrophoresis of Proteins in Polyuc~ylamide and Sturch Gels, North-Holland/ American Elsevier, Amsterdam, New York, 1975. H. Glossman and D. M. Neville, J. Biol. Chem., 246 (1971) 6339. B. G. Hudson and R. G. Spiro, J. Biol. Chem., 247 (1972) 4229. K. Weber and D. J. Kuter, J. Eiol. Chem., 246 (1971) 4504. J. A. Reynolds and C. Tanford, Proc. Nut. Acad. Sci. US.,66 (1970) 1002. J. A. Reynolds and C. Tanford,J. Eiol. Chem., 245 (1970) 5161. A. K. Dunker and R. R. Rueckert,J. Eiol. Chem., 244 (1969) 5074. J. P. Segrest and R. L. Jackson,Methods Enzymol., 28B (1972) 54. J. M. Morgan, Exp. CellRes., 65 (1961) 7. W. H. Evans and J. W. Gurd, Biochem. J . , 128 (1972) 691. W. H. Evansand J. W. Gurd, Biochem. J., 133 (1973) 189. R. Brimacombe, J. M. Morgan and R. A. Cox, Eur. J. Biochem., 23 (1971) 52. W. H. Evans, D. 0. Hood and J. W. Gurd, Biochem. J., 135 (1973) 819. R. Lim and E. Taddayyon, Anal. Biochem., 34 (1970) 9. B. Ballou, G. Sundharadas and M. L. Bach, Science, 185 (1974) 531. R. T. Swank and K. D. Munkres, Anal. Biochern., 39 (1971) 462. M. R. Salaman and A. R. Williamson, Biochem. J . , 122 (1971) 93. U. E. Loening, Biochem. J., 113 (1969) 131. M. Inoue,J. Eiol. Chem., 246 (1971) 4834. D. P. Blattler and F. J. Reithel, J. Chromatogr.,46 (1970) 286. G. Gorin, M. F. Butler, 1. M. Katyal and J. E. Buckley, Proc. Okla. Acad. Sci., 40 (1960) 62. R. R. Traut, P. B. Moore, H. Delius and A. Tissieves, Proc. Nut. Acad. Sci. U.S., 57 (1967) 1294. J. Margolis and K. G . Kenrick, Anal. Eiochem., 25 (1968) 347. G. G. Slater,Anul. Chem., 41 (1969) 1039. H. Waldman-Meyer, Biochim. Biophys. Acta, 261 (1972) 148.

This Page Intentionally Left Blank

Chapter 5

Zone electrophoresis (except gel-type techniquesand immunoelectrophoresis) W.OSTROWSKI

CONTENTS Paper techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General considerations .................................................. in filter-paper . . . . . . . . . . .... .............. Electroosmosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hydrodynamicliquid flow ................................................. Adsorption and diffusion ........... ......................... Other factors . . . .................................................... Equipment . . . . . . . . ....................................................

......................

.....

.............

70 70 70 71 72

73 73 73 74 75 75 76 76 79 79 80 81 81 82 a4 84 85

Power supply ............................. ...................... Densitometers and scanning devices .......................................... Low-voltage technique. . . . . . . . . . . . . ...... Proteins and enzymes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nucleic acids and polynucleotides . ............ Compounds of low molecular weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inorganicions ........................................................... High-voltage technique ...................................................... Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... Twodimensional separations . . . . . . . . ....................................... Twodimensional electrophoresis ............................................ Electrophoresis combined with chromatography ................................ driving forces acting simultane Twodimensional separation with t Cross paper electrophoresis . . . . . . . . ....................... Separations on cellulose acetate and n cellulose membranes . . . . . . . Separations on ionexchange papers. ........................... 92 Ultramicroelectrophoresis on single fibres or on thin threads. ......................... Separation in non-aqueous solvents ....................................... 93 Thin-layer electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 .................................... 96 General considerations ... ..................... Equipment,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 One-rJimemional separation ................................................... 97 ................................. 98 Twodimensional separation. ............. Electrophoretic separation of inorganic ions in d salts. . . . . . . . . . . . 100 References . . . . . . . . . . . . . . . . . ............................................. 103

70

ZONE ELECTROPHORESIS

PAPER TECHNIQUES General considerations At the same time as A. Tiselius [ l ] published his classical work on free boundary electrophoresis in 1937, Konig [2] indicated the possibility of using filter-paper for the electrophoretic separation of protein mixtures. In the 1950s several workers described various electrophoretic techniques with paper as the anticonvection agent for the separation mixtures of amino acids, peptides, proteins and other substances [3-61. Since then, filter-paper has been widely used as the most practicable material as supporting medium for electrophoresis. Undoubtedly, the development of paper chromatography in the 1940s and later led to the production of special types of filter-paper that were used simultaneously for paper electrophoresis. The use of filter-paper and other solid supports also introduces several factors involving electrophoretic mobility of charged molecules, viz., electroosmosis, hydrodynamic liquid flow, adsorption, chromatographic effects and molecular sieving effects. From this point of view, zone electrophoresis cannot be used for precise determinations of the electrophoretic mobility, isoelectric point or other physical properties of migrating substances without a suitable standardization procedure for the particular technique. The ideal support would be one that is free of impurities, chemically inactive with respect t o the separated substances, free of adsorptive activity and with the lowest electroosmotic effect possible. No type of filter-paper fulfils all of these conditions. However, new support media have been introduced that have much more favourable properties, e.g., cellulose acetate, polyacrylamide gel and agarose gel. Electrophoretic mobility in filter-paper

It was quickly found that the mobility of charged particles on filter-paper is lower than that measured in free boundary electrophoresis. Two theories have been developed to explain this difference: tortuous path theory [4] and the barrier theory [7]. The tortuous path theory describes the lowered mobility in filter-paper as being due to an increased path length around the matrix particles (tortuosity factor) and to a decreased electric field strength acting along this longer pathway. The observed mobility, (uncorrected for sorption and related interactions), is related to the free solution mobility, 4,by the relationship [8]

where 1 is the effective length of the tortuous path and L is the length of a direct, straight path (Fig. 5.la). In the barrier theory, the support is treated as being composed of randomly distributed barriers (Fig. 5.1 b) which slow the migrating molecules by collisions or by mechanical restriction (Fig. 5 . 1 ~ )of the free pathways [9]. Both theories have been extensively studied [lo, 111 and a comprehensive discussion was given by Edward [12]. In addition to the tortuosity factor, 7 , two other factors were defined: a constrictive factor, C, and retardation factor, KR, which define the obstructive factor, to:

71

PAPER TECHNIQUES

b

0

C

Fig.S.l. Schematic representation of (a) tortuosity, (b) retardation and (c) constrictive effects during migration of an ion in the stabilized medium. to =

ccKRY-2

The extent of the decrease in mobility can be measured by an experimental determi. nation of the obstructive factor, to :

s' = K U. & = A (&)macro

where K , and K , are the specific conductivity in the presence and absence of solid support medium, respectively, and E is the void fraction or porosity of the matrix. Hence, in the absence of an electroosmotic flow, can be obtained from electrophoretic experiments provided that K,, K , and Care known. T h s shows clearly that the rate of migration of a charged molecule on filter-paper in an electric field will be the vector of the sum of a driving force (electric potential) and a number of additional factors that affect the separation.

Elecrroosmosis The filter-paper used for paper electrophoresis and paper chromatography is produced from a mass of pressed cellulose fibres containing a large amount of hydroxyl groups. During the processing of cellulose, some carbon atoms are oxidized to carboxyl groups (one -COOH group per ca. 500 glucosyl residues), which are negatively charged at neutral and alkaline pH, i.e., at the pH at which electrophoresis is usually carried out. In addition t o the possible presence of ionizable groups in the support, most stabilizers used for electrophoresis acquire a net negative charge due to the zeta potential at the boundary of the support and the buffer solution. The natural tendency of these negatively charged groups is to move towards the anode, but this would be impossible as the medium, in this instance filter-paper, is a stationary phase. To counter this effect, a movement of positively charged water molecules, H30+, towards the cathode takes place, which gives the effect of an osmosis of solvent towards the cathode and usually results in an apparent migration of neutral molecules towards the cathode. The electroosmotic flow is most pronounced in alkaline solution and varies with the type of support medium used. For instance, glass-fibre paper or thin layers of aluminium oxide or titanium oxide show a positive electrokinetic potential and the movement of water molecules occurs in these media in the opposite direction, i.e., towards the anode [13, 141. The electroosmotic flow of the electrolyte can be calculated:

72

ZONE ELECTROPHORESIS

4

where vo, is the electroosmotic mobility, 3' the zeta potential, E the electric field strength, 77 the viscosity of the medium and E the dielectric constant of the medium. The electroosmotic movement of the buffer can be measured experimentally by using neutral compounds such as dextran or glucose [4], hydrogen peroxide. [ 151 and many other substances [161. In practice, electroosmotic flow is frequently disregarded when considering the electrophoretic migration of a separated mixture. The relative mobility, (Llis)rel, or the apparent separation of substances being analysed is the parameter of practical importance in most instances. (U& is calculated as the ratio of the mobility of the unknown t o that of a standard substance [17, 181 and it includes the actual electrophoretic mobility in addition to electroosmotic and hydrodynamic flow.

Hydrodynamic liquid flow A strip of filter-paper moistened with an electrolyte has a resistance that depends on, among other factors, the chemical composition of the buffer and the amount of it on the paper, i.e., on the "wetness" of the strip [7].If the strip of paper or other support medium is not sufficiently protected from evaporation, the Joule heat evolved by the passage of the current will result in a loss of water from the surface which will be compensated for by a hydrodynamic flow of buffer liquid from the ends of the strip immersed in the electrode vessels. The replacement of evaporation losses results in an increase in the electrolyte concentration and conductivity in the supporting medium. As can be seen in Fig. 5.2, the hydrodynamic liquid flow has the lowest value at the middle point of the paper strip (zone n) and the velocity of flow on each side is directly proportional to the distance from the centre of the strip. This flow will affect the mobility of the molecules being investigated, so the apparent mobility of the substance may depend on its position on the paper. It is possible to reduce the buffer flow in various ways, but it is also possible t o calculate the effects of the evaporation of water, and the theory of electrorheophoresis [19,20] enables the resultant buffer flow to be taken into account in evaluating the mobility [21-231 and other properties of migrating ions [24, 251. Electrophoretic migration

4

J

Fig. 5.2. Schematic representation of the forces involved in migration of an ion in filter-paper or thin layer 1241. The polarity of the electrophoretic and the electroosmotic migration depends on the experimental conditions. The velocities of the hydrodynamic flow from both sides are directly proportional to the distance from the midpoint, n, of the strip.

PAPER TECHNIQUES

73

Adsorption and diffusion Owing to the negative charge on the paper strip, positively charged molecules are adsorbed on paper and their mobility is reduced. The adsorption effect is manifested by “tailing” or the formation of ”comets” of the solute ions observed after completion of electrophoresis. The formation of non-symmetrical zone profiles during electrophoresis on filter-paper can also be explained on the basis of obstructivity [26, 271, chromatographic effects [28] and concentration-dependent equilibria 1291. If the pH of the buffer solution approaches the isoelectric point of the migrant, the effect of adsorption would decrease to zero 1301. The adsorption and tailing can be almost eliminated by addition to the buffer solution of large polyvalent ions (for instance, 2-aminonaphthalene-6,8sulphonic acid) when low-molecular-weight ions are separated on filter-paper [29]. Electrophoresis on ion-exchange papers takes advantage of selective adsorption to give a resolution of different solute mixtures-p1 I. The extent of adsorption can be determined prior to electrophoresis by the method described by Kunkel [32]. During the course of an electrophoretic separation, the zones will broaden as a result of molecular diffusion in the direction of electrophoretic movement [33]. In the separation of macromolecules such as proteins and nucleic acids, the adverse effect of diffusion on resolution is generally small. To prevent diffusion of small ions and to facilitate resolution, high-voltage techniques can be used. The application of the sample in as narrow a band as possible usually gives sharp and better resolved zones.

Other factors Reproducible electrophoretic separations on a paper strip will depend on the homogeneity of the electric field along the strip. Usually after a long period of electrophoresis and continuous evaporation of water, an increase in temperature is observed, which leads to an increase in salt concentration near the zone 1~ of the strip (when a non-volatile buffer is used), to an increase in conductivity and a decrease in potential gradient. An additional complicating factor is the application of the sample, itself an electrolyte, which may result in a significant local increase in conductivity if the concentration is high enough. The local field will fall accordingly, with the effect that the charged molecules will move only very slowly or even be completely immobilized. It is therefore necessary to dilute the sample or to dialyse, so that its conductance is as close as possible to that of the buffer solution. In conclusion, the actual velocity of a molecule being studied is a sum of electrophoretic migration, electroosmotic flow, and hydrodynamic flow (see eqn. 1.17). The electrophoretic mobility will, of course, depend on the pH and the ionic strength of an electrolyte, the size and shape of the molecule, the number of charges, the viscosity of the solution and the solubility of the solute. These factors are discussed in other chapters. Equipment The apparatus for paper electrophoresis consists of two main components, an electrophoretic tank or chamber and a source of electric power. The electrophoretic tanks

74

ZONE ELECTROPHORESIS

are varied in construction, depending on a horizontal or vertical arrangement of the strip and other needs characteristic of particular experiments. Different tanks are described below in the discussion of various electrophoretic techniques. Here the basic design of the tanks for low-voltage electrophoresis will be considered.

Electrophoretic tank The apparatus commonly used for low-voltage paper electrophoresis is of two types, with either a horizontal or a vertical arrangement of the strip. In the horizontal type of tank (Fig. 5.3a), the paper strip is clamped horizontally between two or more shoulder glass rods or nylon threads. The ends of the paper strip are immersed in the buffer vessels to the same depth and the levels of liquid in opposite compartments must be equalized so as to prevent siphoning through the strips during the run. The whole trough is covered with a tightly fitting lid, made of Plexiglas.

I

I

I

I

b

I& ' 1

d

a

Fig. 5.3. Schematic drawings of the tanks used for paper electrophoresis: (a) horizontal type; (b) vertical type; (c) immersed strip type; (d) enclosed strip type.

In the vertical type, a freely hanging strip of filter-paper is folded in two and supported V by a glass rod (Fig. 5.3b). In both types of electrophoretic tank (open strip technique) [3,6,27,34-401, evaporation is not suppressed completely, but is minimized. For further limitation of evaporation, a lower voltage can be used, electrophoresis can be performed in the cold, the paper strip can be immersed in a liquid immiscible with water (immersed strip technique, Fig. 5 . 3 ~ )[41-431 or the strip can be enclosed between two glass plates which are cooled to the desired temperature (enclosed strip technique, Fig. 5.3d) [5,44,45]. The latter tanks are usually used at higher voltages. as an inverted

PAPER TECHNIQUES

75

Chambers with controlled evaporation of water from the paper strip have also been described and have been used in electrorheophoretic techniques for the physicochemical characterization of proteins [22,23,46,47]. The electrodes used for the construction of the electrophoretic apparatus are made from different materials, such as graphite [48], silverlsilver chloride [49], palladium [50] or platinum. The type of electrode must be selected on the basis of the kind of electrolyte in which the separation is to be carried out. Graphite electrodes easily polarize and disintegrate, leading to contamination of the buffer solutions. Silver/silver chloride electrodes must be reversed after each use and they cannot be used in eledtrolytes that react with the silver or silver chloride. The best results are obtained with platinum electrodes made from wire or foil, which can be used with various electrolytes at different pH (excluding electrophoresis in molten salts, see p. 100). To prevent contamination with electrode products, the paper strip should not be dipped directly into the electrode compartments but into adjacent buffer troughs, an electric contact being made through filter-paper wicks, cotton-wool plugs, agar bridges or a special labyrinth system. The value of electrolyte in the vessels should be sufficiently large to avoid the influence of electrolysis products on the electrophoretic separation. The volume V that the hydrogen ion transports during the experiment can be calculated:

where UHt is the mobility of IT, I (A) is the current, t (s) is the time and K is the conductivity of the solution. With dilute electrolytes and a long separation time, the buffer should circulate continuously between the electrode compartments [ 5 1,521.

Power supply Direct-current power supplies can be built in the laboratory or can be obtained commercially. They must be free of a.c. impurities so as not to disperse the solute ions by convection, and should be capable of providing either a constant current or a constant voltage, preferably both. For most conventional separations a power supply capable of delivering a current up to 50 mA and 500 V is adequate. Special power packs with an automatic safety switch and programmed potential and constant current delivery have been constructed [53-551. The maximal current that can be used for unit strip area can be calculated:

where I(mA) is the current, P (cm’) is the area of the filter-paper strip measured between surfaces of the electrolyte in the compartments and U (V) is the potential.

Densitometers and scanning devices For the quantitative determination of separated substances on the paper strip, both direct and indirect methods can be used. In the former method the substance being

76

ZONE ELECTROPHORESIS

measured must absorb UV light or be fluorescent or radioactive. In the indirect method the separated zones must first be stained and the absorbance measured with a recording densitometer. Several semi- and fully automatic densitometers have been constructed [56-611 that plot transmitted or reflected light from the stained electropherogram plot versus distance along the strip in the form of electrophoretic curves. The area under the curve is then determined automatically by scaled integration, by the use of a planimeter or by cutting out the plotted areas and weighing them. Computerized scanners have also been described [59,62] that allow a complete quantitative analysis of the electropherogram within a few seconds. Low-voltage technique If the potential gradient along the paper strip is less than 20 V cm-' the technique is called low-voltage electrophoresis. Adequate control of evaporation and sufficient resolution can usually be achieved simply by means of a tight-fitting cover o r by cooling the tank with tap water. The filter-paper strip is dipped in buffer solution, blotted and positioned in the electrophoresis tank. The sample is then applied with the aid of a micropipette or a suitable applicator. Two types of spots are employed - drops or streaks. Drops are simple to use when many origins are required, especially in two-dimensional separations. This technique of application is used when the initial concentration of the solute is low and it may be necessary to apply successively a number of drops to one origin, drying between the application of each drop. Streaks are used for a better resolution of close-running substances, where a circular origin may produce spots that overlap after electrophoresis. After the separation is completed, the paper strips are removed from the electrophoresis chamber and dried in an oven at about 100°C. The separated components are then usually located by histochemical staining with an appropriate dye. The excess of dye is washed from the strips with a suitable solvent and the amount of the dye adsorbed by the various components can be determined by elution or by the direct scanning method. Recently, low-voltage paper electrophoresis has been mostly used for the analysis of protein mixtures, particularly blood serum proteins and lipo- and glycoproteins. In this respect paper electrophoresis played an important role in diagnostic investigations where changes in albumin and globulin fractions during pathological processes are of great clinical significance [63-691. The technique is also widely used for the separation of lowmolecular-weight compounds and inorganic ions. A few examples are discussed below.

Proteins and enzymes In human plasma more than 100 different proteins have been identified [70];most of them are present in very low concentrations and cannot be detected electrophoretically even by using very sensitive polyacrylamide gel electrophoresis and immunoelectrophoresis. Using paper electrophoresis, in normal human serum at pH = 8, six different fractions can be separated, viz., albumin and ol-, 0 2 - ,PI-, p2- and y-globulins (Fig. 5.4). The most frequently used buffer solution for the electrophoretic separation of serum proteins for clinical purposes is veronal buffer, pH 8.6 (1.84g of veronal and 10.30g of

PAPER TECHNIQUES

77

Fig. 5.4. Electrophoretic pattern of human blood serum proteins separated on filter-paper, stained with amino black and scanned using a Tecnical A-2 semi-automatic densitometer (Italy).

sodium veronal dissolved in 1 1 of water, ionic strength 0.05). A high resolution of serum proteins can be achieved in Tris-EDTA-boric acid buffer, pH 8.9 (60.5 g of Tris, 6.0g of EDTA and 4.6 g of boric acid dissolved in 1 1 of water). In this solution both LY- and &globulins are each resolved into three subfractions and the prealbumin fraction is also well separated [71]. The use of a pH gradient on a paper sheet [72] and discontinuous buffer systems for the better resolution of proteins and other substances have been used by several workers [73-751 in a method called multiphasic zone electrophoresis [76]. A computer program was prepared for generating the composition and properties of more than 4000 discontinuous buffer systems [77] for the electrophoretic separation ofvarious compounds in different support media. The separated protein components on the paper strip are usually located by staining with an appropriate dye. Commonly used dyes are Amino Black 10B [78,79], bromophenol blue [80], Azocarmine B [81,82], Light Green SF [83] and Ponceau Red 2R [84]. After staining, excess of dye is washed from the strip with a suitable solvent. For more effective destaining, detergent can be added,to the solvent [85] or washing can be carried out at about 80°C [86]. The protein fractions of cerebrospinal fluid [87-891, urine [90-921, gastric and pancreatic juice [90,93-9.51 and other body fluids [96,97] have been analysed by paper electrophoresis in many laboratories. Prior to electrophoresis, the proteins in these fluids must be enriched to 4-670, i.e., 100-to 150-fold, by using concentration dialysis [98], ultrafiltration [99], acetone precipitation [loo], the tannin method [ l o l l or centrifugation [ 1021. The electrophoretic patterns of blood serum proteins and other body fluids in normal subjects and in patients suffering from various diseases may be very different. Because most observations were obtained by paper electrophoresis, it has been found that in many diseases there are often appreciable differences in the mobilities of the various components in addition to those in the concentrations of a particular fraction. It is beyond

78

ZONE ELECTROPHORESIS

the scope of this chapter to discuss the changes in the electrophoretic patterns of serum proteins resulting from diseases. Nevertheless, only general conclusions can be drawn that there are distinct changes during acute and subacute liver diseases, diseases of the kidneys, cancer, genetic disorders and diseases of blood-forming systems (Fig. 5.5). It must be pointed out, however, that no disease-specific value is attached to changes in proteins arising from a particular sickness. Such changes, if observed, must always be assessed in relation to all of the other clinical data.

a

b

C

e

f

9

Fig. 5.5. Characteristic examples of separation of pathological blood serum proteins: (a) normal; (b) acute inflammation; (c) subacute chronic inflammation; (d) cirrhosis of the liver; (e) nephrotic syndrome; (t] p-myeloma; (g) y-plasmocytoma.

The lipoprotein components of blood serum can be detected on paper by staining with Sudan Black B [I031 or Oil Red 0 [104]. Using these procedures, four lipoprotein fractions can be separated and detected: chylomicrons, a-lipoproteins, pre-0-lipoproteins and P-lipoproteins [105]. The best resolution of lipoprotein fractions in blood serum is obtained in agarose gel [lo61 and cellulose acetate membranes [107]. Glycoprotein components are detected on paper by the Schiff-.periodic acid reagent for neutral glycoproteins [ 1081, diphenylamine for mucoproteins and glycosaminoglycans [109], toluidine blue and Azure A for acid glycoproteins [ 1101 and alcian blue for acid glycosaminoglycans [ 1 11, 1 121. Normal sera yield 5-6 glycoprotein fractions, the concentration of which is changed during pathological processes of the liver, kidneys and central nervous system and during tumour diseases. Among the proteins fractionated by paper electrophoresis are haemoglobins [68], cellular proteins [113-1161 and many enzymes. The enzymes can be detected on the paper strip by two different procedures. In the first technique, the enzymic activity is determined on fractions eluted from the wet electropherogram at about 4°C. The second method is the staining technique, which is in widespread use for the location of different

PAPER TECHNIQUES

79

groups of enzymes in various electrophoretic media. The wet electropherogram with the separated enzyme or izozymes is brought into contact with an agar gel layer which contains a suitable substrate and is incubated under the desired conditions. The surface of the gel is then sprayed with a reagent that reacts with the product formed or the excess of substrate to give coloured zones [ 1 171. Tetrazolium salts, particularly nitro blue tetrazolium chloride, have been used for the detection of izoenzymes of lactate dehydro-

Nitro blue tetrazolium chloride.

genase owing to the ease with which they undergo reduction to soluble, but intensely coloured formazans [I 181. By coupling the dehydrogenase reaction with another system, this technique can be extended to the study o f other izoenzymes, such as aspartate aminotransferase [119] . For detecting esterases, 0-naphthyl esters (acetate, phosphate) are used as substrates and the liberated alcohol is coupled with tetrazotized o-dianisidine to produce an intense purple dye [120]. Modifications of the above technique for the detection o f phosphatases [I 211, cholinesterases [122] and other esterases [123] have been described. The direct spectrophotometric determination of dehydrogenases on agar gel layers was described by Wieme [ 1241 and the procedure has been called enzymoelectrophoresis [125, 1261. Many specific tests for detecting different enzymes on paper and in other media have been applied [123, 1271. Paper electrophoresis is in widespread use to control the homogeneity of purified enzymes [128-1311 and for the determination of the isoelectric point [132], izoenzymic composition [ 133,1341 and other properties [ 1351.

Nucleic acids and polynucleotides The best resolution of nucleic acids and nucleoproteins can be achieved in polyacrylamide, agar and agarose gels; paper electrophoresis has occasionally been applied for characterization and identification and to follow the interaction of polynucleotides with other substances [136-1391. Nucleic acids can be stained on a paper strip with Pyronine Y-methyl green reagent [140]. RNA can be differentiated from DNA by prior treatment with RNAase. DNA can also be detected with a modified Feulgen reaction after acid hydrolysis at 60°C. Polynucleotides labelled with radioactive phosphorus (32P),carbon (I4C) or tritium (3H) can be analysed on the filter-paper by a direct radioactivity scanning method.

Compounds of low molecular weight The best separation of low-molecular-weight substances such as amino acids, peptides, nucleotides and sugars is obtained when a high-voltage technique is used (see below).

80

ZONE ELECTROPHORESIS

Nevertheless, many organic and inorganic substances can be fractionated and characterized by low-voltage paper electrophoresis. Sugars and polyalcohols can be separated electrophoretically when they are converted into ionic complexes with boric acid in alkaline solution. Under these conditions chelated oxy derivatives of boron are formed from two &OH groups with tetraboric acid:

R-0-B-OH

\ o ’

0 ‘’

\OH

I

n

m

Negatively charged ionic species produced by complexing of carbohydrateswith boric acid.

The amount of borate that carbohydrates will bind depends on the steric configuration of the monosaccharides and on the number and type of carbohydrate units comprising an oligosaccharide [ 141, 141a]. Monosaccharides and oligosaccharides can be detected on the paper with a variety of reagents but the most frequently used are aniline hydrogen phthalate [I 421 and ammoniacal silver nitrate [143]. The separation of phenols can be accomplished at alkaline pH where the phenolic groups are partially ionized or in an electrolyte consisting of inorganic oxy acids such as boric acid, which forms complexes with polyphenolic compounds with specific structures to confer a negative charge [144]. Ketones and aldehydes can be separated on paper impregnated with NaHCO,, as the a-hydroxysulphonates migrate to the anode in 0.1 M hydrogen sulphite solution [145]. Organic acids are separated in various electrolytes such as acetic acid, ammonia, sodium tetraborate, pyridine-acetate buffers and other solutions [146]. The best resolution is obtained in an electrolyte with a pH near the pK value of the acid. Steroids are separated as sulphate and glucuronide conjugates [ 1471, complexes with succinic acid [148], hydrazones [I491 and phosphoric acid esters [150, 150al. The glucuronide zones can be detected by immersion of the strip in triphenyltetrazolium chloride solution [151] and sulphate conjugates with the Zimmerman reagent [152]. Paper electrophoresis is widely used for the separation of alkaloids [ 153- 1571, antibiotics [ 158- 1621, vitamins and coenzymes [ 163- 1681, sulphur and phosphorus compounds [169-1721, dyes [173-1771 and organometallic substances [178-1821. Inorganic ions These are rapidly separated on paper because they have a comparatively high mobility (about half of the mobility in free solution)[l83]. The best resolution o f inorganic cations is obtained in complexing electrolytes containing C1-, citrate, lactate or formate anions or EDTA. The mobility of the cations depends on the concentration of the complexing anion. At low concentrations of the anion (below 0.01 M), metal ions migrate to the cathode, whereas at higher concentrations of an anion (more than 0.9M) they move to the anode [184]. In order to eliminate adsorption of ions, especially multivalent cations, on the paper, the paper strip can be impregnated with aluminium oxide or chromium

PAPER TECHNIQUES

81

oxide [185]. Inorganic anions can be resolved in solutions of ammonium carbonate [186] or sodium hydroxide [ 1871. Detection of organic ions on the paper can be carried out with a wide variety of reagents, the most commonly used being ammonium sulphide [188], 8-hydroxyquinoline [183], alizarin [ 1891 and silver nitrate [190], or by measuring the radioactive properties of the ions [191, 1921. Paper electrophoresis of inorganic ions is very useful for mobility measurements, for investigation of their complexing properties [ 193, 1941, for preparative separation of isotopes and for desalting of the biological material. High-voltage technique Low-voltage paper electrophoresis has a limited value in the study of low-molecularweight substances because of the diffusion of zones that takes place during the prolonged period required for separation. In the high-voltage technique, a much higher potential difference, up to 200 Vcm-’, is applied for a short period. The heating effect of the applied current is then much greater, and intense cooling of the paper strip is necessary. However, in contrast to low-voltage electrophoresis, almost constant migration rates are maintained during a separation, zone broadening by diffusion is minimal and the resolution of a mixture is much sharper. In addition, this technique often allows separations to be achieved that would not be possible by either the low-voltage method or chromatography.

Apparatus At present, two types of equipment for high-voltage paper electrophoresis are used: (a) immersed strip type, by surrounding the strip with a buffer-immiscible, non-conducting liquid that is denser than water and cooled by means of a cooling coil; (b) enclosed strip type, by placing the strip between two cooled metal plates. (a) The strip or sheet of filter-paper, wetted with buffer solution, is placed in a tank (Fig. 5.6a)[41, 1951 filled with a large volume of liquid coolant such as toluene, carbon tetrachloride, chlorobenzene, heptane, Varsol (Esso white spirit 100) or low-viscosity silicone oil. For better heat dissipation, a special cooling coil, through which tap water or another pre-cooled liquid is pumped, is mounted close to the paper sheet. (h) The enclosed strip type apparatus [44, 1961 is complex (Fig. 5.6b). The moistened filter-paper lies horizontally between two metal plates, usually made of an aluminium alloy, which are machined to achieve uniform flatness to within 0.002 mm. An uniform pressure of ca. 1 atm/cm* (“pressure-plate electrophoresis cell”) is then applied to the plates and cooling is effected by circulating coolant through the labyrinth of channels that have been arranged inside the plates to ensure uniform cooling over their whole surface area. Air at a pressure controlled by a gauge can be fed directly from the laboratory air supply line. Buffer vessels are situated at each end of the plate assembly, equipped with screened platinum electrodes. Contact between paper and buffer compartments is effected by means of paper wicks (four thicknesses of Whatman 3 MM papers) moistened with the buffer solution. lnterlocking relays are fitted to the apparatus on the cover and on the cable connections. If the cover is not properly closed or the cables are

82

ZONE ELECTROPHORESIS

P €3

E

E W

u

o! a

b

Fig. 5.6. Schematic representation of high-voltage electrophoresis apparatus. (a) Immersed paper strip (sheet) type. B = Buffer compartments; P = filter-paper strip; C = cooling liquid; E = electrodes; W = cooling coil. (b) Enclosed paper strip (sheet) type. P = Filter paper; B = fdter-paper wicks; I = polythene insulation foils; C = cooling metal plates; W = cooling coil systems.

disconnected, the voltage supply will automatically switch off. The advantage of this apparatus is the possibility of using it in one- and two-dimensional separations at very high potential gradients. Another type of apparatus has been described in which only one, i.e., the lower, plate is extensively cooled. These tanks were constructed by Wieland and Pfleiderer [197], Werner and Westphal [198],Prusik and Keil [199] and Ostrowski and Krawczyk [200]. A schematic diagram of the latter type is shown in Fig. 5.7. The apparatus permits the use of potentials up to 8000 V and can be easily built at low cost. The thin (2 mm) glass plate (60 x 40 cm) is bent in the form of an inverted V and mounted on a support made of Plexiglas. Two lateral Plexiglas plates well fitted to the glass plate form a box to which a metal tube (2 cm in diameter) is fixed just below the bend of the glass plate. The metal tube contains at its upper end many small holes, acting as an injector when cooled liquid is forced into the tube from a refrigerating system. Injected liquid flowing down along the inner walls of the glass plate removes heat and returns in the closed system to the cooling machine. This cooling system is very efficient and the apparatus is very convenient for one- and two-dimensional separations at high potential gradients. Several d.c. power supplies are commercially available that provide up to 100mA at potentials variable from 0 to 10,000V and that have safety features.

Applications The great advantage of the high-voltage technique is that the very subtle charge differences between different groups of substances can be fully exploited and separated in a very short time. A further advantage is that prior desalting of the sample is unneccessary, as salts rapidly move off the paper during the run and the sample is automatically desalted. Separations can be performed in one direction or in two-dimensional procedures in which electrophoresis is followed by chromatography or again by electrophoresis at the same or a different pH in a second direction (fingerprint technique, diagonaI electrophoresis). A large sheet of filter-paper (60 x 30 cm) also gives the possibility of analysing

PAPER TECHNIQUES

83

a

b

Fig. 5.7. High-voltage electrophoresis tank with one cooling plate [200]: (a) perspective view; (b) cross-sectional view. 1 = Lid from Plexiglas; 2 = filter-paper sheet; 3 = thin glass plate bent as an inverted V; 4 = platinum electrode; 5 = buffer compartment; 6, 7 = metal tube with small holes; 8 = outlet of cooling liquid.

several samples by one-dimensional electrophoresis simultaneously and under the same conditions. The method can be used for qualitative and quantitative separations. For the quantitative analysis of amino acids, special volatile buffer systems [196,201] and developing reagents [202] have been prepared. The main fields of application of high-voltage paper electrophoresis are in the separation and identification of amino acids and their derivatives [203-2081, peptides [2092191, nucleotides [220-2231, sugars [224-2271, phosphoric acid esters [228], urine constituents [229] and inorganic ions [230]. The technique has been used extensively for studies of protein [23 1-2331 and nucleic acid [234-2361 sequences.

84

ZONE ELECTROPHORESIS

Two-dimensional separations Two-dimensional separations using electrophoresis in both directions or electrophoresis followed by chromatography give increased resolutions of different groups of substances compared with one-direction separations using either technique alone. The method is based on the fact that certain charged substances, such as amino acids, peptides and oligonucleotides, migrate with different relative mobilities at different pH values or are characterized by different RF values when chromatography is carried out at right angles to electrophoresis. The two-dimensional separation was introduced by Durrum [237] and Strain [238], who also gave general descriptions of the method now commonly used. There are several modifications of the technique, the most common being (a) two-dimensional electrophoresis, (b) electrophoresis combined with chromatography and (c) simultaneous separation under the influence of two driving forces at right angles to each other.

Two-dimensional electrophoresis This technique can be used for the separation of proteins at low voltages, and especially for the fractionation at higher voltages of mixtures of peptides, nucleotides and other organic and inorganic compounds. The separation is carried out on a paper sheet first in one direction and then at right angles to it. If the separations in both dimensions are performed under identical conditions, no additional resolution will be obtained and all of the components will lie along a diagonal line (diagonal electrophoresis). When resolved components in the first direction are chemically modified on the paper and then separated in the second direction under the same conditions, the modified components of the mixture move away from the diagonal line. Mike5 and Holeybvsky [239] and later Hartley and co-workers [240,241] used two-dimensional electrophoretic separation for the elucidation of sulphur amino acids and disulphide peptides in proteolytic enzymes. The tryptic hydrolysate of the protein was separated on a filter-paper strip at pH 6.5. The peptides on the strip were then oxidized with performic acid vapour by means of which all S-S bridges were converted into the pair of peptides containing cysteic acid residues. The strip was then dried in a current of air to remove excess of performic acid, stitched to another sheet of paper and subjected to high-voltage electrophoresis under the original conditions but at right angles to the first direction. Each peptide should then lie on a 45" diagonal, whereas every cysteic acid peptide having an extra negative charge will lie off the diagonal line, vertically below each other (Fig. 5.8). The advantage of diagonal electrophoresis is that it allows one to recognize at once which pairs of cysteic acid peptides were originally disulphide-bridged, even if they are incompletely resolved. By using the above technique, cystine peptides of immunoglobulin [242,243],prostatic acid phosphatase [244], pancreatic ribonuclease [245], urease [2461 and several proteinases [2472501 have been isolated and analysed. Naughton and Hagopian [25 11 separated tryptic peptides of proteins in the first dimension, treated them with carboxypeptidase B to split off C-terminal lysine or arginine residue, then, after electrophoresis in the second dimension, the peptides that had been modified by carboxypeptidase B moved off the diagonal line and were isolated

85

PAPER TECHNIQUES

0

0 Fig. 5.8. Schematic map of some cysteic acid peptides separated by using the diagonal technique. Series of peptides (I-V) lying below the horizontal line bear a negative net charge and those above the line a positive charge.

and characterized. Two-dimensional paper electrophoresis at different pH values in each direction was used for the separation of oligonucleotides and related substances (see below).

Electrophoresis combined with chromatography This method is widely used for the separation of complex mixtures of amino acids, peptides and nucleotides and for the differentiation of closely related proteins and polynucleotides. Ingram [252] was the first to show the great advantage of combined separation by paper electrophoresis and chromatography of the mixture of peptides obtained by tryptic hydrolysis of haemoglobin A and S. The sample of peptides was placed on a sheet of paper cut as shown in Fig. 5.9 and separated by high-voltage electrophoresis at pH 6.5. After drying the paper in air, ascending chromatography was performed at right angles in n-butanol-acetic acid-water (200:30:75, v/v/v). The two-dimensional map of the

I

I

I

Fig. 5.9. Diagram illustrating how the paper sheet is cut for two-dimensional separation: electrophoresis and chromatography.

86

ZONE ELECTROPHORESIS

separated peptides obtained was called a “fingerprint”, a term that is now commonly accepted. The fingerprint technique was later modified by staining peptides with isatin instead of ninhydrin (Fig. 5.10)[253] or with specific stains for particular amino acids in peptides such as histidine (Pauly test) or arginine (Sakaguchi test)[253a]. Isatin and

n

a

I

0

-0

High-Voltage electrophoresis

b

v

\I/

High-voltage electrophoresis

Fig. 5.10. Fingerprints of tryptic peptides obtained from (a) human and (b) horse haemoglobin. Peptides were first separated by high-voltagepaper electrophoresis at pH 6.4, then ascending chromatography was carried out at right angles in n-butanol-acetic acid-water (3:1:1, v/v/v). Peptides were stained with isatin solution.

specific stains for particular amino acids in the analysed peptides give a greater possibility of differentiating the peptides obtained from very closely related proteins. The conventional procedure can be reversed by carrying out chromatography in the first dimension followed by electrophoresis in the second, and other support media can also be used (for instance, thin layers; see below).

PAPER TECHNIQUES

87

A three-dimensional mapping system was described by Chen and Krause [254]. Peptides from a proteolytic digest are separated initially by electrophoresis at pH 6.5 into acidic, basic and neutral peptides. After locating the three groups of peptides on a guide strip by staining, electrophoresis in the second dimension is performed at pH 2.1 for acidic and basic peptides. The third map is prepared from the neutral peptides by chromatography in the first direction and electrophoresis at pH 2.1 in the second dimension. The three-map system is useful in the study of the homology of large proteins with hundreds of amino acid residues. The fingerprint technique is frequently used to characterize the differences between proteolytic hydrolysates of different haemoglobins [253, 2551, mutant proteins [256], enzymes [257] and high-molecular-weight proteins [258] and for mapping oligonucleotides [259] and chemically modified peptides [260]. Two-dimensional separation with two driving forces acting simultaneously There are several such techniques in which an electric field is combined with an equal potential or with other types of driving forces acting at right angles and leading to twodimensional separations of various mixtures. There are two-dimensional separations with crossed currents employed simultaneously, the curtain technique, centrifugally accelerated electrophoresis and electromagnetophoresis. McDonald and Urbin [261] have described apparatus for two-dimensional separations in which simultaneous crossed currents are used. The electrophoretic chamber has four electrodes, two of which are positive and two negative. When a potential is applied, two forces of equal potential are acting at right angles to each other. The charged particle will then move off at an angle of 45' with respect to the sides of the paper sheet (continuous diagonal electrophoresis). The technique was used for separation of DNP-amino acids and McDonald et a1.[7] were able to calculate the molecular weights of amino acids based on their electroacceleration in the paper. A second technique, in which solute molecules are separated continuously by the flow of electrolyte and electrophoresis at right angles, was described by Svensson and Brattsten [262] and Grassmann and Hannig [263], and will be discussed in Chapter 11. Centrifugally accelerated electrophoresis, described by McDonald et al. [264], depends on the fact that the centrifugal force propels the solution in a straight line from the centre of the sheet of paper to the collecting cups and the electric field, which acts at right angles to the centrifugal field, drifts the charged components to the positive and negative poles. The method has been applied to the separation of mixtures of amino acids. A combination of electric and magnetic fields for the separation of amino acids derivatives was used also by McDonald et al.[7]. The technique was explored later by Kolin [265] and adapted for the separation of neutral and charged particles in free solution. Then an interesting modification of the method was suggested [266,267] for two-dimensional separations in two simultaneously acting fields, i.e., electric and magnetic in filter-paper or in thin layers. Kowalczyk [267], using an electric field at a potential gradient of 20 Vcm-' and a magnetic field of 10,00Ogauss, obtained a distinct deviation of charged molecules and particles (mitochondria, starch particles) with respect to their electrophoretic migration and two-dimensional separation was achieved. This

88

ZONE ELECTROPHORESIS

technique, named electromagnetophoresis, created new possibilities for the zone electrophoresis of charged and neutral particles. Cross paper electrophoresis In this technique, two components of the system are forced by an electric field to cross each other as they migrate towards the cathode and anode. When the paths of the components cross, there will be a deviation from the original direction if there is any interaction between two substances. If there is no interaction, the components will move unhindered along their original paths. Two modifications of the technique have recently been described: one-dimensional and two-dimensional cross paper electrophoresis. Support media other than paper can also be used. General features of the technique are illustrated schematically in Fig. 5.1 1 . In the one-dimensional technique, two reactants, a+ and b-, are applied on the paper strip along lines AB and CD inclined toward the electric field; in a modification, one line is parallel to the field. If the studied substances a+ and b- differ in mobility, they will cross each other and, if there is an interaction between them, a complex ab will be formed that will deviate from the original paths of a+ and b- (crossing at a moving point, Fig. 5.1 1 A) or they will form a “depression” at the crossing (crossing at a fixed point, Fig. 5.1 1 B) on the path of one reactant. The deviation or depression is unequivocal proof of the interaction between a+ and b-, i.e., the formation of a complex ab.

I

I

I

7

Fig. 5.11. Diagrams showing the principle of cross electrophoresis. A , Crossing at a moving point; B, crossing at a fixed point.

A two-dimensional technique is used when one or both reactants are not pure substances. It is analogous to two-dimensional electrophoresis with changing of the buffer solution in the second run. Two-dimensional cross diagrams are usually prepared for immunological studies of pathological serum proteins [268]. For cross electrophoresis, ordinary horizontal and vertical types of equipment can be used. To use cross paper electrophoresis for quantitative purposes, Nakamura [269] adapted the theory of paper chromatography for calculating concentration distribution curves from the depression-deviation spot area. The method of cross paper electrophoresis is commonly used in immunochemical studies [270,271] and studying enzyme-substrate (inhibitor) complex formation [272, 2731 and interactions of proteins with a variety of low-molecular-weight substances [274].

PAPER TECHNIQUES

89

Separations on cellulose acetate and nitrocellulose membranes The use of cellulose acetate membranes as the support medium was suggested by Kohn [275]. Cellulose acetate foil is obtained by treatment of cellulose with acetic anhydride and cellulose triacetate is most commonly used for electrophoresis (Oxoid, 0 x 0 Ltd., London, Great Britain). The membrane (ca. 12 pm thick with diameter of pores ca. 0.4 pm) is a homogeneous material containing only trace amounts of impurities. Adsorption of proteins and other substances on this medium is minimal, so that the tailing of zones is almost eliminated. The strips from cellulose acetate are insoluble in water, alcohols, ethers, dilute acids and alkalis and in most hydrocarbons, but are readily soluble in phenol, glacial acetic acid, dichloromethane and acetone, and particularly in chloroform-ethanol (90: 10, v/v) and dichloromethane-acetone (50:50,v/v). The strips can be rendered translucent by immersion in cottonseed oil, decalin, liquid paraffin or Whitemore oil 120, which is important for optical scanning of stained bands on the foil. They can be impregnated and stored permanently in glycerol, which prevents the destruction of the stained spots. For electrophoretic separations of proteins and nucleic acids, Pfistoupil [276,277] introduced nitrocellulose membranes. These membranes, with a pore size of 0.1-1.2 pm, are produced by various manufacturers (e.g., VCHZ Synthesia, Czechoslovakia) and are commonly used in the analysis of nucleic acids, especially for RNA-DNA and DNA-DNA hybrids. They show selective adsorptivity towards proteins and nucleic acids under appropriate conditions. Nitrocellulose strips ( 5 x 1 cm), when used for protein electrophoresis, must be first treated for 5 min with 2% Tween 60 solution in a buffer composed of veronal, citrate and oxalate, pH 8.6 [278]. Detergent-treated nitrocellulose membranes d o not adsorb proteins and can be used for protein separations at high potential gradients of 15-20 V cm-' for 10-1 5 min. The best stain for proteins on the nitrocellulose foils is nigrosine. At neutral and alkaline pH, RNA migrates rapidly in the foil to the anode, whereas native DNA remains on the start. Denatured, single-stranded DNA migrates to the anode [279]. Nitrocellulose membranes with different pore sizes, impregnated with detergents or solutions of serum albumin, can be used for the separation of proteins, polynucleotides and nucleoproteins [280,281]. The basic design of the electrophoresis tanks used for cellulose membranes resembles that for the horizontal paper technique and excellent ones are commercially available. Some of them have cooling arrangements [37, 1271 and can be used for membranes and thin-layer electrophoresis. Separations can be performed both at low voltages and at high field strengths. To minimize evaporation, glycerol can be added to the buffer solution. Following electrophoresis the strip is dried at 100°C for about 10 min. The staining and detection methods used with various substances on cellulose acetate foils are the same as on filter-paper. Aqueous solutions of dyes are preferred to alcoholic solutions and the dyes usually should have lower concentrations than when applied to paper. Cellulose acetate and nitrocellulose membranes are ideal for absorptiometric and fluorodensitometric examinations and the detection of radioactive substances by autoradiography or by the direct scanning method. Cellulose membranes are used routinely for the separation of plasma proteins [282],

90

ZONE ELECTROPHORESIS

glyco- and lipoproteins [283-2941, haemoglobins [295-2971, enzymes [298-3101 and other proteins [311-3131. The small size (50 x 10mm) of membrane strips available gives the possibility of effecting complete protein analyses on 0.25 pl of blood serum (microzone electrophoresis) in 50 min. Details of the different techniques for electrophoretic separations on cellulose acetate membranes were described by Chin [I071 and on nitrocellulose foils by Pfistoupil [313a]. Separations on ion-exchange papers Jermyn and Thomas [3141 were pioneers in obtaining and applying ion-exchange papers in paper chromatography. Then Huehns and Jakubovic [315] used CM-cellulose powder for the thin-layer electrophoresis of haemoglobin variants. Finally, Yaron and Sober [316] showed that CM-, DEAE- and ECTEOLA-cellulose papers and thin layers can be used for the electrophoretic separation of homologous series of oligopeptides and oligonucleotides. Closely related compounds, differing only in chain length, can be rapidly separated by the two-dimensional technique using electrophoresis in both dimensions or electrophoresis in one direction followed by ion-exchange paper chromatography in the other. In particular, this work has been extended by Sanger and co-workers [317-3201 in the study of complex mixtures of oligonucleotides. Differential retardation induced by an anion-exchange adsorbent combined with electrophoretic migration gives the possibility of examining the composition and sequence of nucleotides. In the method of Sanger and co-workers, initial fractionation of an oligonucleotides mixture is carried out by high-voltage electrophoresis on cellulose acetate strips, usually at pH 3.5 (acetate buffer). The partially separated nucleotides are then transferred to a sheet of DEAE-cellulose paper by placing on a cellulose acetate strip and dissolving it with chloroform-ethanol (9: 1, v/v); electrophoresis is then performed in 7% formic acid (pH 1.9) at right angles. When nucleotides are radioactive, they can be detected autoradiographically and eluted with 20% triethylamine carbonate solution for sequence analysis. Because the migration distance of oligonucleotides on DEAE-cellulose paper is a function of time and chain length, diagrams can be constructed that are useful for the identification of unknown spots under previously calibrated conditions (Fig. 5.12). As can be seen in Fig. 5.12, the spots of oligonucleotides obtained from depurinated DNA are arranged in five series, containing nucleotides with one, two, three, four and five T residues, respectively. Nucleotides containing no or one T residue moved very fast in the 7% formic acid system. The largest nucleotide in the first series was TC4, in the second T2C4,in the third T3C4, in the fourth T4C4 and in the fifth T5C3,etc. Thus, as can be seen from this scheme, some nucleotides can be identified on the basis of their positions on the fingerprint. Murray [32 11 used DEAE-cellulose paper for the electrophoretic separation of deoxyribonucleotides in both directions. Separation in the first dimension was carried out with triethylamine carbonate solution (pH 9.7) and in the second dimension with formic acid-acetic acid (pH 1.9). Brownlee et al. [322] used a different two-dimensional technique to separate oligonucleotides up to 25 residues long. The first stage is highvoltage electrophoresis on a cellulose acetate strip, and subsequent separation on

PAPER TECHNIQUES

91

0

-

2nd

pH 35 (Acetate cellulose strip)

Fig. 5.1 2. Diagram illustrating a two-dimensional electrophoretic separation of the oligonucleotides obtained from depurinated fdDNA [395].

DEAE-cellulose paper was performed by homochromatography using a concentrated mixture of non-radioactive oligonucleotides (commercial yeast RNA dissolved in 9 M urea at pH 7.5) as the eluent. The oligonucleotides in the sample were labelled with 32P. A homomixture of nucleotides saturates the DEAE-cellulose groups, which displace one another to produce a series of fronts. The small nucleotides of lower valency are displaced by the larger ones and thus move faster. The analysed radioactive nucleotides move with the different fronts and are fractionated according to their affinity for DEAE-cellulose ion-exchange paper. Using this procedure Brownlee et al. were able to determine the nucleotide sequence of 5 S RNA of Escherichia coli. In addition to DEAE-cellulose paper, other ion-exchange cellulose papers can be used for the separation of nucleotides and peptides, e.g., polyethyleneimine-cellulose [323] and aminoethyl-cellulose, which have much better wet strengths than DEAE-cellulose paper [321]. Because DEAE-cellulose paper is very fragile, recently DEAE-cellulose

92

ZONE ELECTROPHORESIS

powder has been used more frequently in the form of thin layers for two-dimensional separations. Nevertheless, ion-exchange cellulose papers and powders are commonly used for the fractionation of oligonucleotides and their sequencing [324-33 11. Ultramicroelectrophoresis on single fibres or on thin threads An ultramicro method can be used for the electrophoretic separation and identification of picogram (lO-'*g) amounts of a substance. Edstrom [332,333] described a method for the electrophoretic separation on single cellulose fibres of nucleotides obtained from RNA or DNA of a single cell. The separation is carried out on a cellulose fibre, ca. 20 mm long and 10-20 pm thick, placed on a plate of quartz glass. The fibres before use are treated with 8.1% NaOH, then with viscous buffer containing 4 Nsulphuric acid and glycerol (sp. gr. 1.261) for 50-75 min at 100°C. The quartz plate with the cellulose fibre is then placed in a paraffin oil chamber, the sample is applied by means of a micromanipulator and a potential of 1800 V is applied for 5-1 0 min. The fibre is then photographed under a W microscope at 257 nm, and the photographic plate is examined with a microdensitometer, which registers the light intensities photographically. The method of Edstrom, which with experience gives a reproducibility of more than 95%, has been used for the quantitative analysis of the nucleotide composition of DNA of a single animal cell [334], for the DNA analysis of a single chromosome or even its frag. ments [335]and for the localization of nucleolytic enzymes in subcellular fractions of sea urchin oocytes [336]. Marchalonis and Nossal [337] described an ultramicroelectrophoretic method for the characterization of immunoglobulins isolated from a single myeloma cell. A protein sample isolated from a single cell and rendered radioactive in microdroplets by iodination with '*ST in 5-6Murea was placed under a dissecting microscope ( 5 x lO-'ml) on microstrips of cellulose acetate and separated at 35 Vcm-' for 60 min at room temperature. The buffer solution was 8 M urea in 0.05 M veronal (pH 8.2). After electrophoresis, the strip was washed with distilled water-methanol-glacial acetic acid (5:5:1, v/v/v), which served to fuc protein to the strip and to remove the free '''I. The strip was then exposed to X-ray film and, after development, was analysed with a double-beam recording integrating densitometer. The authors were able to show that a single normal antibodyforming cell also synthesizes a myeloma-like population of identical protein molecules. An ultramicro method for the separation of inorganic cations on cellulose acetate microstrips was described by Tufts [338] and modified by Van Dijk et al. [339]. The microstrips were placed on a piece of wet filter-paper supported by a microscope coverglass. The electrolyte was 5 ml of water containing 1 drop of conc. HCl and 20 mg of NaC1. A potential gradient of 20 Vcm-' was applied for 1.5-2 min. At the cathode, in the slightly acidic solution of NaCl, hydroxyl ions were produced by electrolysis. These ions, migrating towards the anode, met the cations migrating towards the cathode. The cations were then precipitated as oxides or hydroxides in an arc around the cathode at a distance dependent on their precipitation pH. The strip was then sprayed with a saturated solution of rubeanic acid in 50%aqueous methanol. Iron, copper, nickel and cobalt gave yellow, green, blue and pink colours, respectively. Matioli and Niewisch [340] used thin threads of polacrylamide gel for the separation

93

PAPER TECHNIQUES

of haemoglobins in a single erythrocyte. The thread (1 00 pm long and 10pm thick) was placed between two microscope cover-glasses (Fig. 5.1 3) and connected to electrode vessels by means of agarose bridges. A single red cell was placed on the thread under a microscope and a potential gradient of 1000 V cm-' was applied for 4-5 min. The separation was carried out in 0.03M Tris-HCI-glycine buffer (pH 8.5). Separated zones of haemoglobins were observed under the microscope with monochromatic light at 41 8 nm. In normal human erythrocyte several haemoglobins were identified, namely Hb A, AZ,C, S and J. HydCn et al.[341] and Neuhoff and co-workers [342,343] used capillaries fixed in a vertical position (Fig. 5.14) filled with polyacrylamide gel for the separation of proteins and enzymes. Capillaries, ca. 50 mm long and 200pm in diameter (volume 1-5 pl), were loaded with about 3 x lO-"g of protein mixture and fractionated at 80 V cm-' for 15 min. The method was used for the separation of mixtures of different proteins in continuous and discontinuous gel gradient systems [344,345]. The sensitivity of the method can be increased by using tritium-labelled proteins [341].

1-

Fig. 5.1 3. Scheme of the apparatus for ultramicro electrophoresis on a single fibre of polyacrylamide gel (3401. A, Agarose bridges; E, electrode vessels; F, polyacrylamide gel fibre on the microscope cover-glass. Fig. 5.14. Simple arrangement for ultramicro electrophoresis in polyacrylamide gel in a vertical capillary [ 3441.

Separationin non-aqueous solvents Non-aqueous solvents, their mixtures and mixtures of organic solvents with water are being increasingly used for the separation of different substances in electric fields. Electrophoresis in such systems is of both theoretical and practical interest because of the relationship between charge and dielectric constant. In organic buffers, pH-dependent H' ionization of potentially dissociable groups of proteins and other substances can occur and the separation is based mainly on charge differences. Changes in dielectric constant and the conductivity of the solvent may influence the electrophoretic separation by the following phenomena:

94

ZONE ELECTROPHORESIS

(1) it can lead to a decrease in the degree of dissociation of the solute components and thus a better differentiation of the motion of ions in the electric field; (2) medium or low dielectric constants may promote the formation of ion pairs between the charged groups of the solute and counter ions of the solvent; nascent ions of such complexes may be better separated from other components of the mixture that do not react with the solvent ions; (3) the presence of some organic solvents in the medium makes possible the separation of very complex protein and polynucleotide mixtures, for instance proteins of membranes and cellular structures that are insoluble in pure aqueous solutions; (4) because of the low conductivity of organic solvents and the low currents used for separation, high potential gradients can be applied without the need for special cooling arrangements. Pohl and co-workers [346,347] applied strong, non-uniform electric fields for continuous separations of various solid particles in liquid dielectrics. Such a motion of polarizable particles towards the region of greatest field intensity is termed dielectrophoresis [346], as distinct from electrophoresis, which is motion caused by the action of an electric field on a charged object. Dielectrophoresis, which is carried out in free solutions, has been applied to the separation of inorganic particles [347] and cells of living organisms [348] and the concentration of dilute protein solutions [349]. In recent years, two systems of organic solvents have been used for electrophoretic separations of charged molecules: (a) dry, non-aqueous solvents and (b) organic solventwater mixtures. Both systems have been used for the separation of inorganic ions, amino acids, nucleotides, polynucleotides and proteins. (a) Paper electrophoresis in non-aqueous solvents of substances insolubIe in water was first used by Paul and Durrum [350]. They found that substances such as cholesterol, fatty acids and steroid hormones in absolute ethanol and in nitromethane-acetic acid mixtures in an electric field migrate towards the anode, but the components were not resolved sufficiently. Oehme and Rauschenbach [351], using high-voltage paper electrophoresis, separated dyes in organic solvents. It was found that a basic dye such as crystalline violet moved towards the cathode, and an acidic dye such as eosine moved towards the anode in all of the solvents used. The authors found that addition of dimethylformamide to the absolute ethanol caused a decrease in the electroosmotic effect and that at a concentration of 20% (v/v) it is virtually zero. Viscous solutions of sucrose in anhydrous formamide have been employed for the electrophoretic separation of nucleoside phosphates and oligonucleotides [352]. All mono- and oligonucleotides tested migrated towards the anode with good resolution. Liquid ammonia was used for the separation of osmium ion complexes by Preetz and Pfeifer [353]. High-voltage paper electrophoresis was carried out at -50°C in a specially cooled chamber. It was shown that the electrophoretic mobility of [OsBr6]*-was 10%higher in liquid ammonia than in aqueous solvents, and hexachloro complexes were clearly separated from hexabromo complexes in this solvent. (b) organic buffers are most frequently used for the high-voltage paper electrophoresis of amino acids, peptides and nucleotides. They consist of volatile constituents such as pyridine, ammonia, acetic acid, formic acid or non-volatile components such as formamide, phenol, glycerol and polyethylene glycol. The advantage of the volatile buffers is

PAPER TECHNIQUES

95

that they are readily removed from the paper by drying without leaving salts or other impurities. The latter group of organic buffers give an increased resolution of some complex mixtures and minimize evaporation, allowing much higher voltages to be used. Cordon et al. [354], using 30%formamide to which dimethylaminoethanol, urea and acetic acid were added to achieve a pH of 4.0, obtained good resolutions of glutamic acid, aspartic acid, alanine, 0-alanine and y-aminobutyric acid. Brattsten et al. [335] used phenol-acetic acid-water (1 :1:1 or 2 :1:1, v/v/v) for the electrophoretic separation of o-polypeptidyl-tRNA extracted from ribosomes synthesizing proteins. The addition of strong electrolytes, such as cetylpyridinium bromide or sodium bromide, to the above solutions led to the dissociation of nascent proteins and tRNA from ribosomes and differentiation in mobility on the paper of both components studied. Phenol is frequently used as the component of buffer solutions for electrophoretic separations of protein mixtures in paper and other support media [356-3601. Staynov and co-workers [361,362] described the separation of mixtures of RNA and other polynucleotides in deionized formamide by the polyacrylamide gel technique. The formamide was buffered with 0.02M diethylbarbituric acid and adjusted to pH 9.0 by means of 1 N NaOH or 1 N HCl. Separation of the polynucleotides in this medium is mostly based on conformational homology and not on the charge of the ion, and the electrophoretic mobilities are independent of the base composition. In completely dry formamide, double-stranded RNA melts spontaneously at room temperature on dissolution and the melted chains remain stable indefinitely after cooling [363] . Thus, the formamide method of separation of polynucleotides affords a high resolution, better than in aqueous systems, and is analogous in scope to SDS-gel electrophoresis for the study of proteins. The method of Staynov and co-workers has been used for the separation of a-and 0-chain haemoglobin mRNAs, which cannot be separated in aqueous gels [364], viral DNA fragments [365] and poly-A spacers from mRNA obtained from yeast spheroplasts [366, 3671. Kitaoka et al. [368] separated Congo red dyes in aqueous dimethylformamide-sodium chloride solutions. At a ratio of organic solvent to water of 5: 1 (v/v), migration of all components greatly increased, which indicates that the interaction of the dyes with the paper decreased in the presence of higher concentrations of dimethylformamide. Pompowski and Kowalczyk [369] prepared mixtures of acetone, methanol, dioxane, formic acid and water that were characterized by different dielectric constants, and the solutions were used for paper electrophoresis of metal ions. With dielectric constants in the range 30-50, the best resolution was obtained for Fe2+,Fe3+ and Cu+ mixtures. Similar conditions were applied by Ruis and Grass [370] for the separation of compounds of hexavalent uranium. Separation was carried out on the filter-paper in 5 M HC1-ethanol and/or acetone mixtures. Addition of organic solvents increased the formation of complexes between uranium ions and the solvent anions as a result of a decrease in the dielectric constant of the medium.

ZONE ELECTROPHORESIS

96

THIN-LAYER ELECTROPHORESIS General considerations When thin-layer chromatography and I&-layer electrophoresis were introduced several years ago, their convenience and increased sensitivity (1 0- to 50-fold) in comparison with paper techniques seemed to make them promising methods for the analysis of protein and nucleic acid compositions. Thin-layer electrophoresis combined with thinlayer chromatography is now used routinely for the separation of small amounts of amino acids, nucleotides, inorganic ions and other low-molecular-weight substances. The use of thin-layer support media offers the following advantages over the paper technique: (1) it requires much simpler equipment for preparing fingerprints; (2) the resolution is generally superior; (3) separations can be performed at hgh potentials within very short periods; (4) smaller amounts of sample are usually needed and both quantitative and micropreparative work can be carried out more easily; (5) aggressive spraying reagents can be used and the limits of detection are lower. Equipment For thin-layer electrophoresis, conventional horizontal type apparatus can be used, similar to those used in paper, cellulose acetate or agar gel electrophoresis. For separations at higher potentials and for two-dimensional separations, suitable chambers have been described [37 1-3741 that are equipped with cooling systems and other arrangements. The apparatus for thin-layer electrophoresis with a cooled lower plate is shown schematically in Fig. 5.15. A water-cooled flat block made of Dural or brass is insulated by means of a replaceable glass plate 0.8 mm thick or Melinex film (Imperial Chemical Industries, London, Great Britain). The thin-layer plate is placed on the insulated block and paper wicks (Whatman 3MM), cut to overlap each side of the thin layer by 1 cm, make contact with buffer troughs. A polythene sheet, 250-300pm thick, placed between the thin layer and the lid, protects the thin layer from drops of condensed water vapour. A sheet of plate glass, 5 mm thick, covers the whole chamber. The short distance between the thinlayer plate and the cover plate acts as a moist chamber and prevents the thin layer from drying out. When thin-layer electrophoresis is carried out on pre-coated plastic foils (Eastman Chemagram cellulose, MN Polygram 300 cellulose), an immersed type of apparatus can be

f

f

Fig. 5.15. Chamber for thin-layer electrophoresis. (a) Water-cooled flat plate made of Dural; (b) Melinex insulating foils; (c) thin layer on a glass plate; (d) paper wicks (Whatman 3 MM); (e) plate glass, 5 mm thick; (f) buffer compartments with electrodes; (w) cooling coil system.

97

THIN-LAYER ELECTROPHORESIS

used as described by Gassen [375]. When longer chromatographic runs are needed, a chamber with a continuous flow of the solvent over the thin-layer plate can be used [376]. For rapid separations on thin layers a simple apparatus for circular electrophoresis has been described [372]. To obtain the sample from the thin-layer electropherogram in a small volume for further analysis, a special technique was developed [375] . The tip of a 0.4-ml plastic testtube is cut off to leave a small hole in the bottom and a small pad of glass-wool is pressed into the tube to act as a filter (Fig. 5.16). The tube is connected to a,vacuum line and the thin layer, which has been loosened around the spot with a spatula, is sucked into the tube. The tube containing the sample is placed on the top of another 0.4-ml tube and loop1 of a suitable solvent are poured into the top tube. The whole assembly is shaken, then the extraction is repeated with an additional loop1 of solvent. The bottom tube contains the eluted spot in a 2OO-pl volume, which can be concentrated to dryness by freeze-drying, if necessary. W-absorbing substances can be scanned directly on thin-layer fluorescent plates with suitable filters (F254 fluorescence indicator plates; Merck, Darmstadt, G.F.R.) [377]. This is not a precise method because of the irregularities in the thckness of the layer, the distribution of the fluorescent indicator and the shape of the spots [378], but it is very useful for rapid examinations and rough estimations. Eluent (100 p c )

Cellulose powder Cotton plug

Fig. 5.16. Apparatus for elution o f spots from cellulose thin layer [375].

One-dimensional separation The most representative papers describing this technique concern the separation of amino acids, nucleotides and inorganic ions. Katz and Lewis [371] separated amino acids on cellulose-coated plates in 0.1 5 M formic acid (pH 2.8) or in 0.01 M ammonia-0.0033M acetic acid buffer (pH 10.2). Relative mobilities for all amino acids in both solutions and in other buffers were calculated [379]. Cellulose-coated plates have been used in the one-dimensional technique for the separation of ribonucleotides [380,381]. In 0.1 M ammonium formate (pH 2.5) trinucleotides were well resolved, then eluted from the thin layer and sequenced. Baguley and Staehelin [380j separated tRNA oligonucleotides on cellulose plates in the same buffer solution at 50 V cm-'. Thin-layer plastic foils (Eastman Chromagram; Eastman-Kodak, Rochester, N.Y., U.S.A.) have been applied to the separation of trinucleotides derived from T1 ribonuclease hydrolysates of tRNAPhe

98

ZONE ELECTROPHORESIS

from E.coli [375]. A similar technique has been applied for the separation of polynuclear aza-heterocyclic compounds [382] and the vitamin B6 group, including 5-hydroxyindoles [383,384]. Muir and Jacobs [385] used compressed glass-fibres (Whatman GF/B) as the thin layer at pH 7.2 for the preparative separation of protein-polysaccharides of chondroitin 4-sulphate extracted from pig laryngeal cartilage. Criddle et al. [386] and Pastuska and Trinks [387] separated synthetic dyes by electrophoresis on thin layers of alumina, kieselguhr and silica gel. Inorganic cations have been separated in thin layers of cellulose powder, silica gel “Fertigplatten” or kieselguhr, using as the solvents 0.01 N HCl for iridium compounds [388] and 0.05M lactic acid, 0.1 N NaOH and 0.15M citric acid for most inorganic cations and anions [389]. Two-dimensional separation Three different modifications can be applied: electrophoretic separation in both dimensions; electrophoresis combined with thin-layer chromatography; and electrophoresis combined with thin-layer gel filtration. Thin-layer techniques have proved very useful for preparing peptide maps or fmgerprints of proteins to be characterized. Glass plates (20 x 20 cm) coated with the silica gel are most frequently used for this purpose. A tryptic digest, 0.05-0.5mg of the protein being analysed is applied at a corner of the plate in the form of a spot not exceeding 4 mm in diameter (1 pl), then dried. The plate is sprayed with a selected buffer and subjected to electrophoresis in the first direction. In some laboratories chromatography is performed as the first step. Following electrophoresis, the plate is heated for 10 min at 100°C, then ascending chromatography is carried out within about 60 min. The plate is dried again and sprayed with developing reagent or autoradiographed. Bates et al. [374] described a method for labelling peptides by the amidination reaction using methyl [‘“C] acetimidate. Sub-nanomole amounts of peptides can then be detected on a plate. Cook and Bieleski [390] used a mixture of cellulose powder and silica gel (25:10, w/w) for the separation of amines and amino acids. Polygram Cel300 or 400 thin layers have been applied for peptide mapping and sequence determination of ribosomal protein from E.coli [391]. Mixtures of nucleotides, nucleosides and bases derived from DNA were well resolved on cellulose-coated plates [392] when electrophoresis was carried out in 0.05 M formate buffer (pH 3.4) and chromatography in ammonium sulphate-sodium citrate-isopropanol solution. Silica gel G plates and MN-cellulose or Whatman cellulose CC-41 were applied for fractionation of the complex mixture of phosphoric acid esters [393]. Ionexchange thin-layer chromatography in combination with electrophoresis on Cellogel strips gave satisfactory results in separation and sequencing studies on oligonucleotides. After separation on Cellogel first, the strip is placed on a pre-wetted thin-layer plate, firmly smoothed and chromatography or electrophoresis is carried out in the second direction. The separation of higher oligonucleotides is usually performed in pyridine-acetic acid buffer (pH 3.5-4.5) containing 7Murea [394]. The use of urea in the buffer overcomes interferences caused by secondary, non-ionic binding forces. The addition of 7 M urea to the buffer systems in ion-exchange thin-layer electrophoresis leads to a marked increase

99

THIN-LAY ER ELECTROPHORESIS

in the mobility and resolution of the hexa-, hepta- and octanucleotides [3 161. In the twodimensional mapping procedure the best results were obtained when the oligonucleotides were labelled at the 5tterminus as described by Szekely and Sanger [395]. In this method, radioactive phosphate from y-32P-labelled ATP is transferred to the 5tOH group of each nucleotide by the action of polynucleotide kinase. One part of the oligonucleotides sample is first treated with bacterial alkaline phosphatase, which hydrolyses 3'- and 5'phosphate groups but does not attack 2':3'-cyclic phosphates. RNA can be hydrolysed by pancreatic ribonuclease A, by ribonuclease TI or by both enzymes at the same time. Under these conditions, series of (pN)npN-OH and (pN)npNp isopliths can be obtained and each of the oligonucleotides occupies a unique and identifiable position in the fingerprint [323,396-3991. Proteins and other macromolecular substances can be separated by using thin-layer gel filtration-thin-layer electrophoresis. According to the molecular weights of the sample, Sephadex Superfine (3-25 to G-200 can be applied. The Sephadex, equilibrated with an appropriate buffer, is spread on a glass plate in the usual way. The thin-layer plate is placed in a suitable apparatus [400], shown in Fig. 5.17, and equilibrated by allowing 0

C

i

s

Fig. 5.1 7. Apparatus for two-dimensional separation using gel filtration and electrophoresis on Sephadex thin layers [400]. (a) Supporting system; (b) slits for filter-paper connections; (c) watercooled brass platform; (d) trough containing the buffer or solvent used for the development of the electropherogram; (e) stand with arrangements for changing position of the supporting plate; (f) holes for the screw g.

the buffer to flow through the gel layer. A sample of protein solution is then applied by means of micropipette at the upper corner of the plate. The layer should be connected with a buffer reservoir using a sheet of Whatman No. 1 filter-paper at the upper end, and the plate is inclined so as to form an angle of 15-20"with the horizontal. A sheet of Whatman 3MM paper is attached to the lower end of the layer for removal of the filtrated fluid. After the filtration is completed, the plate is arranged horizontally and electrophoresis is performed at right angles to the first direction. To obtain a print of the protein

100

ZONE ELECTROPHORESIS

spots, the plate is carefully covered with a sheet of Whatman No. 1 paper and after 5 min the contact sheet is removed from the thin layer, dried at 100°C and stained with amino black or another protein dye. Dextran gel filtration-thin-layer electrophoresis has been used for the separation of serum proteins [400], different tetrazolium dyes [401], urinary amino acids and peptides [401a] and other substances [401b]. Electrophoretic separation of inorganic ions in fused salts The determination of the electrophoretic mobility of inorganic ions creates many possibilities for obtaining information on the electric charge of the ion, the nature and size of the ion, the nature and extent of solvation, etc. One of the methods by which the above properties of inorganic ions can be studied is electrophoresis in molten salts, which can be used for the investigation of cations, anions and the separation of isotopes. Electrophoretic separations in fused salts can be carried out on paper strips resistant to high temperature, on thin layers and in columns. The most frequently used supports are paper strips made from glass-fibres [402,403], asbestos fibres [404] and quartz fibres [405]. For thin-layer and column electrophoresis, ceramic oxides, glass powder and quartz powder have been used. All of these materials must first be well washed with a strong mineral acid, then with distilled water to remove the excess of acid and finally with a concentrated salt solution, for instance 4-6M KNOJ at pH 8-9 [406]. Such a treatment of the strip eliminates impurities and ensures better wetting by the melt at the temperature of the experiment. Electrophoresis on paper strips and in thin layers in molten electrolytes is carried out in a special tank made of Pyrex glass (Fig. 5 .18). The chamber is placed in a thermoregulated furnace and, during passage of current, the chamber must be continuously

Fig. 5.18. Cross-sectionalview of apparatus for electrophoresis in fused salts (modified from ref. 412). (a) Electrophoresis chamber; (b) Pyrex glass plate; (c) Pyrex vessels; (d) glass-fibre paper; (e) capillary with a screw to move it in all directions; (f) furnace with heating and temperature-regulating system; g.i. = gas inlet tube; g.0. = gas outlet tube.

washed out with dry neutral gas in order to remove the gaseous products evolved at the electrodes. The choice of the electrodes depends on their resistance to corrosion under the experimental conditions. Platinum wire or foil can be used as the anode, while materials such as tungsten, nickel, copper and graphite of suitable purity have been used as the cathode, as platinum may be attacked by the alkali metal produced during the electrolysis of the melt. Before application of the sample, the impregnated support must be equilibrated for about 1 h with the molten salt of the reservoirs with an applied

101

THIN-LAYER ELECTROPHORESIS

electric field to achieve the same temperature of the strip and the surrounding atmosphere. The sample to be studied is placed on the strip by means of a capillary, the tip of which can be moved in all directions by an external device fixed in the wall of the chamber, As the solvents, molten alkali metal (Li, Na, K) nitrates and chlorates and also their eutectics have been widely used on account of their relatively low melting points. Electroosmotic flow in molten salts is believed to be very small because the electrical double layer is negligible at the temperatures used [407]. A small electroosmotic movement of the molten salts is independent of the type of support used and depends on the characteristics of the salt alone. The solvent flow in molten salts cannot be easily measured because of the lack of suitable markers. Most of the work up to now has been concerned with the determination of the mobility of metal ions, particularly alkali metal ions. To determine the electrophoretic mobility of metal ions, the oxides of which are almost insoluble in most of the fused solvents, the addition to the molten salts of a suitable amount of an ammonium salt is necessary, which acts as a strong solubilizing agent for metal oxides [408]. By tlus method it was possible to determine the electrophoretic mobility of different ions, even di- and trivalent ions in nitrate melts (cf. Table 5.1). Lithium ions derived from molten lithium salts have a strong influence on the mobility of other alkali metal ions in the fused salts. It was demonstrated that the mobility of Na', K+ and Li' as the dissolved metal ions increases with an increase in the molar fraction of LiN03 in the melt [409]. This phenomenon can be explained by assuming that the Li' ion, having a small Stokes radius, interacts strongly with the anions of the solvent, decreasing their activity. Under these conditions, the interaction of other alkali metal ions with the anions of the solvent becomes smaller and their cationic mobility increases [410]. The different extent of interaction of the metal ions of the solute with anions of the solvent causes the metal ions to exhibit a neutral or even anodic mobility. On the other hand, the electrophoretic mobility of inorganic anions dissolved in fused salts is generally lower compared with other media, especially when Li+ ions are present in the melt [411,412]. Some anions move towards the cathode or remain at the point of application (cf. Table 5.1). TABLE 5.1 ELECTROPHORETIC MOBILITIES OF SOME INORGANIC CATIONS AND ANIONS IN MOLTEN SALTS Prepared from data published by Alberti and Allulli [412]. + indicates movement towards the cathode and - towards the anode. No.

Ion

Melt composition (mole-%)

Temperature ("C)

(u

NaN0,-KNO, (45.5:54.5) NaNO,

250

+ 2.24

350

+ 3.16

NaN0,-KNO,

270

t 1.77

Cations

1

Li

2

Na

3

Na'

+

+

(50:SO)

Mobility

x

Support

10')

Asbestos paper Glass-fibre paper Ceramic thin layer

(Continuedon p. 102)

I02

ZONE ELECTROPHORESIS

TABLE 5.1 (Continued) No.

4

Ion

K‘

Melt composition (mole-%)

Temperature (”C)

(U x 10‘)

,

450

+ 3.16

,

270

+ 4.47

NaN0,-KNO, (5O:SO) 10% NH,NO, NaNO ,-KNO, (50:SO) NaN0,-KNO, (5050) NaN0,-KNO, (5O:SO) NaNO ,-KNO (5050) NaNO ,-KNO, (5O:SO) LiCI-KCI (59:41) KCNS

250

+ 0.9

270

+ 1.6

270

+ 0.1

2 70

i0.45

350

+ 1.25

270

+ 0.79

450

+ 1.9

210

- 0.14

250

- 0.25

250

+ 0.25

250

i 0.15

Glass-fibre paper

250

+ 0.45

Glass-fibre paper

255

+ 0.28

250

+ 0.8

250

- 0.7

NaNO ,-KNO (5O:SO)

5

Rb’

NaN0,-KNO (5O:SO)

6

CS’

7

TI+

8

Mg’+

9

Ca2+

10

Sr2+

11

Ba2+

12

Mn”

13

co2+

14

Ni”

15

cu2+

16

Zn’+

17

Cd”

18

Sn’+

19

Pb”

Anions 20

c1-

21

Br-

22

NO;

23

NO;

24

NO;

+

,

NaN0,-KNO, ( 5 0 5 0 ) + 10% NH,NO, NaNO ,-KNO , (50:SO) + 10% NH,NO, NaN0,-KNO, (5O:SO) + 10% NH,NO, NaN0,-KNO,

+

(5050) 10% NH4 NO, LiNO ,-KNO (43: 5 7) NaN0,-KNO, (5O:SO) + 10% NH,NO,

,

NaNO ,-KNO (5O:SO) NaNO ,-KNO, (5O:SO) LiCIO, -KCIO, (76:24) NaNO ,-KNO, (50: SO) KNO,

,

270 300

275 450

Mobility

Support

Glass-fibre paper Ceramic thin layer Glass-fibre paper Ceramic thin layer Ceramic thin layer Ceramic thin layer Asbestos paper Ceramic thin layer Glass-fibre paper Glass-fibre paper Glass-fibre paper Glass-fibre paper

Glass-fibre paper Glass-fibre paper

Glass-fibre paper - 0.9 Ceramic thin layer + 2.6 Glass-fibre paper - 0.95 Glass-fibre paper - 2.45 Glass-fibre paper (Continued on p . 103)

103

REFERENCES TABLE 5.1 (Continued) No.

Ion

25

cr0:-

26

Cr,O:-

Melt composition (mole-%)

Temperature C)

e

Mobility x lo4)

Support

(u

LiC1O4-KC10, (76 :24) NaN0,-KNO, (5 0: 50)

300

+ 0.9

250

- 0.8

Glass-fibre paper Glass-fibre paper

The technique of electrophoresis in fused salts is very useful for separations of isotopes. Because the ions in the molten salts are in an anhydrous state, the relative mass differences of the isotopes are higher in fused solvents than in aqueous solutions, and it is thus possible to obtain greater differences in their electrophoretic mobilities. For example, when isotopes of lithium (6Li-'Li), sodium (22Na-24Na), rubidium (ssRb-87Rb) and caesium ('31Cs-137Cs) were fractionated in the molten nitrates, a sufficient resolution was achieved with high enrichment factors [413,414].

REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

20 21 22 23 24 25 26 27

A. Tiselius, Trans. Faraday SOC., 33 (1937) 524. P. Konig, Actas e Trabalhos Congr. Sud.-Amer. Chim., I l l Congr., Rio de Janeiro, 2 (1937) 334. T. Wieland and E. Fischer, Narurwissenschaften, 35 (1948) 29. H. G. Kunkel and A. Tiselius, J. Gen. Physiol., 35 (1951) 89. H. D. Cremer and A. Tiselius, Biochem. Z., 320 (1950) 273. W. Grassmann and K. Hannig, Hoppe4eyler's 2. Physiol. Chem., 290 (1952) 1. H. J. McDonald, R. J. Lappe, E. P. Marbach, R. H. Spitzer and M. C. Urbin, lonogruphy: Electrophoresis in Stabilized Media, Year Book Publ., Chicago, Ill., 1955. J. T. Edward, J. Chromatogr., 1 (1958) 446. J. C. Giddings and J. R. Boyack, J. Theor. Biol., 2 (1962) 1. J. R. Boyack and J. C. Giddings, Arch. Biochem. Biophys., 100 (1963) 16. J. C. Giggings and J. R. Boyack,Anal. Chem., 36 (1964) 1229. J. T. Edward, Advan. Chromatogr., 2 (1966) 63. J. L. Engelke and H. H. Strain, Anal. Chem., 26 (1954) 1872. J. Kowalczyk, J. Chromatogr., 14 (1964) 41 1. S. E. Wood and H. H. Strain, Anal. Chem., 26 (1954) 1869. E. W.Bermes, Jr. and H. J. McDonald, J. Chromatogr., 4 (1960) 34. J. D. Raacke and C. H. Li, J. Biol. Chem., 215 (1955) 277. J. D. Raacke,J. Amer. Chem. SOC.,80 (1958) 3055. M. Macheboeuf, P. Rebeyrotte, J. M. Dubert and M. Brunerie, Bull. SOC.Chim. BioL. 35 (1953) 334. M. Macheboeuf, J. M. Dubert and P. Rebeyrotte, Bull. SOC.Chim. Biol.. 35 (1953) 346. R. Audubert and S. de Mende, The Principles of Electrophoresis, Hutchinson, London, 1959, p. 60. K. Schilling and H. Waldmann-Meyer, Acta Chem. Scund., 13 (1959) 1. H. Waldmann-Meyer and K. Schilling, Acta Chem. Scund., 13 (1959) 13. H. Waldrnann-Meyer, Biochim. Biophys. Acta, 261 (1972) 148. Y.Kiso, M. Kobayashi, Y.Kitaoka, K. Kawamoto and J. Takada, J. Chromatogr., 36 (1968) 215. H. H. Strain and T. R. Sato,Anal. Chem., 28 (1956) 687. H. Waldmann-Meyer, Methods Biochem. Aiial., 13 (1965) 47.

104

ZONE ELECTROPHORESIS

28 W. Preetz, Fortschr. Chem. Forsch., 11 (1969)375. 29 W. Preetz and H. Homburg,J. Chromatogr., 54 (1971) 115. 30 H. J. McDonald, R. J. Lappe. E. P. Marbach, R. H. Spitzer and M. C. Urbin, CZin. Chem., 5 (1953) 35. 31 R. Gijbels, L. de Meyer and J. Hoste, Anal. Chim. Actu, 31 (1964) 159. 32 H. G. Kunke1,Methods Biochem. A n d , 1 (1954)141. 33 J. C. Giddings,J. Chem. Phys., 31 (1959) 1462. 34 F. G. Williams, E. G. Pickels and E. L. Durrum, Science, 121 (1955)829. 35 F. V. Flynn and P. de Mayo, Lancet, 261 (1951)235. 36 G. B. Marini-Bettolo and J. A. Coch-Frugoni, J. Chromutogr., 1 (1958)182. 37 L. Darlage,PToc. Indiana Acad. Sci., 76 (1967)223. 38 J. D. Herron,J. Chem. Educ., 46 (1969)527. 39 D.Racusen and L. H. White, J. Chem. Educ., 49 (1972)439. 40 H. Przybyh and Z. Kwapniewski, Zesz. Nauk. Politech. Krakow., Chem., 3 (1968) 11. 41 H. Michl, Monatsh. Chem., 82 (1951)489. 42 F. Turba, H.Peltzer and H.Schuster, Hoppe-Seyler's Z. Physiol. Chem., 296 (1954)97. 43 A. M. Katz, W. J. Dreyer and C. B. Anfinsen, J. Biol. Chem.,'234(1959)2897. 44 D. Gross,J. Chromutogr.., 5 (1961)194. 45 K. Dose,J. Chromatogr., 2 (1959)575. 46 N. F. Hollinger and R. K. Lausing, J. Chromatogr., 5 (1961)38. 47 R. H.Hackman and M. Goldberg, And. Chem., 36 (1964) 1220. 48 B. Paletta, CZin. Chim. Acta, 5 (1960)490. 49 G. T. Franglen, N. H. Martin and J. D. Treherne, J. Clin. Parhol., 8 (1955)144. 50 R. Neihof, J. Colloid Interface Sci., 30 (1969) 128. 51 G. Schneider and G. Sparmann, Naturwissenschuften, 42 (1955) 156. 52 J. B. Sieh and C. Sterling, Biochim. Biophys. Actu, 184 (1969)281. 53 P. D. Hoagland, A n d . Biochem., 26 (1968) 194. 54 H.E. Schaffer and F. M. Johnson, Anal. Biochem., 51 (1973)577. 55 I. M.Williams and N. Catsimpoolas, Anal. Biochem., 71 (1976)555. 56 W. Ostrowski and A. Mikucki, Actu Physiol. Polon., 3 (1952)277. 57 K. H. Kimbel and H. Bunte, Klin. Wochenschr., 33 (1955) 187. 58 L. Gbrski and K. Wbjcik, Chem. Anal., 3 (1958) 137. 59 W. Schmidtmann, R. Kaempffer and L. Reschke, Aerztl. Lab., 15 (1969)165. 60 J. Winkelman and D. R. Wybenga, Clin. Chem., 15 (1969)708. 61 R. Tykva and F. Franek, Anal. Biochem., 78 (1977)572. 62 M. D.Poulik and L. Pinteric, Nature (London), 176 (1955)1226. 63 H. J. Antweiler, Die Quantitative Elektrophorese in der Medizin, Springer, Berlin, Gottingen, Heidelberg, 1957. 64 J. A. Owen,Advan. Clin.Chem., 1 (1958)237. 65 D. H. Moore, in M. Bier (Editor), Electrophoresis, Vol. 1, Academic Press, New York, 1959, p. 369. 66 L. P. Ribeiro, E. Mitidieri and 0. R. Affonso, Paper Electrophoresis, Elsevier, Amsterdam, 1961. 67 F. Scheiffarth, G. Berg and H. Gotz, Papierelektrophorese in KIinik und Praxis, Urban und Schwarzenberg, Munich, Berlin, 1962. 68 L. A. Lewis, Electrophoresis in Physiology,Charles C. Thomas, Springfield, Ill., 1960. 69 L. P. Cawley, H. P. Chin. E. Epstein and R. L. Xelsey, Clinical E1ectrophoresis:A Manual, Gelman Instrument Co., Ann Arbor, Mich., 1969. 70 H. E.Schultze and J. F. Heremans,Molecular Biology of Human Proteins, Elsevier, Amsterdam, 1966. 71 T.Aronsson and A. Gronwall, Sci. Tools, 5 (1958)21. 72 M. V. Narurkar, L. M. Narurkar and M. B. Sahasrabudhe,AnaL Biochem., 26 (1968)174. 73 M. D. Orr, R. L. Blakley and D. Panagou, Anal. Biochem., 45 (1971)68. 74 D. Racusen and M. Foote, h a l . Biochem., 25 (1968) 164. 75 J. N. Miller, A. H.Newton and D. T. Burns, Clin. Chim. Actu, 31 (1971)427.

REFERENCES

105

76 T.M.Jovin, Biochemistry, 12 (1973)871,879and 890. 77 T. M. Jovin, M. L. Dante and A. Chrambach, Multiphusic Buffer Systems Output, Federal Scientific and Technical Information, U.S. Dept. of Commerce, Springfield, Va., 1971. 78 W. Grassmann and K. Hannig, Angew. Chem., 62 (1950)170. 79 E. Kawerau. Anulyst (London), 79 (1954)681. 80 E. L. Durrum.1 Amer. Chem. SOC.,72 (1950)2947. 81 H. Pliickthum and H. Gotting, Klin. Wochenschr., 29 (1951)415. 82 G. Korver, Klin. Wochenschr., 29 (1950)693. 83 L. A. Rideout and R. W. Prichard, Science, 121 (1955)374. 84 H. Rottger, Nuturwissenschuften, 39 (1952)451. 85 K. Heide and H. Biel, Aerztl. Lob., 2 (1956)360. 86 2.Puhr, Hoppe-Seyler's Z. Physiol. Chem., 296 (1953)62. 87 F.C.Courtice, in M. Bier (Editor), Electrophoresis, Vol. 2,Academic Press, New York, 1967, p. 241. 88 A. Lowenthal, Agar Gel Electrop-ho&finNeurology, Elsevier, Amsterdam, 1964,p. 18. 89 H. Bauer, Angew. Chem., 65 (1953)90. 90 G. B. J. Glass, in M. Bier (Editor), Electrophoresis, Vol. 2,Academic Press, New York, 1967, p. 311 91 L. Bonomo, U.Gillardi, F. P. Schena, L.Pinto and F. Dammacco, Recenti hog. Med., 42 (1967) 215. 92 H. Wachter, G. Sallaberger, W.Guetter and W. Raffeiner, Protides Biol. Fluids, 16 (1969)597. 93 G.B. I. Glass, M. Rich and L. Stephanson, Gastroenterology, 34 (1958)598. 94 G . B. J. Glass, Advun. Clin. Chem., 7 (1964)373. 95 B. J. Haverback, B. Dyce, H. Bundy, H. F. Wirtschaffer and H. A. Edmondson, J. Clin. Invest., 41 (1962)972. 96 S. Shulman and P. Bronson, Fertil. Sten'l., 19 (1968)549. 97 T. Katsuki, H. Shimura, C. G. Johnston and H. Miyake, Proc. Soc. Exp. Biol.Med., 103 (1960) 272. 98 F. W. Aly, Biochem. Z . , 325 (1954)505. 99 T. Aronson, Scund. J. Clin. Lab. Invest., 9 (1957)95. 100 T.Bucher, D. Matzelt and D. Pette, Klin. Wochenschr., 30 (1952)325. 101 W. Mejbaum-Katzenellenbogen,Act4 Biochim. Polon.. 6 (1959)375. 102 R. C.R. Barreto and M. B. Mano, J. Chrornatogr., 7 (1962) 346. 103 B. Swahn, Scand. J. Clin. Lab. Invesr., 4 (1952)98. 104 H. Weller, Klin. Wochenschr., 36 (1958)563. 105 R. S. Lees and F. T. Hatch, J. Lab. Clin. Med., 61 (1963)518. 106 A. J. Houtsmuller, Agarose Gel Electrophoresis of Lipoprotiens, Van Gorcum, Assen, 1969. 107 H. P. Chin, Cellulose Acetate Electrophoresis, Ann Arbor-Humphrey Science Publ., Ann Arbor, Mich., London, 1970,p. 55. 108 J. M.Matthieu and R. H. Quarles, A n d . Biochem.. 55 (1973)313. 109 A. J. Anderson and N. F. Maclagen, Biochem. J., 59 (1955)638. 110 J. Takeuchi, Stain Technol., 36 (1961)159. 1 1 1 R. E. S. Prout, Biochim. Biophys. Actu, 177 (1969)157. 112 D. J. Newton, J. E. Scott and P. Whiteman, Anal. Biochem., 62 (1974)268. 113 H, E. Homann and K. H. Nierhaus, Eur. J. Biochem.. 20 (1971)249. 114 D. Kabat, Biochemistry, 10 (1971) 197. 115 H. A. Klein and S . 0choa.J. Biol. Chem., 247 (1972)8122. 116 F.C. H.Chon, C. H. J. Chou, R. Shapira and R. F. Kibler,J. Biol. Chem., 251 (1976)2671. 117 C. L. Market and F. MoUer, Proc. Nut. h a d . Sci. U S . . 45 (1959)753. 118 W.Gerhardt, J. Clausen, E. Chrostensen and J. Riishede, Actu Neurol. Scund., 39 (1 963) 85. 119 T. R. C. Boyde and A. L. Latner, Biochem. J., 82 (1961)51P. 120 G.Gormori, Microscopic Histochernistry, Chicago Univ. Press, Chicago, Ill., 1952,p. 21 1. 121 D. W.Moss, Nature (London), 193 (1962)981. 122 J. Bernsohn, K. D. Barron and A. R. Hess, Proc. Soc. Exp. Biol. Med., 108 (1961)71.

106

ZONE ELECTROPHORESIS

123

G.1.Brewer and C. F. Sing, An Introduction to Isozyme Techniques, Academic Press, New

124 125 126 127 128 129 130 131 132 133 134 135 136

York, London, 1970, pp. 62-137. R. J. Wieme, Clin. Chim. Acto, 4 (1959) 46. R. J. Wieme and Y. Van Maercke, Ann. N.Y.Acad. Sci., 94 (1961) 898. J. Kamaryt and Z. Zhzvorka, Sci. Tools, 10 (1963) 21. R. J. Wieme, Agar Gel Electrophoresis, Elsevier, Amsterdam, 1965, pp. 148-165. T. Ueda, E. T. Lode and M. J. Coon,J. Biol. Chem., 247 (1972) 2109. E. Kuhn and K. Brand, Biochemistry, 11 (1972) 1767. T. H. Chin and D. S. Feingold, Biochemistry, 8 (1969) 98. G. L. Dale and G. Popjak, J. BioL Chem., 251 (1976) 1057. A. M. Crestfield and F. W.Allen,J. Biol. Chem.. 211 (1954) 363. 0. Tsolas, J. De Castro and B. L. Horecker,Arch. Biochem. Biophys., 143 (1971) 516. J. S. Schutzbach and H. Anke1,J. Eiol. Chem., 247 (1972) 6574. J. Larner, C. Villar-Palasi and N. E. Brown, Biochim. Biophys. Acta, 178 (1969) 470. Y. Lapidot, D. Elat, S. Rappoport and N. De Groot, Biochem. Eiophys. Res. Commun., 38

137 138

(1970) 559. G. Beaud and D. H. Hayes, Eur.

J. Eiochem., 20 (1971) 525. R. G. Crystal, A. W. Nienhuis, N. A. Elson and W. F. Anderson,J. Biol. Chem., 247 (1972)

5357.

C. Ilgen, L. L. Kirk and J. Carbon,J. Biol Chem.. 251 (1976) 922. N. Schiimmelfeder and W.Heyer,Naturwssenschaften, 4 1 (1954) 164. A. B. Foster and M. Stacey,J. Appl. Chem., 3 (1953) 19. V. Stephanovich,J. Chromatog., 31 (1967) 466. S. M. Partridge,Nature (London), 164 (1949) 443. L. Hough, Nature (London), 165 (1950) 400. M. Vietti-Michelina, Z. Anal. Chem., 157 (1959) 266. H. Berbalk,Monatsh. Chem., 89 (1958) 536. E. Schumacher and T. Studer, Helv. Chim. Acta, 47 (1964) 957. G. Cavina and L. Tenton, Clin. Chim. Acta. 3 (1958) 160. K. D. Voigt and I. Beckmann,Acia Endocrinol., 15 (1954) 251. W.P. McKinley, Science, 121 (1955) 139. P. Troen, B. Nilsson, N. Wiquist and E. Diczfalusy, Acta Endocrinol., 38 (1961) 361. 150a 0. J. Schmid, A. Oriol-Bosch and K. D. Vioght, Clin Chim. Acta, 4 (1959) 599. 151 G. L. Cohen and P. K. Bondy, J. Eiol. Chem., 234 (1959) 31. 152 C. J. Migeon and J. E. Plager, Recent Progr.Horm. Res., 9 (1954) 235. 153 S. K. Babich and B. T. Mendibaev,Fom. Zh., 23 (1968) 14. 154 G. BOIS, I. Popa and A. Voicu, Farmacia (Bucarest), 19 (1971) 207. 155 J. L. Frahn,Aust. J. Chem., 22 (1969) 1655. 156 M. Sinibaldiand B. Rinalduzzi, Anal. Lett., 4 (1971) 125. 157 S. N. Tewari,Mikrochim. Acto, 7 (1968) 390.. 158 S. Ochab, Diss.Pharm Pharmacol,, 24 (1972) 205. 159 S. Shaltiel, R. Mizrahi, Y. Stupp and M. Sela, Eur. J. Biochem., 14 (1970) 509. 160 T. P. Hetthger and L.C. Craig, Biochemistry, 9 (1970) 1224. 161 T. Yajima, M. A. Grigg and E. Katz, Arch. Biochem. Biophys., 151 (1972) 565. 162 U. Melcher, Biochim. Biophys. Acto, 246 (1971) 216. 163 J. L. Peel, Eiochem. J., 69 (1958) 403. 164 K. Yagi, Methods Biochem. Anal., 10 (1962) 319. 165 1. Takeda, Agr. Biol. Chem., 33 (1969) 122. 166 S. DuprB, M. T. Graziani, M. A. Rosei, A. Fabi and E. Del Grosso, Eur. J. Biochem.. 16 (1970)

139 140 141 141a 142 143 144 145 146 147 148 149 150

571. 167 168 169 170

T. Urushibara, H. S. Forrest, D. S. Hoare and R. N. Patel, Eiochem. J., 125 (1971) 141. J. M. Gawthorne, Ausr. J. Exp. Biol. Med. Sci., 47 (1969) 311. H. Kondo, H. Anada, K. Ohsawa and M. Ishimoto,J. Biochem. (Tokyo), 69 (1971) 621. S. J. Friedberg, A. Heifetz and R. C. Greene, J. Biol. Chem., 246 (1971) 5822.

REFERENCES

171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 21 1 212 21 3 214 215 216 217 218 219 220

107

D. Lipkin, B. E. Phillips and J . W. Abrell, J. Org. C/7em., 34 (1969) 1539. C. P. Milstein, Biochem. J., 110 (1968) 127. J. GraspariE, J. Chromatogr., 54 (1971) 436. A. Cee and J. GaspariE, Collect. Czech. Chem. Commun., 33 (1968) 1091. T. Schmidt, Acta Histochem., 39 (1971) 69. E. P. Shvartsval'd, Vopr. Med. Khim., 17 (1971) 327. C. M. Wilson, Anal. Biochem., 53 (1973) 538. H. Yoneda and T. Miura, Bull. Chem. SOC.Jap., 43 (1970) 574. F. PetrS and F. Jursik, Collect. Czech. Chem. Cornmiin., 34 (1969) 3153. F. Jursik and B. Hijek, Collect. Czech. Chem. Commun., 37 (1972) 1801. E. Could and J . J. Karolus, Anal. Biochem., 67 (1975) 515. J. L. Frahn,J. Chromatogr., 54 (1971) 103. D. Gross,J. Chromatogr., 10 (1963) 221. S. K. Shukla and I. P. Adloff,J. Chromatoy., 8 (1962) 501. M. Lederer,Anal. Chim. Acta, 6 (1952) 521. M. Lederer, Anal. Chim. Acta, 17 (1957) 606. G. Grassini and M. Lederer, J. Chromatogr., 28 (1 959) 1560. H. H. Strain, J. F. Binder, G. H. Evans, H. D. Frame and J. J. Hines, Anal. Chem., 33 (1961) 527. G. de Vries, G. P. Schiitz and E. van Dalen, J. Chromatogr., 13 (1964) 275. F. H. Pollard, G. Nickless and K. W. C. Burton,J. Chromatogr., 8 (1962) 507. W. Friedli and E. Schurnacher, Helv. Chim. Acta, 44 (1961) 1829. J. P. Adloff and R. Bentrand, J. Electroanal. Chem., 5 (1963) 461. B. Kozjak, Z. Marinic, Z. Konrad, L. Musani-MarazoviE and Z. PuEar, J. C/rromatogr., 132 (1977) 323. M.Casillo, M. Lederer and L. Ossicini, J. Chromatogr., 135 (1977) 256. A. P. Ryle, F. Snager, L. F. Smith and R. Kitai, Biochem. J., 60 (1955) 541. S. Blackburn, Methods Biochem. Anal., 13 (1965) 1. T. Wieland and G. Pfleiderer,Angew. Chem., 67 (1955) 257. G. Werner and 0. Westphal, Angew. Chem., 67 (1955) 251. Z. Pruslk and B. Keil, Collect. Czech. Chem. Commun., 25 (1960) 2049. W. Ostrowski and A. Krawczyk, Acta Biochim. Polon., 13 (1966) 121. K. Dose and A. Caputo, Biochem. Z., 328 (1956) 376. G. N. Atfield and C. J. 0. R. Morris, Biochem. J., 81 (1961) 606. Y. Kakimoto, T. Nakajima, A. Kumon, Y. Matsuoka, N. Imaoka and 1. Sano,J. Biol. Chem., 244 ( 1969) 6003. W. M. Livingston and P. Leder, Biochemistry, 8 (1969) 435. W. Stoffel and G. Assmann, Hoppe-Seyler's Z. Physiol. Chem., 351 (1970) 1041. G. L. Moore and R. S. Antonoff, Anal. Biochem.. 39 (1970) 260. B. L. Ong and J. F. Jackson, Anal. Biochem., 42 (1971) 289. E. Moczar, B. Robert and L. Robert,Anal. Biochem., 45 (1972) 422. D. Gibson and G. H. Dixon, Nature (London), 222 (1969) 753. D. Lalka and N. Back, Biochim Biophys. Acta, 236 (1971) 47. C. J. Bailay and J. A. M. Ramshaw, Biochem. J.. 135 (1973) 889. R. V. Wenn, Biochem. J.. 145 (1975) 281. P. Dimroth, Eur. J. Biochem., 64 ( I 976) 269. R. Radcliffe and Y. Nemerson, J. Biol, Chem., 251 (1976) 4797. F. C. Stevens, M. Walsh, H. C. Ho, T. S. Teo and J. H. Wang, J. Biol. Cliem., 251 (1976) 4495. P. Hohmann, R. A. Tobey and L. R. Gurley, J. Biol. Chem.. 251 (1976) 3685. J. J. Jackson, A. Hagopian, D. J. Carlo, G. A . Limjuco and E. H. Eylar, Biochim. Biophys. Acta, 427 (1976) 251. J. Clarke and L. M. Shannon, Biophys. Acta, 427 (1976) 428. T. Muramatsu, M. Ogata and N. Koide,Biochim. Biophys. Acta, 444 (1976) 53. M. J. Silver, I. Rodalewicz, Y. Duglas and D. Park, Anal, Biochem., 36 (1970) 525.

108

ZONE ELECTROPHORESIS

221 A. Orengo, J. Greenspan and J. D. Love, A n d . Biochem., 63 (1975)559. 222 K. Prabhakararao and M. E. Jones, Anal. Biochem., 69 (1975)451. 223 Z. W. Hall and I:R. Lehman,J. Biol. Chem., 244 (1969)43. 224 M. Kosakai and Z. Yosizawa, Anal. Biochem., 69 (1975)415. 225 A. S. Rae and K. P. Strickland, Biochim. Biophys. Actu, 433 (1976)564. 226 J. C. Schultz and K. Takayama, Biochim. Biophys. Acta, 433 (1976)564. 227 Y. C.Lee and J. R. Scocca, J. Biol. Chem., 247 (1972)5753. 228 A. Crevat, M. L. Jacquet-Francillon and P. Bernard, Ann. Phys. Biol. Med., 2 (1968)73. 229 G. Holmgren, J. 0. Jepson and G. Samuelsen, Scund. J. Din. Lab. Invest., 26 (1 970) 31 3. 230 E.Blasius and W . Preetz, Chromotogr. Rev., 6 (1964)191. 231 M. Sajgd, in T. Devhyi and J. Gexgely (Editors), Amino Acids, Peptides und Proteins, Akaddmiai Kiadd, Budapest, 1974,pp. 96-1 16. 232 K. E.Van Holde, in S. B. Needleman (Editor),Protein Sequence Determination, Springer, Berlin, Heidelberg, New York, 1975,pp. 5-29. 233 J. M. Denuck, in D. H. Moore (Editor),PhysicolTechniques in Biological Reseurch, Vol. 2A, Academic Press, New York, 1968,pp. 151-216. 234 K. Mueller and H. Bremer. J. Mol. Biol., 38 (1968)329. 235 G. Haegeman, W.Min Jou and W. Fiers, J. Mol. Biol., 57 (1971) 597. 236 L. Clarke and J. Carbon, J. Biol. Chem., 249 (1974)6874. 237 E. L. Durrum,J. Colloid Sci., 6 (1951)274. 238 H.H. Strain, Anal. Chem., 24 (1952)356. 239 0.Mikeg and V. Holeygovskf, Collect. Czech. Chem. Commun., 23 (1958)524. 240 J. R. Brown and B. S . Hartley, Biochem. J., 101 (1966)214. 241 B. S. Hartley, Biochem. J., 119 (1970)805. 242 W. E.Gall and G. M. Edelman, Biochemistry, 9 (1970)3188. 243 C. Milstein and B. Frangione, Biochem. J., 121 (1971)21 7. 244 E. Luchter-Wasyl and W. Ostrowski, Biochim. Biophys. Acta, 365 (1974)349. 245 G. R. Stewart and K. J . Stevenson, Biochem. J., 135 (1973)427. 246 C.J. Bailey and D. Boutler, Biochem. J.. 113 (1969)669. 247 B. S. Hartley, J. R. Brown, D. L. Kauffman and L. B. Smillie, Nature (London), 207 (1965) 1157. 248 L. B. Smillie and B. S . Hartley, Biochem. J., 105 (1967) 1125. 249 J. R. Brown, D. L. Kauffman and B. S. Hartley, Biochem. J.,, 103 (1967)497. 250 J. Tang and B. S . Hartley, Biochem. J , 118 (1970)611. 25 1 M. A. Naughton and H. Hagopian, And. Biochem., 3 (1962)276. 252 V. M. Ingram, Biochim. Biophys. Actu, 28 (1958)539. 253 A. Krawczyk, Bull. Acad. Polon. Sci., CI. II, 10 (1962)87. 253a C. W. Easley, B. J. M. Zegers and M. De Vijlder, Biochim. Biophys. Acto, 175 (1969)211. 254 K. C. S.Chen and R. M.Krause,Anal. Biochem., 69 (1975) 180. 255 C.Baglioni, Biochim. Biophys. Acta, 48 (1961)392. 256 B. C. Carlton and C. Yanofsky, J. Bid. Chem., 240 (1965)690. 257 M. Telegdi, F. Fabian, S. M. El-Sewedy and B. F. Straub, Biochim. Biophys. Actu, 429 (1976) 860. 258 M. G. Weigert and A. Garen, J. Mol. Biol., 12 (1965)448. 259 C. Branhnt, A. Krol, J. Sriwidada, J. P. Ebet, P. Sloof and R. A. Garrett, Eur. J. Biochem., 70 (1976)457. 260 N. A. Elson, R. G. Crystal and W.F. Anderson, Anof. Biochem., 54 (1973) 161. 261 H. J. McDonald and M. C. Urbin, Fed. Proc.,Fed. Amer. SOC.Exp. Biol., 12 (1953)243. 262 R.Svensson and I. Brattsten, Ark. Kemi, 1 (1949)401. 263 W.Grassmann and K. Hannig.Angew. Chem, 62 (1950)170. 264 H. J. McDonald, L. P. Ribeiro and L. J. Banaszak, Anal. Chem., 31 (1959)825. 265 A. Kolin,Science. 117 (1953)134. 266 J. Kowalczyk and T. Pompowski, Wiad. Chem., 15 (1961) 107. 267 J. Kowalczyk, Zesz. Nuuk. Politech. cdansk, Chem., 15 (1966) 1.

REFERENCES

268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 313a

109

A. Katsuno, Bull. YamaguchiMed. Sch., 8 (1961) 1. S. Nakamura, Cross Electrophoresis - Its Principle and Applications. Igaku Shoin, Tokyo, and Elsevier, Amsterdam, 1966, pp. 28-38. S. Nakamura, K. Takeo, K. Tanaka and T. Ueta, Hoppe-Seyler's Z. Physiol. Chem., 31 8 (1960) 115. W. Grassmann and L. Hubner, Naturwissenschaften, 40 (1953) 272. S. Nakamura, K. Takeo and 1. Sasaki, Hoppe-Seyler's Z. Physiol. Chem., 328 (1962) 139. S. Nakamura, K. Takeo and I. Sasaki, Hoppe-Seyler's Z . Physiol. Chem., 334 (1963) 95. M. H. Bickel and D. Bovet, J. Chromatogr., 8 (1962) 466. J. Kohn, Biochem. J . , 65 (1957) 9. T. I. Pfistoupil, Nature (London), 212 (1966) 121. T. I. Pfistoupil, Biochim. Biophys. Acta, 117 (1966) 475. T. I. Pfistoupil and V. Fricova,J. Chromatogr., 26 (1967) 331. T. I. Pristoupil, V. Fricova, M. Kramlova, M. Travnicek and J. Sterbikova, J. Chromatogr., 36 (1968) 91. T. 1. Pfistoupil, Chem. Listy, 63 (1969) 577. T. 1. Pfistoupil, Chromatogr. Rev., 12 (1970) 109. J. Kohn, in I. Smith, (Editor), Chromatographic and Electrophoretic Techniques, Interseience, New York, 1968, pp. 84-146. A. S. Kleiner, W. R. Beisel and W. K. Atkins, Amer. J. Clin. PathoL, 50 (1968) 137. E. Wessler, Anal. Biochem., 26 (1968) 439. K. Arai and H. W. Wallace, Anal. Biochem., 31 (1969) 71. A. Gardais, J. Picard and C. Tarasse, J. Chromatogr.. 42 (1969) 396. N. Seno, K. Anno, K. Kondo, S. Nagase and S . Saito, Anal. Biochem., 37 (1970) 197. R. Hata and Y. Nagai, Anal. Biochem., 45 (1972) 462. C. S. Sat0 and F. Gyorkey, Anal, Biochem., 61 (1974) 305. A. Ogamo, T. Ogashiwa and K. Nagasawa, Biochim. Biophys. Acta, 45 1 (1976) 426. J. Partridge, L. Shannon and D.Gumpf, Biochim. Biophys Acta, 451 (1976) 470. R. C. Charman and R. A. Landowne, Anal. Chem., 19 (1967) 177. P. C. Gieson and M. Thiessen, Amer. J. Clin.Pathol., 51 (1969) 523. T. Nikkari and 1. Viikari, Clin. Chim. Acta. 24 (1969) 473. C. A. J. Goldberg, Clin. Chem., 5 (1959) 446. J. Kohn,J. Clin. Pathol., 22 (1969) 109. J. F. Godeau, Y. G. Benzard, J. Cacheleux, C. P. Brizard, A. Gibaud and J. Rosa, J. Biol. Chem., 251 (1976) 4346. P. Meera Khan, Arch. Biochem. Biophys., 145 (1961) 470. T. Matshushima, S. Kawabe, M. Shibuya and T. Sugimura, Biochem. Biophys. Res. Commun., 30 (1968) 565. K. Adachi, L. J. Chen and H. J. Sallach, Biochem. Biophys. Res. Commun., 30 (1968) 343. L. FlohC, Hoppe-Seyler's Z. Physiol. Chem., 250 (1969) 856. S. N. Millner and B. Massie, Anal. Biochem., 32 (1969) 154. L. A. Robinson and T. T. White, Biochim. Biophys. Acta, 222 (1970) 390. E. Bergamini and B. L. Horecker, Anal. Biochem., 44 (1971) 182. W.A. Susor and W. J. Rutter, Anal. Biochem., 43 (1971) 147. A. Gertler, Y. Tencer and G. Tinman, Anal. Biochem., 54 (1973) 270. J. K. Herd and J. Tschida,AnaL Biochem., 68 (1975) 218. J. J. Strandholm. R. D. Dyson and J. M. Cardenas, Arch. Biochem. Biophys., 173 (1976) 125. G. Matteo, G. Di Priseo and G. Romeo, Biochim. Biophys. Acta, 429 (1976) 694. B. Konze-Thomas and W. Ruger, Eur. J. Biochem., 63 (1976) 321. F. Machicao and J. Sonnenbichler, Biochim. Biophys. Acta, 236 (1971) 360. M. H. Siegehman, 1. R. Rasched, K. J. Kunert, P. Kroneck and P. Boger, Eur. J. Biochem., 64 (1976) 131. T. E. Huntley, J. K. Neitzel and M. K. Elson, Biochim. Biophyr Acta, 490 (1977) 112. T. I . Pristoupil, Clzromatogr Rev., 12 (1970) 109.

110

ZONE ELECTROPHORESIS

A. M. Jermyn and R. Thomas, Nature (London), 172 (1953) 728. E. R. Huehns and A. 0. Jakubovic, Nature (London), 186 (1960) 729. A. Yaron and H. A. Sober, Anal. Biochem., 12 (1965) 173. F. Sanger, G. G. Brownlee and B. G . Barrell,J. Mol. Biol., 13 (1965) 373. G . G. Brownlee and F. Sanger,J. Mol. Biol., 23 (1967) 337. G. G. Brownlee, F. Sanger and B. G . Barrell, Nature (London), 215 (1967) 5102. G . G. Brownlee, Determination of Sequences in RNA, North-Holland, Amsterdam, 1972. K. Murray, Biochem. J., 118 (1970) 831. G. G. Brownlee, F. Sanger and B. G. Barrell, J. Mol. Biol., 34 (1968) 379, J. Crossley, J. Mansbridge and R. Williamsin, Biochem. J., 131 (1974) 299. S. K. Dube, K. A. Marcker, B. F. C. Clark and S. Cory, Eur. J. Biochem., 8 (1969) 244. E. M. Southern and A. R. Mitchel1,Biochem. J., 123 (1971) 613. M. W. Konrad, Anal. Biochem., 54 (1973) 213. A. G. Porter, J. Hindley and M. A. Milleter, Eur. J. Biochem., 4 1 (1974) 4 1 3. R. Contreras and W. Fiers, Anal. Biochem., 67 (1975) 319. H. Van Orrnondt and S. Hattrnan, Anal. Biochem., 74 (1976) 207. A. M. Lankowitz and D. J. L. Luck, J. Biol. Chem., 251 (1976) 3081. J. P. Briand, G . Jonard. H. Guilley, K. Richards and L. Hirth, Eur. J. Biochem., 72 (1977) 453. J. E. Edstrorn, Biochim. Biophys. Acta, 12 (1953) 361. J. E. Edstrorn, Biochim. Biophys. Acta, 2 2 (1956) 378. J. E. Edstrom and J. Kawiak,J. Biophys. Biochem. Cytol., 9 (1971) 619. J. E. Edstrorn and W. Beerrnann,J. CellBiol., 14 (1962) 371. H. Sierakowska, J. E. Edstrom and D. Shugar, Acta Biochim. Polon., 11 (1964) 497. J. J. Marchalonis and G. J. V. Nossal, Proc. Nut. Acad. Sci. US.,61 (1968) 860. B. J. Tufts, Anal. Chim. Acta, 25 (1961) 322. H. Van Dijk, F. P. Ijsseling and H. Loman, Anal. Chim. Acta, 27 (1962) 563. G . T. Matioli and H. B. Niewisch, Science, 150 (1965) 1824. H. H y d h , K. Bjurstam and B. McEveri, Anal. Biochem., 17 (1966) 1. R. Riichel, S. Mesecke, D. I. Wolfrum and V. Neuhoff, Hoppe-Seyler's Z. Physiol. Chem., 354 (1973) 1351. 34 3 R. Riichel, S . Mesecke, D. I. Wolfrum and V. Neuhoff, Hoppe-Seylerb Z. Physiol. Chem., 355 (1974) 997. 344 V. Neuhoff,Arzneim.-Forsch., 18 (1968) 35. 345 V. Neuhoff, W. B. Schill and H. Sternback, Biochem. J., 117 (1970) 623. 346 H. A. Poh1,J. Appl. Phys., 29 (1958) 1182. 347 H. A. Pohl and C. E. Plymale, J. Electrochem. SOC.,107 (1960) 390. 348 J. S . Crane and H. A. Poh1,J. Theor. BioL, 37 (1972) 15. 349 A. D. D'Angeac and J. P. Sallei, Bull. SOC.Chim. Biol., 49 (1967) 124. 350 M. H. Paul and E. L. Durrurn, J. Amer. Chem. SOC.,74 (1952) 4721. 35 1 F. Oehme and I. Rauschenbach, Chem. Tech., 8 (1956) 21. 352 A. Krawczyk and E. Dziernbor,Bull. Acad. Polon. Sci., Cl.II, 19 (1971) 81. 353 W. Preetz and H. L. Pfeifer,J. Chromatogr., 4 1 (1969) 500. 354 H. T. Gordon, W. Thornburg and L. N. Werurn, J. Chromatogr., 9 (1962) 44. 355 I. Brattsten, R. L. M. Synge and W. B. Watt, Biochem. J., 9 7 (1965) 678. 356 M. Bagdasarian, N. A. Matheson, R. L. M. Synge and M. A . Youngson, Biochem. J., 91 (1964) 91. 357 W. Thorum and E. Mehl, Biochim. Biophys. Acta, 160 (1968) 132. 358 H. Demus and E. Mehl, Biochim. Biophys. Acta, 203 (1970) 291. 359 A. Pusztai and W. B. Watt,Biochim. Biophys. Acta, 251 (1971) 158. 360 A. Pusztai and W. B. Watt, Anal. Biochem., 54 (1973) 58. 361 D. Z. Staynov, J. C. Pinder and W. B. Gratzer, Nature (London), 235 (1972) 108. 362 J. C. Pinder, D. 2. Staynov and W. B. Gratzer, Biochemistry, 13 (1974) 5373. 363 J. C. Pinder, D. Z. Staynov and W. B. Gratzer, Biochemistry, 13 (1974) 5367. 364 H. J. Gould and P. H. Harnlyn, FEBS Lett., 30 (1973) 301.

314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 34 1 342

REFERENCES

111

T. Maniatis and M. Ptashne,Proc. Nat. Acad. Sci. US.,70 (1973) 1531. G. G . Brownlee, E. M.Cartwright, W. J. Cowan, J. M. Jarvis and C. Milstein,Nature (London), 244 (1973) 236. 367 B. Groner, N. Hynes and S. Phillips, Biochemistry. 13 (1974) 5378. 368 Y. Kataoka, I. Yamase, T. Fukumura and T. Kiato,J. Chromatogr., 132 (1977) 175. 369 T. Pompowski and J. Kowalczyk,PToc. Anal. G e m . Conf., Budapest, 1966, p. 97. 370 A. Ruis and F. Grass,J. Chromatogr., 45 (1969) 291. 371 S. Katz and A. Lewis, Anal. Biochem., 17 (1966) 300. 372 W. J. Van Ooij, J. Chromatogr,, 42(1969) 432. 373 A. R. Trim and P. E. Dickerson, Anal. Biochem., 57 (1974) 93. 374 D. L. Bates, R. N.Perham and J. R. Coggins, Anal. Biochem., 68 (1975) 175. 375 H. C.Gassen, J. Chromatogr. 39 (1969) 147. 376 W. J. Ritschard,J. Chromatogr., 16 (1964) 327. 377 A. Hamann, K. D. Spindler and K. Wortelkamp, Scan, 7 (1976) 9. 378 U. Hezel, Angew. Chem;, 8-5-11973) 334. 379 J. Chudzik and A. Klein,J. Chromatogr., 36 (1968) 262.. 380 B. C. Baguley and M. Staehelin, Biochemistry, 7 (1968) 45. 381 J. P. Monjardino, Anal. Biochem.. 28 (1969) 313. 382 E. Sawicki, M. Guyer and C. R. Enge1,J. Chromatogr., 30 (1967) 522. 383 H. Ahrens and W. Korytnyk, Anal. Biochem., 30 (1969) 413. 384 C. E. Colombini and E. E. McCoy, Anal. Biochem., 34 (1970) 451. 385 H. Muir and S. Jacobs, Biochem. J., 103 (1967) 367. 386 W.J. Criddle, G. J. Moody and J. D. R. Thomas, J. Chromatogr., 16 (1964) 350. 387 G. Pastuska and H. Trinks, Chem.-Ztg.,86 (1962) 135. 388 W. J. Van Ooij and J. W. P. Houtman, Z. Anal. Chem.. 236 (1968) 407. 389 A. Moghissi, Anal. Chim. Acta, 30 (1964) 91. 390 A. R. Cook and R. L. Bieleski, Anal. Biochem., 28 (1969) 428. 391 R. Chen, Hoppe-Seyler’s Z. Physiol. Chem., 358 (1977) 1415. 392 K. Keck and U. Hagen, Biochim. Biophys. Acra, 87 (1964) 685. 393 R. L. Bieleski, Anal. Biochem., 12 (1965) 230. 394 L. N. Ferrier, P. L. Davies and G. H. Dixon, Biochim. Biophys. Acta, 479 (1977) 460. 395 M. Szekely and F. Sanger, J. Mol. BioL, 43 (1969) 607. 396 E. M. Southern and A. R. Mitchell, Biochem. J., 123 11971) 613. 397 E. Dziembor, A. Krawczyk and W. Ostrowski, Bull. Acad. Pol. Sci., CI. II, 17 (1969) 213. 398 D. T. Dubin and R. H. Taylor, Nucleic Acid Res., 2 (1975) 1653. 399 V. G. Gorbulev, T . V. Kutateladze, J. Barciszewski and V. D. A. Axel’rod, Biochem. J., 163 (1977) 409. 400 L. A. Hanson, B. G. Johansson and L. Rymo, C7in. Chim. Acta, 14 (1966) 391. 401 D. C. Locke, J. Chromarogr., 35 (1968) 301. 401a A. Klein, J. Chudzik and M. Sarnecka-Keller, J. Chromatogr., 53 (1970) 329. 401b D. Jaworek, Int. Symp. VI. Chromatogr. Electrophor., Bruxelles, 1971, pp. 118-125. 402 G. Alberti, G. Grassini and R. Trucco, J. Electroanal. Chem., 3 (1962) 283. 403 E. P. Honig and J . A. A. Ketelaar, Trans. Faraday SOC.,62 (1966) 190. 404 H. J. Amikar, C. R. Acad. Sci., 244 (1957) 2241. 405 J. A. A. Ketelaar and J. C. T. Kwak, J. Phjzs. Chem., 71 (1967) 1149. 406 G. Alberti, A. Conte and S. Allulli, J. Chromatogr., 18 (1965) 564. 407 B. R. Sundheim, in B. R. Sundheim (Editor),Molten Salts, McGraw-Hill, New York, 1964, p. 165. 408 J. A. A. Ketelaar and E. P. Honig, J. Phys. Chem.. 68 (1964) 1596. 409 F. Lantelme and M. Chemla, Electrochim. Acta, 10 (1965) 663. 410 G. Alberti, S. Allulli and L. Polazzeschi,J. Chromatogr., 44 (1969) 141. 411 S. Allulli, L. Polazzeschi and L. Corsi, J. Chromatogr., 45 (1969) 464. 412 G. Alberti and S . Allulli, Chromatogr. Rev., 10 (1968) 99. 413 A. Lund6n,Z. Naturforsch. A , 14 (1959) 801. 414 S. Jordan and A. Klemm, Z. Narurforsch. A , 21 (1966) 1584.

365 366

This Page Intentionally Left Blank

Chapter 6

Gel-type techniques Z . HRKAL

CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory of gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Starch gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Working procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrophoretic apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acrylamide gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Disc electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stock solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Working procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrophoretic apparatus and destainer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SDS electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SDS-gelsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrophoresis in gel slabs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Instrument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Working procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gradient gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two-dimensional polyacrylamide gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . Agarose gel electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Workingprocedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agarose-acrylamide composite gels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Working procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel preparation (2%acrylamide. 0.5% agarose) . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

113 114 116 116 116 116 117 117 118 119 121 122 123 124 125 125 125 126 126 127 128 128 129 129 129 130 130 130 130

INTRODUCTION Electrophoresis in its present form represents the movements of charged particles such as macromolecular ions. cells. subcellular organels and small organic molecules under the influence of an electric field . The method was discovered in the last century and since then numerous techniques have been developed for the separation of charged particles in mixtures based on the principles of electrophoresis . In principle there are two types of electrophoretic methods: moving boundary electrophoresis. in which the separation occurs in a free solution. and zone electrophoresis. i.e., electrophoresis on solid supports

114

CEL-TYPE TECHNIQUES

that serve as the anticonvection media limiting free diffusion and occasionally exerting a sieving action. The first practical apparatus for free moving boundary electrophoresis was constructed by Tiselius in 1937 [I], and zone electrophoretic methods were developed in the 1950s when filter-paper was used as the support and anticonvection medium, later being replaced with cellulose acetate sheets which give a higher resolution and reproducibility and which are also transparent. In 1955, Smithies [2] introduced starch gel for the electrophoretic separation of protein mixtures. Electrophoresis in starch gels gave a much higher resolution than electrophoresis on paper or cellulose acetate, small molecules migrating faster than large molecules with the same charge to mass ratio. These findings indicated that in electrophoresis in concentrated gels, in addition to the separation according to charge, separation according to particle size also occurred. This property is called the “filtration” or “sieving” effect. Starch gel electrophoresis achieved great popularity in the 1960s but since then it has largely been replaced by polyacrylamide gel, although it still is in use for special purposes. Raymond and Weintraub [3] first used acrylamide gel layers for electrophoresis and in 1964 Ornstein 141 and Davis [5] introduced a method of electrophoresis in gel rods. They used a discontinuous buffer system, which promoted sharpening of the protein bands at the start as a consequence of the tachophoretic principle, which enables the analysis of even very dilute solutions t o be carried out. Electrophoresis on acrylamide gel is a universal method; by changing the gel concentration, a desired range of molecular weights can be chosen in which a particular separation is to occur. In 1967, Shapiro et al. [6] introduced acrylamide gel electrophoresis in the presence of an ionic detergent, sodium dodecyl sulphate (SDS electrophoresis). Their method allows the determination of the subunit molecular weight of reduced and carboxymethylated proteins by comparison of their mobilities with those of standard proteins of known molecular weight. For the separation of particles with molecular weights above lo6 (RNA, DNA, subcellular particles), electrophoresis in agarose gels [7] or composite agarose-acrylamide gels [8] is employed. Acrylamide gel electrophoresis was further modified by the introduction o f “gradient gels” [9], i.e., of gels with a linear or nonlinear acrylamide concentration (usually 430%). In gradient gel electrophoresis the electric field promotes only the driving force for the macro-ions while their separation occurs on the basis of their different molecular sizes,

THEORY OF GEL ELECTROPHORESIS The main feature of gel electrophoresis is the dependence of the mobility of a macroion on the gel concentration. The basic equation describing the electrophoretic behaviour of the charged particle is Q * E = f*(U. I , Z )macm where Q is the effective electric charge of the macro-ion, E the intensity of electric field, (Ui,z)m, the electrophoretic mobility and f the friction coefficient. f follows from Einstein’s equation:

THEORY OF GEL ELECTROPHORESIS

115

where k B is Boltzmann’s constant, T the absolute temperature and Di,z the diffusion coefficient. Then

Considering the electrophoresis of a single macro-ion in two gels differing in gel density under otherwise identical conditions, we have

Q * E = constant

(6.4)

where the subscripts 1 and 2 refer t o the two gel densities. Assuming that the contributions of the solvent t o f are identical in both gels, then the only reason for the different mobility of the macro-ion in the gels of different density is the contribution of the gel structure t o the friction coefficient. Let us consider the movement of the molecule through the gel of a given pore distribution (presumably Gaussian) under the action o f an electric field. The molecule is restricted in its movement, as it cannot pass through some of the pores. Owing t o thermal agitation, however, it continues t o migrate until it finds a pore that it can penetrate. If this did not occur, a smear from the origin instead of a sharp zone would be observed in gel electrophoresis. Two factors are then responsible for the migration rate of the particle in gel electrophoresis: its charge and the proportion ofpores that it can penetrate. The mobility of the macro-ion would be lower in the gel o f higher concentration as it contains fewer pores that it can penetrate in comparison with the gel of lower density. The total distance that the particle would migrate in each o f the two gels of different concentration, however, would be about the same for an equal time. The empirical relationshp between the electrophoretic mobility, U,., and the total gel concentration,yp, was deduced by Ferguson [lo]: log (Ui)macm = log Cui)Lam - K R Y P

(6.6)

where (Ui)kac, is electrophoretic mobility at zero gel concentration and K R the retardation coefficient. This equation demonstrates the linear relationship between the logarithm of electrophoretic mobility and the gel concentration. Subsequently a theoretical foundation was given by Rodbard and Chrambach [ 1 11, who derived the equation mathematically on the basis of gel structure theory. The validity of Ferguson’s equation was confirmed for acrylamide gels over a range of conditions [ 121 and applied t o various kinds of charged molecules such as proteins, nucleic acids and dyes. Hedrick and Smith [ 131 made use of Ferguson’s equation for the determination of the molecular weights of proteins*. By determining the relative mobilities o f the particular macro-ion in a series of acrylamide gel rods at several gel concentrations, the retardation coefficient, K R , is obtained, which is related t o the molecular weight through a linear relationship.

See also Chapter 4 for details.

116

GEL-TYPE TECHNIQUES

STARCH GEL ELECTROPHORESIS Electrophoresis on starch gel, introduced by Smithies [ 2 ] ,was the first electrophoretic method in which the sieving effect of the electrophoretic medium played a significant role in'the resolution of complex protein mixtures. Nowadays the method has largely been replaced by the simpler and more reproducible acrylamide gel electrophoresis, although for special purposes it is still of importance (abnormal hemoglobin classification, studies of isoenzymes, haptoglobin typing, etc.). Chemicals and solutions Starch hydrolysed for gel electrophoresis (Sigma, St. Louis, Mo., U.S.A.; Serva Feinbiochemica, Heidelberg, G .F.R.).

Stock buffer, pH 8.8 Tris(hy d ro xy met hy1)aminomet hane 109 g Ethylenediaminetetraaceticacid, disodium salt 5.87 g boric acid 2.3 g 11 HzO to The stock solution t o be diluted 1 :7 for the electrode trays and 1 :20 for the gel preparation. Staining solution Amido Black 10B methanol acetic acid (glacial)

10g 450 ml 450ml 100 ml

Des tainer methanol acetic acid (glacial) H2O

250ml 100 ml 750 ml

H2O

Working procedure Gel preparation A 13.3-g amount of hydrolysed starch is added to 100 ml of the gel buffer in a 0.5-1 suction flask and the mixture is heated on an electric heater. As the temperature increases, the viscosity of the gel increases and then suddenly decreases. At this moment the gel mixture is taken off of the heater, degassed with the aid of a water pump and poured in the gel form. The gel is allowed to solidify at 4°C for at least 1 h. After the gel has solidified, slots for sample application, 10 mm wide, are cut with a razor blade. Paper strips (10 x 3 mm) of Whatman No. 3 paper are immersed in the sample solution (1-3% protein solution), the excess is wiped off and the application strip inserted in the slots made in the gel layer. The starch gel is then covered with cellophane foil and positioned

ACRYLAMIDE GEL ELECTROPHORESIS

117

in the central part of the electrophoretic apparatus. The electrophoresis is performed for 4 h at 15 V/cm. After the electrophoretic experiment has been terminated, the gel tray is removed from the apparatus and the gel is cut in half with the aid of a cutting tray, the string of w h c h is inserted in the slot between the two tray frames. Two gel layers are obtained, of which the lower one usually is of better quality and is therefore employed for subsequent staining procedures. The upper layer can be used for parallel staining by other procedures, e.g., for enzymic activity.

Staining and destaining procedure The gel stain is poured on to the gel layer and left to penetrate it for 1 min. Then the gel slab is immersed in the destaining solution with occasional exchange until the gel background is white and opalescent. This takes several days, although on the following day the stronger bands are already visible. Clearing of the gels If it is desirable to have a transparent gel for densitometry, this can be accomplished by heating in glycerol for 20-30 s at 70-80°C. Electrophoretic apparatus The simple electrophoretic apparatus is shown in Fig. 6.1. It consists of the two separate buffer vessels containing platinum electrodes with the gel tray placed in between and connected with the buffer by means of wicks made of seven sheets of Whatman No. 1 fiter-paper. The gel tray consists of a plastic base (9 x 14cm) and two identical frames placed on top of each other, the whole system being held together with two rubber bands. Commercially available instruments suitable for starch gel electrophoresis are produced, for example, by Desaga (Disaphor) and Savant (Model HGE 1312), both possessing provisions for cooling.

ACRYLAMIDE GEL ELECTROPHORESIS The polyacrylamide gel is formed by the polymerization and cross-linking of monomeric substances, acrylamide (CH2=CH-CONH2) and a cross-linking agent, N,N’methylenebisacrylamide (Bis) (CH2=CH-CO-NH-CH2-NH-CO-CH=CH2). As a result of the reaction, a three-dimensional network is formed, the density of which can be determined by the concentration of acrylamide and bisacrylamide in the gelation mixture. While the gel density is determined by the total concentration of acrylamide and bisacrylamide, the degree of cross-linking determines the mechanical properties (elasticity, fragility), which depend on the weight ratio of Bis to acrylamide. The optimal concentration of bisacrylamide (C) in the gelation mixture in relation to the total acrylamide concentration, T(%w/w), is given by the equation of Richards et al. [14]: C = 6.5 -0.3T

Generally, 5-10% gels with 3 - 5 9 >cross-linking are used.

118

GEL-TYPE TECHNIQUES

Fig. 6.1. Starch gel electrophoresis apparatus. (1, 2), Electrode reservoirs; (3) gel tray; (4) paper wicks; (5) platinum electrode.

The polymerization of the monomeric mixture of acrylamide and bisacrylamide is accelerated by the catalytic action of suitable redox systems, ammonium persulphate and riboflavin in combination with N,N,N‘ ,N’-tetramethylethylenediamine(TEMED) or dimethylaminopropionitrile (DMPN) being commonly used. Several techniques of acrylamide gel electrophoresis have been developed, of which the following are the most significant.

Disc electrophoresis Let us consider a simple electrophoretic experiment in which the protein solution is applied on top of the gel rod and the components are forced to migrate under the influence of an electric field. To obtain sharp zones, the solution of the investigated macroionic mixture has to be applied in the form of a narrow band of high protein concentration, otherwise excessive spreading of the bands occurs owing to diffusion. To avoid this unfavourable effect, Ornstein [4] and Davis [S]developed an ingenious technique of electrophoresis on acrylamide gel in a discontinous buffer system in three gels of different acrylamide concentration, extremely sharp zones of disc shape being obtained in the rod-like gels. The principle of their discontinuous electrophoretic system is as follows. A cylindrical column of the “separation gel”, in which the separation of the macro-ions is to occur on the combined basis of different electric charge and molecular size, contains a tris(hydroxymethy1)aminomethane-chloride buffer of pH 8.3. Above this separation gel a small column of large-pore “spacer gel” and large-pore “sample gel” is applied, the latter containing a small amount (5Opg) of the mixture of macro-ions t o be analysed.

ACRYLAMIDE GEL ELECTROPHORESIS

119

These two gels also contain Tris-HC1 buffer. Both ends of the gel rod are immersed in Tris-glycine buffer (pH 8.3). The mobility of the chloride ion is greater than that of any macro-ion of the investigated system, while that of the glycyl anion is the lowest at pH 8.3. When the electric field is applied, zone sharpening and stacking occur as a result of the Kohlrausch principle [ 151. The chloride ion moves to the anode and the migrating macrocomponents segregate in order of decreasing mobilities, followed by the slowest anion in the system, gly-. According to the Kohlrausch principle, each macro-ionic component concentrates in an extremely sharp starting zone. On entering the separation gel of pH 8.9, the mobility of the glycyl anion increases to a value greater than the mobility of the fastest component of the investigated system and migrates immediately behind the chloride ion. This chloride-glycine boundary moves through the gel column, leaving the proteins behmd in Tris-glycine buffer t o separate in a uniform electric field according to their charge and size. The above system was found to be unnecessarily cotnplex and the sample gel is now commonly omitted, being replaced with an increase in sample density by the admixture of sucrose solution. Stock solutions Alkaline gel system (pH 8.9), 7.5% gel (for the separation of macro-ionic components with pZ below 7.5) A HCllN 24 ml 18.2 g Tris TEMED 0.12 ml H2O to 100 ml B HCllN 48 ml 6.0 g Tris TEMED 0.46 ml H20 to 100 ml 29.2 g C Acrylamide N,N'-Methylenebisacrylamide 0.8 g H20 to 100 ml D Acrylamide 10 g N,N'-Methylenebisacrylamide 2.5 g E Riboflavin 0.004g H2O to 100 in1 F Ammonium persulphate 0.15g H20 to 100 ml G Sucrose 40 g H2O to 100 ml Electrode buffer solution (pH 8.3) Tris 6.0 g Glycine 28.8 g H2Oto 11 To the electrode vessels to be diluted 1 : 10 Lower gel composition, 7.5% gel (volume parts): A 1 , C 1, H 2 0 1, F 1

120

GEL-TYPE TECHNIQUES

Upper gel composition : B 1, D 2, E 1, G 4 Polarity: cathode (-)top anode (+)bottom Acidic gel system (pH 4.3), 15% gel (for the separation of macro-ionic components with p l above 7.5) A KOHlN 48 ml Acetic acid (glacial) 17.2 ml TEMED 4.0 ml H2O to 100 ml B KOHlN 48 ml Acetic acid (glacial) 2.9 ml TEMED 0.45 ml H20 t o 100 ml C Acrylamide 29.2 g Bis 0.8 g H2O to 100ml D Acrylamide 10 g Bis 2.5 g H20 to 100ml E Riboflavin 0.004 g H2O to 100ml F Ammonium persulphate 0.28 g H2O to 100 ml Electrode buffer solution (pH 4.5) 0-Alanine 31.2 g 8.0 ml Acetic acid (glacial) H20 to 11 Lower gel composition, 7.5% gel (pH 4.3) (volume parts): A 1, C 2, H 2 0 1 , F 4 Upper gel composition : B 1 ,D 2 , E 1, H20 4 Polarity: anode (+) top cathode (-) bottom Gel systems with urea (for the subunit analysis of proteins under denaturing conditions) Alkaline gel system with urea (pH 8.3), 7.5% gel Stock solutions 10 ml A HCl1 N Tris 18.1 g TEMED 0.25 ml Urea (1 0 M ) to l00ml C Acrylamide 29.2g Bis 0.8 g l00ml Urea (1 0 M) to F Ammonium persulphate 0.3 g 100 ml Urea(10M) to B, D, E - see regular alkaline gel system

ACRYLAMIDE GEL ELECTROPHORESIS

121

Electrode buffer system 3.0g Tris Glycine 14.4g HzO to 11 Lower gel composition, 7.5%gel (pH 8.3): A 2 , C 4 , F 10 Upper gel composition : B 1, D 2 , E 1 , 10M urea 4 Polarity: cathode (-)top anode (+) bottom Acidic gel system with urea (pH 4.3), 7.5%gel Stock solutions A KOHlN 12.5 ml Acetic acid (glacial) 53.2 ml TEMED 1.2 ml Urea (10M) to 100 ml C Acrylamide 29.2 g Bis 0.8 g Urea (10M) to 100 ml F Ammonium persulphate 2.8 g Urea (10M) to 100 ml B, D, E - see regular acidic gel system Electrode buffer system (pH 4.08) Glycine 28.1 g Acetic acid (glacial) 3.06 ml HzO to 11 Lower gel composition, 7.5%gel (pH 4.3): A 2 , C 1 , F 1 Upper gel composition : B 1, D 2 , E 1 , l O M urea 4 Polarity: anode (+)top cathode (- ) bottom Staining solution 1%Amido Black 1OB in 7% acetic acid solution Destaining-storage solution 7% acetic acid

Working procedure The gel tubes (65 x 5 nim) are placed in the polymerizing rack and fixed in the rubber caps pasted on a plastic support, The stock solutions needed for the gel preparation are allowed t o warm to room temperature. The lower gel solution is prepared, degassed if necessary and pipetted into the gel tubes to a height of SO mm. The gel surface is carefully overlayered with water and the gel left to polymerize for 30 min. A shorter polymerization time leads to inhomogeneity of the gel, and in this instance the persulphate concentration should be lowered. The water layer is removed, the gel surface rinsed with some upper gel solution and 0.1 5 ml of the upper gel solution is applied and overlayered with water. The upper gel is photopolymerized under W light or in daylight for 30 min

122

GEL-TYPE TECHNIQUES

(the gel must appear opalescent). The water layer is removed and the sample being investigated is applied in a volume not exceeding 0.2 ml. It is advantageous to increase the density of the analysed macro-ionic solution by the admixture of sucrose solution (C). For the analysis of proteins, the recommended concentration is 1 mg/ml; generally the amount of protein component of interest should be 20pg to obtain a clearly visible zone. The protein solution is then overlayered with the electrode buffer up to the top of gel tube. The tubes are inserted in the upper part of the electrophoretic apparatus, the bottom part is filled with the cool electrode buffer and both parts are fixed. The upper part of the apparatus is filled with the electrode buffer, to which 2 ml of 0.001% Bromphenol Blue for an alkaline gel system or 2 ml of 0.005%Pyronine Y for an acidic gel system are added; the dyes serve as the boundary markers. The power supply is connected and the electrophoretic run performed at 2-4 mA per gel tube. When the dye front has migrated ca. 1 cm from the bottom end of gels, the electrophoresis is stopped, the power supply disconnected and the gel tubes are detached. The gels are released with help of a hypodermic needle, which is inserted between the gel and tube at the bottom side while the tube is rotated. The gels are immediately placed in test-tubes containing the staining solution and left t o stain for 1 h. The stained gels are then rinsed with tap water and destained.

Electrophoretic apparatus and destainer A simple home-made electrophoretic apparatus is shown in Fig. 6.2. It consists of two rectangular plastic boxes that fit on top of each other and form the upper and lower electrode compartments. The upper reservoir (1) has eight holes fitted with rubber grommets for fixing the gel tubes. In its centre a plastic rod with a drilled groove (2) is located, which passes through the bottom into the lower reservoir. It bears the upper (3) and lower (4) platinum electrodes, to which cables passing through the central groove are connected, The lower reservoir (5) contains the electrode buffer. Commercial apparatus for disc electrophoresis is produced, for example, by Desaga, Heidelberg, G.F.R.(Desaga Disc Electrophoresis), Savant Instruments, Hicksville, N.Y., U.S.A. (Acrylamide Gel Cell DEC-12-10) and Bio-Rad Labs., Richmond, Calif., U.S.A. (Model 150A).

Electrophoretic destain ing The electrophoretic destainer [Fig. 6.2 ( 6 ) ] consists of a plastic rack containing gel holders (7) and a pair of platinum electrodes of net shape. The gel holders are inserted in grooves between the electrodes parallel t o their plane. The whole system is placed in a vessel containing destaining solution and connected to a power supply of 24 V and 2 A. The destaining takes approximately 2 h. D$.-ion destaining Although the electrophoretic destaining of acrylamide gels is efficient and fast, it has certain disadvantages. There is a risk of over-destaining, in some instances a distortion of the stained bands may occur and small molecules such as peptides may even be lost owing to the action of the electric field. For these reasons, destaining by diffusion is a more

ACRYLAMIDE GEL ELECTROPHORESIS

123

Fig. 6.2. Disc electrophoresis apparatus. (1) Upper reservoir; (2) central groove; ( 3 , 4 ) platinum electrodes; (5) lower reservoir; (6,7)destainer.

reliable method. The destaining rack described above can also be used for diffusion destaining. It is advantageous to provide efficient circulation of the destaining solution by means of a magnetic stirrer and to incorporate activated charcoal or a mixed-bed resin cartridge in the system to adsorb released dye. SDS electrophoresis

It has been shown [ 161 that the anionic detergent sodium dodecyl sulphate (sodium lauryl sulphate) (SDS) binds to proteins in large amounts in a constant mass to mass ratio, namely 1.4 g of SDS to 1 g of protein. As a result of binding confortnational changes occur, resulting in dissociation of proteins into their constituent subunits, provided that disulphide bonds are ruptured. Owing to the SDS binding, most proteins exhibit a nearly constant charge to mass ratio and therefore identical mobilities in free solution. In gel media, however, their separation occurs according to their molecular sizes. It has been suggested that the protein-SDS complexes behave like prolate ellipsoids or rods, the long axis being proportional to their molecular weight. Shapiro et al. [6] showed that in SDS-containing acrylamide gels the electrophoretic mobility of reduced and carboxymethylated proteins is indirectly proportional to the logarithm of their molecular weight. Employing the calibration graph obtained with proteins of known molecular weight, the molecular weight of the component under investigation can be calculated (see Fig. 6.3 and Chapter 4). SDS electrophoresis is especially suitable for the analysis of multicomponent systems such as membranes and viruses, which are readily soluble in SDS solutions at elevated temperatures in the presence of a reducing agent. Not

124

GEL-TYPE TECHNIQUES Pepsin o Ovalbumin

Bovine Serum albumin Human transterrin ).-Globulin L chain m 7-Globulin H chain

A

A\

O\

T9

10-

X

1

20

60

40

1

00

100

RF

Fig. 6.3. Dependence of log R p (relative mobility) on the molecular weight of proteins in SDS electrophoresis.

only the subunit molecular weight is obtained but also the proportion of the components in the system. The method requires only micromolar amounts of the proteins and the individual components can be further characterized (amino acid composition, amino terminal sequence) after cutting off the zone and elution.

SDS-gel system Stock solutions 29.2 g A Acrylamide Bis 0.8 g H2O to 100 ml Sodium phosphate buffer pH 7 . 2 , l M Sodium dodecyl sulphate 10% (w/v) Ammonium persulphate 0.1 g/ml (to be prepared freshly) Gel preparation: Acrylamide gel solution (A) Sodium phosphate buffer (pH 7 2 ) , 1M SDS, 10% HZO TEMED Ammonium pe rsulphate H2O to

16.7 ml 10 ml 1 ml 70 ml 0.05 ml 1 ml 100 ml

ACRYLAMIDE GEL ELECTROPHORESIS

125

Electrode buffer solution : 0.1 Msodium phosphate buffer (pH 7.2) containing 0.1% SDS Denaturation solution: 0.1 M sodium phosphate buffer (pH 7.2) containing 1% SDS and 1% 2-mercaptoethanol (0.6 ml per 100 ml) Sample preparation: A 1-mg amount of protein to dissolve in 0.1 ml of denaturation solution and to incubate for 3 h at 37°C in a small stoppered test-tube. Then 0.9 ml of water is added. For the analysis, 50-1 00 pl are applied. Staining solution: Coomassie Brilliant Blue R 250 1.25 g Methanol 227 ml 46 ml Acetic acid (glacial) H 2 0 to 500 ml Staining is performed in test-tubes for 2-10 h at room temperature. Destaining solution: Methanol Acetic acid (glacial) HzO to

50 ml 75 ml 1 ml

Electrophoresis in gel slabs Electrophoresis in gel slabs has the advantage that series of samples can be run simultaneously under identical conditions on the same plate, which makes comparison between the components of various samples easy. Efficient cooling of the gel slabs can be assured by circulating cooling liquid through the buffer reservoir. The gel solutions and the buffer systems are as described on p. 118 for disc electrophoresis. Instrument For acrylamide slab gel electrophoresis vertical apparatus is now commonly used as the application of the sample and the use of stacking gel are easier in this arrangement. Workingprocedure The gel is polymerized in a mould created by two glass plates 100-200 mm long and 100-1 50 mm wide. By selecting the appropriate spacers the slab thickness can be varied (usually 1-3 mm). The glass plates are fixed with clamps or sealing tape and the cell so formed is filled with the gel solution by means of a syringe needle. Gradient gels can also be formed by connecting the output of the gradient maker to the gel feeding needle. The top of the separation gel is made flat by overlayering with water. After the polymerization of the separation gel has occurred, the stacking gel is applied. The stacking gel solution is poured on the separation gel and a PTFE well-forming comb is inserted. When

126

GEL-TYPE TECHNIQUES

the photopolymerization of the upper gel is completed, the comb is carefully removed, leaving sample slots in the gel slab. The gel cell is then positioned in the gel slab electrophoretic apparatus and sample solutions are applied. The upper and lower reservoirs are filled with the electrode buffer and the cooling system and power supply are connected. Commercial instruments for slab gel electrophoresis, including destaining accessories, are produced, for example, by Pharmacia Fine Chemicals, Uppsala, Sweden (Gel Electrophoresis Apparatus), Desaga, Heidelberg, G.F.R. (System Havana) and Bio-Rad Labs., Richmond, Calif., U.S.A. (Vertical Slab Gel Cell). Staining and destaining The staining and destaining solutions described in the section on disc electrophoresis are employed. The gel slab is placed in the staining solution for 20 min. As destaining by simple diffusion on a flat disc requires over 48 h, the use of special commercial diffusion or electrophoretic destainers is recommended. Gradient gel electrophoresis Although discontinuous electrophoresis yields initially sharp starting zones, these would blur due to diffusion on migration through the homogeneous gel matrix. This disadvantage of the homogeneous gel system is avoided in a system in which a gradient of gel concentration is employed. With this technique, the gel is prepared almost exclusively in a slab shape with the acrylamide concentration increasing linearly or non-linearly from the cathodic to the anodic side. The samples are applied at the side of low gel concentration and migrate under the action of the electric field in the direction of increasing gel concentration. In this arrangement the electric field promotes only the movement of the particles while the separation occurs on the basis of their different molecular sizes. As the macro-ions migrate through the gel of increasing concentration their interactions with the gel matrix increase and their velocity decreases. After a sufficiently long period (16 h), the migration of particles is so slow that they appear to have stopped. With the continuing separation sharpening of the bands occurs as the trailing molecules of a particular band have less restriction in the direction of migration than the molecules ahead and reach them, with resulting zone sharpening. Gradient gel electrophoresis not only gives an excellent resolution but also provides a means for the determination of protein molecular weights under non-denaturing conditions. This is particularly suitable for studies on enzymes, the activity of which is usually destroyed in the SDS method. As a functional relationship exists between the effective molecular radius of proteins and the limiting pore size [17], the molecular weight can be derived from the migration distance by means of a calibration graph. Experimental procedure A linear concentration gradient can be produced according to Margolis and Kenrick [9] by means of two connected vessels filled with gel mixtures of different acrylamide concentrations. The compositions of the stock solutions for production of 4 or 26% gradient gels are as follows [9]

127

ACRYLAMIDE GEL ELECTROPHORESIS

Reagent

Solution A (26%)

Solution B (4%)

Acrylamide Bis Sucrose Tris-EDTA-borate (pH 8.28) Dimethylaminopropionitrile (DMPN) Ammonium persulphate, 10% solution

76.6 g 4.0 g 12.4g to 302 ml 0.03 ml 7.75 ml

11.8g 0.6 g 3.1 g to 302ml 0.09 ml 7.75 ml

Electrode and gel buffer (pH 8.28) Tris 10.75 g 0.93 g EDTA (Na2 salt) 5.04 g Boric acid H20 to 11

Ammonium persulphate is added t o the gel solutions immediately before pouring the gel mixtures into the vessels. The gradient gel mixture is fed into the slab gel cell, preferably from the bottom by puncturing the sealing bar with a syringe needle. The proteins are applied to the side of low gel concentration (upper side). As no stacking gel is employed, pre-electrophoresis for 15 min at 125 V is necessary. The electrophoretic run takes 16 h , then staining and destaining are performed as described in the section on slab gel electrophoresis (pp. 125- 126). Artifacts often occur in laboratory-made gradient gels owing to the gradient inhomogeneities, with resulting uneven migration fronts. These difficulties can be avoided by using commercial gradient gels (Pharmacia, Uppsala, Sweden), which give a high resolution and reproducibility but are expensive. Two-dimensional polyacrylamide gel electrophoresis Two-dimensional techniques of gel electrophoresis are used to combine the advantages of separations in two different buffer systems or in two different media. The common combination is electrophoresis in a regular gel in one direction and in an SDS-containing gel in the perpendicular direction. The method is of growing importance for the separation of membrane proteins. For the analysis of ribosomal proteins, electrophoresis in urea containing polyacrylamide gel followed by SDS electrophoresis has been successfully employed [ 18]. The combination of electrofocusing and SDS electrophoresis resolved 1100 protein components from Escherichia coli [19]. A combination of charge separation in low-density gels followed by size separation in gradient gels has also been described. Technically, two-dimensional electrophoresis is a combination of electrophoresis in gel rods and gel slabs. The material is electrophoresed in the first direction in a gel tube. The gel is removed from the tube, laid on top of a gel slab and fixed to it by the addition of hot 1% agarose solution. A 5-mm tube ofgel can be used with a 3-mm slab of gel. Then the sample is run in the second direction in a regular slab gel electrophoretic apparatus.

128

GEL-TYPE TECHNIQUES

ACAROSE GEL ELECTROPHORESIS Electrophoresis in agarose was first described by Hjertbn [7] in 1961, and represented a considerable improvement over paper and agar electrophoresis as agarose does not exhibit adsorption effects and its lower content of charged groups results in lower electroendoosmosis. Although the resolution of agarose gel electrophoresis is lower than that of polyacrylamide gel electrophoresis as no sieving effect is involved in the separation, it is adequate, for example, for routine examinations of serum or other biological fluids for screening purposes and in clinics. Both the sensitivity and the resolution of agarose gel electrophoresis can easily be increased and the quantitative information obtained by combination with one of the immunotechniques (see Chapter 7). The lower resolution of agarose may even be an advantage as the complex pattern of, eg., serum proteins on acrylamide gel electrophoresis is not yet of diagnostic value. Agarose gel extends the useful range of polyacrylamide gel electrophoresis, making it possible to separate the largest molecules and cellular particles. Many viruses, enzyme complexes, lipoproteins and nucleic acids that are beyond the usable pore size of polyacrylamide gel (ca. 50 pm in 2.6% gel) can be examined by using agarose or agaroseacrylamide composite gels. Agar is a polysaccharide with a structure of repeating units of 1-4 linked 3,6-anhydroa-L-galactopyranose and 1-3 linked 0-D-galactopyranose. The gel is formed by dissolving a dry polymer in boiling water and cooling the mixture to room temperature. A noncross-linked network of highly porous gel is formed that is rich in hydrogen bonds. Usually 1.5% agarose solutions are used; the 7-8% gel was shown to give an excellent resolution, comparable t o that with acrylamide gel [20]. Chemicals and solutions [ 2 1 ] Buffer system: Barbital-sodium barbital buffer (pH 8.6), 0.075M(for plasma proteins also containing 2 mM calcium lactate) Fixing solution: Picric acid (14g) is suspended in 1 1 of distilled water, warmed t o 35-40°C and filtered, then 200 ml of glacial acetic acid are added. Staining solution: Amido Black 10B (or Coomassie Brilliant Blue R-250) 5g Ethanol (96%) 450 ml Acetic acid (glacial) 100 ml H2O 450 ml The dyes are added to the ethanol-acetic acid mixture and left at room temperature overnight. The solution is filtered and water added, Destainer: Ethanol (96%) Acetic acid (glacial)

H2O

250 ml 100 ml 450 ml

AGAROSE-ACRYLAMIDE COMPOSITE GELS

129

Apparatus The apparatus for horizontal gel electrophoresis described in Chapter 7 is suitable for agarose gel electrophoresis and for immunoelectrophoresis. It consists of a support plate (120 x 220 mm) with an in-built cooling coil, platinum electrodes mounted underneath the cooling coil in protecting grooves, a basic container forming the buffer reservoirs and a protective lid with safety connectors. Working procedure [21] A 1.5-g amount of agarose is suspended in 100 ml of barbital buffer and the suspension is heated with continuous stirring until a clean solution is obtained. The agarose is cooled to 60°C and 25 ml are spread on a 110 x 200 x 1 mm glass plate placed on a levelling table in a horizontal position. The glass plate should be cleaned with detergent, washed with ethanol and warmed to 50-60°C. The slit-forming comb is then placed on a plate and left there for 15 min until a gel has formed. The slit former is carefully removed and the samples are immediately applied; 5pl of serum or another sample containing 10-100 pg of proteins are placed in the slits. The gel plate is positioned in the electrophoretic apparatus and the sides of gel are connected to the buffer reservoirs with the aid of paper wicks (7-10 pieces of Whatman No. 1 chromatographic paper). A better electrical connection is secured if the wicks are made of 1.2% agarose gel poured into the frame slots for moulding gel wicks w h c h are piaced next to the gel plate. The buffer reservoirs are filled with buffer and the levels equalized with the aid of a level regulator (a U-shaped glass tube which is filled with electrode buffer and the ends immersed into the buffer reservoirs until the levels equalize). The unequal liquid level in the buffer vessels would cause a flow of buffer through the gel layer. The electrophoretic separation can be performed at 20 V/cm if adequate cooling is ensured. When the marker dye (0.1% bromphenol blue) has migrated ca. 50 mm, the run is interrupted. The plate is transferred into the fixing solution for 15 min, washed with ethanol for 2-3 min and covered with Whatman No. 1 filter-paper to dry the gel layer. The dried plate is immersed in the staining solution for 5 min and destained by two or three washes with the staining solution. The plate is allowed to dry in a stream of warm air.

AGAROSE-ACRY LAMIDE COMPOSITE GELS The poor mechanical properties of low-density acrylamide gels (below 3%T) prevent the application of acrylamide gel electrophoresis for the separation of molecules with molecular weights above lo6.The addition of agarose t o the acrylamide gel mixture to give composite gels results in excellent mechanical properties. The separation range of the composite gels can be varied by changing the gel porosity, which can easily be achieved by varying the concentration of acrylamide in the gel mixture. In this fashion composite gels can be formed that are especially suitable for the separation of nucleic acids. In principle, the composite gels are prepared by the following procedure. The solution of acrylamide and agarose is mixed together with the polymerizing catalyst at a

130

GEL-TYPE TECHNIQUES

temperature higher than the gelling point of agarose. After polymerization of acrylamide has occurred, the temperature is decreased and agarose gels in the network of the acrylamide gel. By varying the acrylamide concentration, gels of different porosity can be prepared. Chemicals and solutions Stock solutions Acrylamide monomer solution, 20% (19g acrylamide, 1 g Bis, H20 to 100ml) Dimethylaminopropionitrile, 6.4% (v/v) Ammonium persulphate, 1.6%(w/v) Electrode buffer (pH 8.3) Tris 108g EDTA (Na2 salt) 9.3 g Boric acid 55 g H2O to 11 Working procedure [ 8 ] Gel preparation (2%acrylamide, 0.5%agarose)

A 1.O-gamount of agarose is placed in an erlenmeyer flask with 100 ml of water and the mixture is refluxed for 15 min at 100°C. The agarose solution is then cooled to 40°C. A 20-ml volume of stock buffer solution, 10 ml of DMPN solution and 20 ml of acrylamide monomer solution are mixed, the mixture is warmed to 35°C and 10 ml of ammonium persulphate solution are added. The two solutions are mixed and the gel solution is poured rapidly into an electrophoretic cell maintained at 2OoC. After 1 h, when both the acrylamide and agarose gels have formed, the gel is cooled to 5°C and pre-electrophoresis is performed for 45 min. The samples are applied and the electrophoretic run is performed in gel slab apparatus as described on pp. 125-126. The agarose-acrylamide composite gels are suitable for the electrophoresis of DNA molecules with molecular weights in the range from 1 x lo6 to 2 x lo6. From the electrophoretic mobility, the molecular weight of the RNA (DNA) molecules can be obtained [81REFERENCES 1 2 3 4. 5 6 7 8

A. Tiselius. l h n s . Famdzy SOC.,33 (1937) 524. 0. Smithies, Biochem. J., 61 (1955) 629. S. Raymond and L. Weintraub, Science, 130 (1959) 711. L. Ornstein,Ann. N. Y.Acad. Sci., 121 (1964) 321. B. Davis, Ann. N.Y. Acad. Sci., 121 (1964) 404. A. L. Shapiro, E. Vinuela and J . V. Maizel, Biochem. Biophys. Res. Commun.. 28 (1967) 815. S. Hjertkn, Biochim. Biophys. Acta, 53 (1961) 514. A. C. Peacock and C. W. Dingman, Biochemistry, 7 (1968) 668.

REFERENCES 9 10 11 12 13 14 15 16 17 18 19 20 21

J. Margolis and K. G. Kenrick, Anal. Eiochenz., 25 (1968).347. K. A . Ferguson, Metabolism, 13 (1964) 985. D. Rodbard and A. Chrambach, Proc. Nut. Acad. Sci. U . S . , 65 (1970) 970. D. Rodbard and A. Chrambach, Anal. Biochem., 40 (1971) 95. J. L. Hedrick and A. J. Smith,Arch. Eiochem. Eiophys., 126 (1968) 155. E. G. Richards, J. A. Coll and W. B. Gratzer, Anal. Biochem., 12 (1965) 452. F. Kohlrausch, Ann. Phys. Chem., 62 (1897) 209. J . A. Reynolds and C. Tanford,J. Eiol. Chem., 245 (1970) 5161. C. Manwell, Eiochem. J . , 165 (1977) 487. L. J . Mets and L. Bogorad, Anal. Biochem., 57 (1974) 200. P. H. O’Farrell, J. Eiol. Chem., 250 (1975) 4007. H. Hoch and C. G. Lewallen,Anal. Eiochem., 78 (1977) 312. B. G. Johansson, Scand. J. Clin. Lab. Invest., 29 (1972) Suppl. 124, 71.

131

This Page Intentionally Left Blank

Chapter 7

Quantitative immunoelectrophoresis P. JUST SVENDSEN

CONTENTS Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antibodies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus and accessories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Procedure for crossed immunoelectrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . Procedure for crossed immunoelectrophoresis with an intermediate gel . . . . . . . . . . . Procedure for fused rocket immunoelectrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . Procedure for rocket immunoelectrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . Procedure for crossed-line immunoelectrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

133 134 134 134 136 136 139 139 141 143 148 150 153

INTRODUCTION Classical immunoelectrophoresis, introduced by Grabar and Williams [ 11, was a great advance in the analysis of complex biological fluids, For the first time it was possible to resolve further components separated by electrophoresis in agar gel. However, the method did not give precise quantitative data and the resolution was adversely affected by a diffusion step. This problem was solved by Laurell, who introduced rocket immunoelectrophoresis [2] and crossed immunoelectrophoresis [3]. In the classical immunoelectrophoresis devised by Grabar and Williams [I], the antigens separated by electrophoresis are permitted to diffuse against a reservoir of precipitating antibodies. A number of arches will appear, one for each antigen in the sample, provided that the corresponding antibody is present in the antibody preparation. In rocket immunoelectrophoresis [2], specific antibodies are evenly distributed in an agarose gel, the samples are applied into sample wells punched out in this gel and a current is immediately passed through the gel. The antigens will migrate through the gel, and the antigen corresponding to the antibody in the gel will precipitate along the edges of the migrating sample, forming a peak when all antigen has been precipitated. The precipitate will appear with contours like that of a rocket, which led to the name of the method. The area enclosed by the precipitation line, or the peak height, is nearly proportional to the amount of antigen applied in the sample well. In crossed immunoelectrophoresis [3] the antigens are separated by agarose gel electrophoresis. The samples are applied in narrow application slits and, after electrophoresis, thin strips are cut out from the separation tracks and transferred to another plate.

134

QUANTITATIVE IMMUNOELECTROPHORESIS

Antibody-containing gel is poured along the strips and a current is passed through these plates perpendicular to the direction of the current in the initial separation. A series of precipitates will occur in the antibodycontaining gel, each revealing the distribution of the corresponding antigen. This technique is primarily used for studying the heterogeneity of antigens and is only semi-quantitative, as not all of the antigen is transferred to the second-dimension plate. To obtain complete quantitative information, the crossed immunoelectrophoresis introduced by Laurell [3] was modified by Clarke and Freeman [4],who applied the sample in a circular application well and transferred the entire separation track to the seconddimension plate. They showed that the area enclosed by the precipitation line is proportional to the amount of antigen applied in the sample well. Kr6ll introduced line immunoelectrophoresis [5]. In this method, the antigens are evenly distributed in a rectangular gel slab in conjunction with the antibody-containing gel. A current is passed through the gels, and the precipitation lines will appear as straight lines parallel to the antigen-containing gel. The distance between the lines and the edge of the gel that contained the antigens at the start of the experiment is proportional to the amount of antigen introduced into the sample gel. Rocket immunoelectrophoresis according to Laurell [2], crossed immunoelectrophoresis according to Clarke and Freeman [4] and line immunoelectrophoresis according to KrQll [5] are the basic methods of quantitative immunoelectrophoresis. These three methods yield quantitative information that is proportional to the antigen concentration and inversely proportional to the antibody concentration. By combining and modifying these principles a great variety of methods have been developed for application in both routine clinical studies and research [6-91. This chapter describes the use of quantitative immunoelectrophoresis as a tool for following the fractionation of an antigen to purity as an example. An apparatus for conducting the experiments and the chemicals and solutions used are discussed.

CHEMICALS AND SOLUTIONS Chemicals The following chemicals are used: 2-amino-2(hydroxymethyl)propane-l,3-diol(Tris), calcium lactate, 5,5-diethylbarbituric acid (Diemal), sodium azide, glacial acetic acid, sodium chloride, agarose Type H S A, Coomassie Brilliant Blue R 250, ethanol (96%), rabbit anti-human serum immunoglobulins, rabbit anti-human transferrin immunoglobulins. Solutions Tris-Diemal buffer, pH 8.6,ionic strength 0.1 (stock solution): 5,5-Diethylbarbituric acid (Diemal) 224 g 443 g Tris Calcium lactate 5.33 g Sodium azide 10 g Distilled water to 10 1

CHEMICALS AND SOLUTIONS

135

(Dilute 1 part of stock solution plus 4 parts of distilled water to obtain an ionic strength of 0.02) The chemicals dissolve quickly in 5 1 of distilled water with magnetic stirring and no heating is necessary. This buffer is essentially the same as that recommended elsewhere (ref. 6, p. 25), except that the sodium ion has been replaced with Tris, thus increasing the buffering capacity several fold. There were some problems with the earlier buffer, as the buffering capacity often was exhausted, and as a consequence the pH in the cathodic vessel increased to over 11, resulting in dissolution of the immunoprecipitates from the cathodic side. This new buffer does not alter the immunoprecipitin pattern in quantitative immunoelectrophoresis, but stabilizes the reproducibility. 1 % Agarose in Tris-Diemal buffer, pH 8.6, ionic strength 0.02: Agarose ( 2 g) is added to 200 ml of the diluted buffer and dissolved by gentle heating with magnetic stirring. To ensure that the agarose is completely dissolved the solution should boil for 5-6 min. The solution is kept fluid at 56°C in a water-bath and is ready for use after temperature equilibration. Agarose (1%) can be stored at 4°C for several weeks and liquified repeatedly by heating. The agarose should exhibit an MR value of 0.1 3 to obtain good results with rabbit antibodies and Tris-Diemal (pH 8.6, ionic strength 0.02). Under these circumstances the average migration velocity of rabbit antibodies is close to zero, The MR value is a measure of the electroendosmosis and is obtained by subjecting a mixture of human albumin and polydextran to electrophoresis in the actual system. The shift of the polydextran is measured (the migration direction taken into consideration: negative for a cathodic shift, positive for an anodic shift), and this distance is related to the total distance between the polydextran and human albumin. The Litex agarose used in this laboratory is guaranteed to maintain the same electroendosmosis from batch to batch, which for Type H S A is 0.13%. 0.1 M sodium chloride for washing excess of protein out of the plates after immunoelectrophoresis. Staining solution: Ethanol (96%) 4500 ml Distilled water 4500 ml Glacial acetic acid 1000ml Coomassie Brilliant Blur R 250 50 g The dye is dissolved and the solution is heated to 6OoC, cooled to room temperature and finally filtered through filter-paper. Destaining solution: 4500 ml Ethanol (96%) Distilled water 4500 ml Glacial acetic acid 1000ml The destaining solution can be used repeatedly if filtered through activated charcoal after use.

136

QUANTITATIVE IMMUNOELECTROPHORESIS

Antibodies Crude antisera can always be used directly in experiments. However, to obtain low background staining, a crude antibody preparation should be subjected to salting-out [250g of (N&)2S04 per 1000ml of rabbit anti-serum], dialysis and chromatography on DEAE-Sephadex A-50at pH 5 .O in sodium acetate-acetic acid buffer with an ionic strength of 0.05.The purified gamma-globulin fraction is then dialysed against 0.1 M NaC1-15 mM NaN3. This preparation will lose less than 2% of its activiiy per year when stored at 4OC. (For further details, see Chapter 23 of ref. 6.)

APPARATUS AND ACCESSORIES We use 1mm thick glass plates to support the agarose gels. These glass plates are made in the following standard sizes: 5 x 5 , 7 x 7 , 7 x 10,lO x 10 and 11 x 20.5 cm. The apparatus shown in Fig. 7.1A was therefore designed in such a fashion that it will accommodate several of the standard glass plates on the cooled surface (2 in Fig. 7.1B). The cooled surface measures 22 x 12 cm and is capable of holding one 1 1 x 20.5 cm plate, two 10 x 10 cm plates, three 7 x 10cm plates, three 7 x 7 cm plates or eight 5 x 5 cm plates. The design is similar to that described by Johansson [lo], and cooling is obtained by a serpentine-shaped cooling channel cut in a thick Perspex plate, leaving 1 mm of material between the cooling channel and the cooling surface. With a flow-rate of 1 l/min the cooling capacity of the apparatus is sufficient for quantitative immunoelectrophoresis. The temperature of the cooling water should be 10-1SoC. The experiments described below were conducted at a temperature of the cooling water of 15OC. The cooling water is circulated by a cooling thermostat. The cooling plate is attached to two supports (4in Fig. 7.1 B), with moulds attached for casting agarose gel connections to the electrode vessels (1 in Fig. 7.1B), each containing 1 1 of Tris-Diemal buffer (pH 8.6, ionic strength 0.02). The inner walls of the moulds are level with the cooling surface, and the outer walls of the moulds are 1 cm higher than the cooling surface and are level with the end pieces attached to the cooling plate. This central unit is placed in the electrode vessels, which are separated by a Perspex wall. When the lid (5 in Fig. 7.1B), is attached, the platinum-wire electrodes supported by PVC plates (3 in Fig. 7.1B), are automatically connected to the wire and jack-plugs (6 in Fig. 7.1B). The lid rests on the edge of the central unit and is approximately 1 mm from the upper edge of the electrode-vesselunit, thereby sealing the electrophoresis chamber completely from the surrounding air. Even when filter-paper wicks are used to connect the agarose gel plates to the electrode buffer, the gel moulds should not be removed from the central unit. As these moulds are immersed in the electrode buffer, they ensure complete sealing of the electrophoresis chamber. It is an advantage to keep the volume of the electrophoresis chamber small and well sealed from the surrounding air, including that above the electrode buffer, as water would otherwise condense on the cooled agarose gel on humid days or with very low cooling temperatures. If water condenses on the immunoelectrophoresis plate during the experiment, distorted immunoprecipitates will be formed. The lid (5 in Fig. 7.1B) is furnished with five pairs of holes 4cm apart, which provide

APPARATUS AND ACCESSORIES

137

Fig. 7.1. (A) Electrophoresis apparatus for quantitative immunoelectrophoresis, assembled. (B) Main parts of electrophoresis apparatus. 1 = Electrode vessels; 2 = cooled surface; 3 = electrodes on PVC supports; 4 = support with moulds for agarose gel connections; 5 = lid with electric connectors; 6 = jack-plugs for power supply.

access for a test probe (3 in Fig. 7.2) that is used to measure the potential gradient directly in the gel when the current has been switched on. The test probe is connected to the power supply (1 in Fig. 7.2), and a switch on the power supply sets the voltmeter so that the potential gradient can be read directly in volts per centimetre. At the same time, the correct polarity of the current is checked. The electrophoresis unit (2 in Fig. 7.2) is connected to one of the four power outlets on the power supply. Three channels in the power supply have the capacity of delivering 300V and lOOmAd.c., which is sufficient for quantitative immunoelectrophoresis. The fourth channel can deliver 300 V and 200 mA d.c., for use with buffers of higher ionic strength, e.g., electrophoresis in agarose gel using Tris-Diemal buffer (pH 8.6, ionic strength 0.075).

138

QUANTITATIVE IMMUNOELECTROPHORESIS

Pig. 7.2. Apparatus and accessories for quantitative immunoelectrophoresis. 1 = Power supply; 2 = electrophoresis apparatus; 3 = test probe for measuring the potential gradient; 4 = adjustable glass table for casting gels; 5 = device for levelling electrode buffers; 6 = gel punchers; 7 = adjustable template for punching samples wells; 8 = vessel for washing, staining and destaining; 9 = racks for holding glass plates; 10 = U-shaped frame for casting gels between glass plates; 1 1 = long razor blades for cutting and transferring gel slabs; 12 = slit formers.

The gels are usually cast on levelled glass plates, and for this purpose a levelling glass table (4 in Fig. 7.2) is used in combination with a precision spirit level. To ensure that no hydrodynamic flow occurs during immunoelectrophoresis due to a different level of the buffers in the electrode vessels, the buffer in the vessels is levelled by means of a U-tube (5 in Fig. 7.2). The U-tube is removed before immunoelectrophoresis. Through a template (7 in Fig. 7.2), the sample wells are punched out by means of a puncher [ 1 I ] (6 in Fig. 7.2), which in a single operation punches the well and removes the agarose gel plug. The puncher is connected to a suction device, and is made of two annular steel tubes. The outer tube has a sharp edge and cuts the gel on impact with it. The inner tube, which is spring-loaded, slides downwards when a mechanical pressure is applied, and sucks up the agarose gel plug. Slits in the outer tube serve as air inlets to release the vacuum introduced through the inner tube, thus avoiding damage to the walls of the well. The template is made of a base plate to which a sliding ruler is attached. The ruler has a pattern of holes with the same diameter as the upper part of the puncher, for punching all the patterns of sample wells needed in quantitative immunoelectrophoresis. The punchers are available for sample wells with diameters of 2.0,2.5,3.0 and 4.0 mm. Racks (9 in Fig. 7.2) for holding the various plate sizes fit into the vessels (8 in Fig. 7.2) for washing, staining and destaining purposes. For rocket immunoelectrophoresis on 1 1 x 20.5 cni glass plates, a matrix is made of a 1.5 mm thick U-shaped frame (10 in Fig. 7.2), set between two glass plates, held together by strong paper clamps. With a pipette the antibody-containing solution is poured into the matrix. If the rear glass plate is shifted a few millimetres relative to the front glass plate, the filling is greatly facilitated. In two-dimensional immunoelectrophoresis agarose gel slabs are transferred from one glass plate to another, by using long razor blades (1 1 in Fig. 7.2) measuring 2 x 15 cm. If narrow slits are desired, slit formers (1 2 in Fig. 7.2) are used. The slit formers rest on four legs, which are adjusted in such a way that the slit-forming blades are approximately

139

PRACTICAL APPLICATION

I

f

I Fig. 7.3. Cross-sectional area of double constriction pipette.

0.2 mm from the glass plate. This prevents the sample from leaking between the gel and glass plate, as the slit is sealed in the bottom. When the gels have been poured on to the levelled glass plates, the slit formers are placed into the gels before they gel, and are removed just before sample application. Sample application is performed by using double constriction pipettes (Fig. 7.3). The double constriction pipette has the advantage over the single constriction pipette of not being fully emptied. This prevents air being blown into the filled wells, causing overflow and contamination of the surrounding gel. The double constriction pipette is filled to the upper constriction and emptied by blowing gently until the meniscus reaches the lower constriction. The tip of the pipette must touch the bottom of the sample well in order to deliver precise and reproducible volumes. These pipettes are available in various sizes from 1 p1 to several millilitres.

PRACTICAL APPLICATION Quantitative immunoelectrophoresis is an almost indispensable tool for following the fractionation of a protein t o purity, and for testing the isolated material for purity. Antibodies to the crude antigen mixture, in the following human serum, must be available; if not commercially available it may be raised in rabbits, following the procedure described in Chapter 23 of ref. 6 . The purification of human serum transferrin is described below. The first step is to investigate the crude antigen mixture by crossed immunoelectrophoresis according to Clarke and Freeman [4]. Procedure for crossed immunoelectrophoresis A 10 x 10 cm glass plate is washed with detergent, rinsed with ethanol and dried. To obtain a 1.5 mm thick agarose gel, 15 ml of agarose solution at 56°C is poured on to the glass plate, placed on the levelled glass table. After congelation of the agarose, five wells are punched out by means of a 2.5-mm gel puncher using the template shown in Fig. 7.4A. A 2 4 volume of human serum is applied into the sample well(s) S1 to S4 using a double constriction pipette. The fifth well, M, is used to apply a marker, e.g., albunin stained with bromophenol blue. The glass plate with the agarose gel is placed on the cooling surface of the electrophoresis apparatus. A conducting connection to the electrode vessels is obtained with buffer-saturated filter-paper wicks (Whatman No. 1, eight layers, 10 x I0 cm). It is important to ensure that the wicks have good contact with the gel. Electrophoresis in the first dimension is run with a field strength of 15 V/cni. This is tested in every run with a test probe in order to obtain the same potential gradient in the runs. The tinie for the run in the first dimension is approximately 55 min at I5'C. The electrophoresis is discontinued when the blue marker has migrated 5.5 cm. The agarose

140 A

QUANTITATIVE IMMUNOELECTROPHORESIS

1.DlMENSlON Cr+

B

2.DIMENSION Cr+

1

I

I

I

I

I

I I I

I I I I I

I

I

I

I I

I I I I

I

I I

I

I I I

I

I

I

I I I I

I I

I I I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I 1

I

I I I I I I

I

I I I

I

o Mc

I

I I I1

I

0

3

f

I I

I I I I

I I

I

I

0 5 2 1 0

I

2

;0 2 4

I

I

I

I

I

I

I

I

Fig. 7.4. Crossed emunoelectrophoresis. (A) Template for the first-dimensiongel. (B)Template for the seconddimension plate. The hatched area indicates the antibodycontaining gel. (C) Crossed immunoelectrophoresis of 2 pl of human serum into l000pl of anti-human serum (1 2.5 pl/cmagel).

gel is cut according to Fig. 7.4A following the dotted lines. The track with the marker is discarded. The first-dimension gel slabs are then transferred with a long razor blade to the seconddimension glass plates, which have been coated with agarose solution and dried (see Fig. 7.4B). The glass plates are then placed on the levelled glass table, and 12 ml of antibodycontaining solution (1 1ml of 1%agarose 1 ml of antihuman serum) are poured on to the upper part of the glass plate, corresponding to the shaded area in Fig. 7.4B. After 10 min the gel is ready, and the plate is placed on the electrophoresis apparatus and connected to the electrode buffer by means of filter-paper wicks (five layers of Whatman No. 1,lO x 10 cm). Immunoelectrophoresis in the second dimension is performed at 3 V/cm for 18-20 h. After electrophoresis, non-precipitated proteins must be removed, which can be ,effected efficiently by pressing the gel under six layers of filter-paper. Air trapped under

+

PRACTICAL APPLICATION

141

the filter-paper will damage the gel, and therefore the application wells must be filled with distilled water prior to the pressing procedure. The first layer of filter-paper is applied wet, followed by five layers of dry filter-paper. A 6-8 mm thick glass plate should be placed on top of the filter-paper to sustain a smooth pressure all over the gel. After 5 min the upper five layers of filter-paper are renewed and the gel is pressed for another 10 min. A polyspecific antibody preparation of high titre usually contains a larger amount of proteins than a monospecific reagent. Therefore, to obtain low staining of the background, the plate should be rinsed for I 5 min with 0.1 M sodium chloride solution, pressed under filter-paper and rinsed for a further 15 min in distilled water. After a final pressing, the filter-paper is moistened with distilled water and removed. The plate is dried in a stream of hot air and then stained in Coomassie Brilliant Blue R 250 for 5 min, rinsed three times with ethanol-water-acetic acid and dried. Many enzymes remain active in immunoprecipitates, and if enzymes are to be stained with specific enzyme substrates the plate should be dried in cold air and developed before normal protein staining. Crossed immunoelectrophoresis performed as described here with 2 pl of human serum and 1 ml of anti-human serum is shown in Fig. 7.4C. Identification of the antigens causing the individual precipitates is performed immunochemically by using monospecific antibodies or pure antigens if available. Alternatively, the antigens can be identified by histochemical methods such as staining for lipid or staining for enzymatic activity, or by autoradiography for the detection of the sepcific binding of radioactively labelled substances. With Clarke and Freeman’s modification of Laurell’s crossed immunoelectrophoresis, individual proteins are quantitated by measuring the area below the corresponding peak. Quantitation by measuring the peak height may not always be possible as many of the peaks are asymmetric, and therefore the peak area is measured electronically. The stained plate is enlarged nine times in an enlarger, and with a light-spot planimeter the curves are integrated. The areas are expressed in arbitrary units on an electronic counter, and when compared with a standard serum the method can be used for quantitative determinations. The next step in the fractionation experiment is to establish which of the numerous peaks appearing on the plate in Fig. 7.4C corresponds to transferrin. This can be done immunochemically by employing the intermediate gel technique introduced by Svendsen and Axelsen [12], if a monospecific antibody preparation is available.

Procedure for crossed immunoelectrophoresis with an intermediate gel The firstdimension agarose gel electrophoresis and transfer of the agarose gel slabs t o the coated seconddimension glass plates is the same as for crossed immunoelectrophoresis. The casting of the antibody-containing gels is shown in Fig. 7.5A. A 1 mm thick glass plate is placed 2.2 cm from the first-dimension gel slab, along the dotted line. A 3.2-ml volume of agarose solution mixed with 75 p1 of anti-human transferrin is poured into the open space between the gel and the glass plate. A reference plate is made in parallel, replacing the anti-human transferrin with 0.1 M sodium chloride solution. When the gel has solidified, the glass plate is cut free, the gel is trimmed along the solid line below the dotted line, and the thin gel strip is discarded. An 8.2-ml volume of agarose solution is mixed with 800pl of anti-human serum and poured on to the remaining part

142

QUANTITATIVE IMMUNOELECTROPHORESIS

Fig. 7.5. Crossed immunoelectrophoresk with an intermediate gel. (A) Template for the seconddimension plate. (B) Reference plate; 75 pl of 0.1 M NaCl were added to the intermediate gel. (C) 75 pl of anti-human serum transferrin were added to the intermediate gel. (B) and (C) 2 p1 of human 1 anti-human serum in the upper gel. serum were applied in the sample well. Antibody: 8 0 0 ~ of

o f t h e glass plate, and the same is done on the reference plate. After immunoelectrophoresis, rinsing, pressing, drying and staining the plates are ready (Fig. 7.5B and C). On the plate in Fig. 7.5B all antigens have passed freely through the intermediate gel. The skewed baseline of the immunoprecipitates is due to decreasing mobilities of the antigens and electroendosmotic transport of slowly migrating antibody molecules. On the plate in Fig. 7.4C all antigens except transferrin have passed through the intermediate gel. As transferrin has been trapped in the intermediate gel, it is missing from the upper gel. A simple comparison of the two plates easily indicates which of the peaks in the reference plate corresponds to transferrin. This peak is marked with a heavy vertical arrow, and the transferrin peak is marked in the same way in all subsequent immunoelectrophoresis plates. The horizontal line in Fig. 7.5C is caused by excess of antigen used to absorb the antibody preparation to make it monospecific. For further details, see Chapter 7 of ref. 6. We have now established which of the peaks in crossed immunoelectrophoresis

PRACTICAL APPLICATION

143

Fig. 7.6. Crossed immunoelectrophoresis. (A) 2 pl of human serum were applied in the sample well. (B) 5 PI of the supernatant from the salting-out experiment described in the text were applied to the sample well. (B) and (C): antibody: lOOOpl of anti-human serum.

corresponds to transferrin. The first step in the fractionation procedure is salting-out with (NH4)$04. To 218 ml of human serum (1000 donor pool) were added 54.3 g (NH4)*S04 and 10.9 mg of FeS04. This mixture was left at room temperature overnight with magnetic stirring, and then centrifuged at 20,000 r.p.m. for 20min. The supernatant was dialysed against water, and finally equilibrated against sodium acetate-acetic acid buffer (pH 5.6, ionic strength 0.05). The supernatant was subjected t o crossed imniunoelectrophoresis and compared with serum (see Fig. 7.61, and was then applied to a column containing 425 ml of DEAE-Sepharose CL-6B, which was equilibrated with the same buffer. The proteins were eluted by increasing the concentration of the buffer to an ionic strength of 0.7 (linear gradient), and fractions were collected in an LKB Ultrorac fraction collector at 2 O C , flow-rate 30 ml/h, one fraction per hour. After the experiment, 68 fractions were collected and analysed by fused rocket immunoelectrophoresis [ 13, 141, which is superior to W monitoring as an elution profile for each individual antigen is obtained in a single experiment. UV monitoring is employed for detecting only where the UV-absorbing substances have been collected. Procedure for fused rocket immunoelectrophoresis A 10 x 10 or 1 1 x 20.5 cm glass plate is used. Volumes for a 1 0 x 10 cm plate are given first, followed by volumes for an 11 x 20.5 cm plate in parentheses. A 10 x 10 cm coated glass plate is placed on the levelled glass table and a 1 mm thick glass plate is placed 2.7 cm from the lower edge of the coated glass plate along the dotted line in Fig. 7.7A.A 4.1 (8.3)ml volume of 1% agarose solution is poured on to the plate and allowed t o congeal. The glass plate is cut free and the gel is trimmed following the solid line below the dotted line in Fig. 7.7A. A 10 (23) ml volume of agarose solution is mixed with 1 (2)ml of anti-human serum, poured on to the remainder of the glass plate and allowed to congeal. Sample wells are then punched out through the template (7 in Fig. 7.2), according

144

QUANTITATIVE IMMUNOELECTROPHORESIS

Fig. 7.7. Fused rocket immunoelectrophoresis. (A) Template for casting the gel and punching the sample wells. (B) Immunoelution profile from ion-exchange chromatography on DEAE-Sepharose C L 6 B ; 2-pl aliquots from the collected fractions were applied in the sample wells. Closed circles indicate the fractions pooled. Antibody: 2000 ~1 of anti-human serum.

to Fig. 7.7 in the antibody-free gel. Thirty-eight sample wells can be punched in a 10 x 10 cm plate and 79 in an 11 x 20.5 cm plate. Samples of constant volume from the collected fractions are then transferred to the sample wells, in the same order as they were collected, by means of a double constriction pipette. The pipette is not rinsed between the individual applications. The plate is left for diffusion for 60 min on the cooled surface in the electrophoresis apparatus. At room temperature in a humid chamber, 30 min will suffice. The diffusion step will cause a continuous precipitation line to be formed for each antigen in the sample, over the sample wells corresponding to the fractions in which they were eluted. The plate is then connected to the electrode buffer by means of five layers of fdter-paper, and immunoelectrophoresis is conducted at 3 V/cm overnight (1 8-20 h). After electrophoresis the plate is pressed, washed, stained and dried as described above. Fig. 7.7B shows fused rocket immunoelectrophoresis of the fractions collected from the DEAE-Sepharose CG6B experiment. A number of elution profiles appear, and

PRACTICAL APPLICATION

145

Fig. 7.8. Crossed immunoelectrophoresis; 5 pl of the pool from the ion-exchange experiment on DEAE-Sepharose CLdB were applied in the sample well. Antibody: 1000 PIof anti-human serum. Fig. 7.9. Fused rocket immunoelectrophoresis; 1-pl aliquots of the fractions from the ion-exchange experiment on CM-Sepharose CL-6B were applied in the sample wells. Antibody: 2000pI of antihuman serum.

transferrin was detected by the intermediate gel immunoelectrophoresis technique (not shown) and is marked with a vertical arrow. The thinner horizontal arrow indicates haemopexin, w h c h is difficult to separate from transferrin. Haemopexin is marked in the same manner in subsequent plates. The fractions marked with closed circles (Nos. 13-35) were pooled and checked for purity by means of crossed immunoelectrophoresis (Fig. 7.8). The pool (640 ml) was then directly applied to a CM-Sepharose CL-6B ion-exchange column (bed volume 125 ml) equilibrated with the same starting buffer as in the preceding ionexchange experiment. (In Fig. 7.7B it can be seen that transferrin passed the matrix without being bound.) The proteins were eluted by increasing the concentration of the buffer to an ionic strength of 0.4 (linear gradient). Fractions were collected in an LKB Ultrorac fraction collector at 2"C, flow-rate 30 ml/h, one fraction per hour. Fortynine fractions were collected and subjected t o fused rocket immunoelectrophoresis and the result is shown in Fig. 7.9. The fractions marked with closed circles (Nos. 38-43) were pooled and tested for purity by crossed immunoelectrophoresis (Fig. 7.10). The amount of impurities had decreased, but a few can still be seen. Preparative isotachophoresis [ 14,15J was then applied to remove the last traces of impurities, especially haemopexin. The pool (1 64 ml) from the experiment on CM-Sepharose CLdB was concentrated to 25 ml on an Amicon filter and dialysed against the terminator, Tris6-aminocaproic acid (pH 8.9). The leading electrolyte was Tris-phosphate (pH 8.1). The supporting medium was 5% polyacrylamide (5% cross-linking) with N,N'-methylenebisacrylarnide as cross-linking agent, cast in the glass column, cross-sectional area 4.0 cm', with central cooling in the apparatus described in ref. 15. A 120O-pl volumeof Ampholine (pZ range 6-8) was added to the sample as spacers. The experiment was run overnight at 5°C with a constant current of 10mA d.c. The elution buffer was Tris-sulphate (pH 7.1, flow-rate 30ml/h). Fractions were collected every 20min. For more details, see ref. 15.

146

QUANTITATIVE IMMUNOELECTROPHORESIS

Fig, 7.10. Crossed immunoelectrophoresh; 2 pl of the pool from the ion-exchange experiment on CMSepharose CLdB were applied in the sample well. Antibody: 1OOOwl of anti-human serum.

Fig. 7.11. (A) Fused rocket immunoelectrophoresis; 1-pl aliquots of the fractions from the isotachophoresis experiment were applied in the sample wells. Antibody: 1OOOwl of anti-human serum. (B) Crossed immunoelectrophoresis; 2 p1 of the pool from the isotachophoresis experiment were applied in the sample well. Antibody: 1000 pi of anti-human serum.

Fused rocket immunoelectrophoresis was employed to analyse the collected fractions, and the plate is shown in Fig. 7.1 1A. It appears that fractions 22-27 can be pooled with transferrin in very high purity. The pool was then checked in crossed immunoelectrophoresis (Fig. 7.1 1B) and only one peak can be seen, except for an extremely small trace of impurity, marked with a thin vertical arrow. To remove the Ampholine collected with the protein, as well as polymers and aggregates, the pool (56 ml) from the isotachophoretic experiment was concentrated to 15 ml and applied on an Ultrogel ACA 44 gel filtration matrix (gel bed 2000 ml) and eluted at 60 d / h . The buffer was sodium phosphate (PH 7.2, ionic strength 0.415). The collected fractions were analysed by fused rocket immunoelectrophoresis, and from the result in Fig. 7.12A it is evident that a small

147

PR An I C AL APPLICATION

Fig. 7.12. (A) Fused rocket immunoelectrophoresis; 1-p1 aliquots of the fractions from the gel filtration experiment on Ultrogel ACA 44 were applied in the sample wells. Antibody: 100Opl of antihuman serum. (R) Crossed immunoelectrophoresis; 4 p1 of the pool from the gel filtration experiment were applied in the sample well. Antibody: l000pI of anti-human serum.

Fig. 7.13. Rocket immunoelectrophoresis. (A) Template. (B) Left: 5-111aliquots of the pool from the gel filtration experiment on Ultrogel ACA 44 diluted 1 20 were applied. Right: 5-111 aliquots of the starting material (human serum, lOOOdonor pool) for the fractionation experiment described in the text diluted 1 t 50, 1 + 20, 1 + 10 and 1 + 5. Antibody: loop1 of anti-human transferrin.

+

impurity has been separated from the transferrin. Fractions 23-29 were pooled (note how precisely fractions can be selected) and finally crossed immunoelectrophoresis of this pool revealed that immunochemically pure human serum transferrin was isolated (Fig. 7.12B). The volume of this pool was 200ml. The recovery of transferrin compared to the starting material can now be measured by Laurel1 rocket immunoelectrophoresis.

148

QUANTITATIVE IMMUNOELECTROPHORESIS

Procedure for rocket immunoelectrophoresis

A 1 0 x 10 cm coated glass plate is placed on the levelled glass table, I 5 ml of agarose solution are mixed with loop1 of anti-human transferrin and the mixture is poured on to the glass plate. When the gel has solidified after approximately 10min the sample wells are punched out through a template according to Fig. 7.13A. When using an 11 x 20.5 cm glass plate, a matrix is made of a 1.5 mm thick U-shaped frame set between two glass plates (1 1 x 20.5 cm, one being coated), held together by strong paper clamps. With a pipette the antibodycontaining solution at 56°C (30 ml) is poured into the matrix. After lOmin at 4"C, the gel is ready, and the uncoated glass plate is removed by gently sliding it off. The U-shaped frame is kept on the glass plate until the pressing and staining procedure. An aliquot of the sample, in this instance the pool from the gel filtration experiment, is diluted as recommended by the antibody supplier, for human serum transferrin 1 20 with 0.1 M sodium chloride solution. To obtain a standard graph, aliquots from the starting material, in this instance the 1000-donor pool from which 1 ml was saved, are diluted 1 5 0 , l q.20, 1 10 and 1 + 5. The four standards are applied with a 5-pl double constriction pipette into four of the sample wells, and the sample dilution is aliquots. Immunoelectrophoresis is applied in one or more sample wells, also in 5-4 immediately started with a field strength of 3 V/cm, and the experiment is run for 18-20 h. After electrophoresis the plate is squeezed under filter-paper. No washing is necessary when using the immunoglobulin fraction of the antiserum rather than the total antiserum, if monospecific antibodies are employed. Therefore, the plate can be dried after pressing it once, stained for 5 min in Coomassie Brilliant Blue R 250, washed three times with destaining solution and dried. The plate is now ready for interpretation (see Fig. 7.13B). If the dilution 1 + 20 is assigned to represent loo%,then the other dilutions will represent the following: 1 5 = 350%, 1 1 0 = 191% (1 + 20 = 100%) and 1 50 = 41%. Laurel1 rocket immunoelectrophoresis offers an almost linear correlation between the antigen concentration and the area delineated by the precipitation line, and as this area is nearly proportional to the peak height the latter is used for the calculation, The peak heights are easily determined on graph paper, measured from the upper edge of the sample well, and a standard graph is drawn with the antigen concentration as the abscissa and the peak height as the ordinate. The peak height of the unknown is measured and interpolated on the diagram. The standard graph obtained from the plate in Fig. 7.13B is shown in Fig. 7.14. The interpolation for calculating the concentration in the transferrin pool is indicated by arrows, and the result is that the pool has a concentration of 45% relative to the starting material and the total yield is 41%. If corrections are made for samples taken during the fractionation and the known loss introduced by omitting to wash the precipitate in the salting-out experiment, then the recovery is 51%. If the precipitate is washed, the total yield would be 48%, so washing the precipitate defmitely has a significant effect on the recovery. As mentioned above, the peaks in crossed immunoelectrophoresis can be identified if a pure known antigen is available or, vice versa, an unknown antigen can be identified if the peaks in crossed immunoelectrophoresis can be recognized.

+

+

+

+

+

+

PRACTICAL APPLICATION

I

149

45

Fig. 7.14. Standard graph for human serum transferrin determined by rocket immunoelectrophoresis. Data from the plate shown in Fig. 7.13B. The arrows indicate the interpolation to obtain the concentration of human transferrin in the pool from the gel fitration experiment on Ultrogel ACA 44. The dilution 1 20 is set as 100%.

+

This can be achieved simply by performing crossed immunoelectrophoresis on 2 p1 of human serum, and for the purpose of demonstration by adding 4 pl of the pool containing the purified human serum transferrin to the sample in the sample well. The result of such an experiment is shown in Fig. 7.1 5B. The plate in Fig. 7.15A is a reference plate, where 4 pl of 0.1 M sodium chloride solution were added to the serum in the sample well. By comparing the two plates, the increase in area of the transferrin peak is easily revealed. Another means of identifying an unknown antigen or a precipitation peak by means of a pure antigen is to absorb the antibodies with the antigen. A 2 - 4 volume of human serum was subjected to crossed immunoelectrophoresis, and 100 pl of the pool containing the pure human serum transferrin were added to the antibody-containing agarose solution before casting the gel. A reference plate was made in parallel, replacing 1OOpl of transferrin with lOOplO.1 Msodium chloride solution. Fig. 7.16B shows the resulting plate, and on comparison of this with the reference plate in Fig. 7.16A, the increase in the area of the transferrin peak is obvious. The background of the plate in Fig. 7.16B is dark, owing to the insoluble transferrinlanti-transferrincomplexes formed by the addition of transferrin to the antibodies. This effect can be prevented by centrifugation of the absorbed antibodies before addition to the agarose solution. However, larger volumes must be used in this instance in order to maintain the precision. Crossed-line immunoelectrophoresis [ 161 can be used for the same purpose.

150

QUANTITATIVE IMMUNOELECTROPHORESIS

Fig. 15. Crossed immunoelectrophoresis. (A) 2 pl of human serum + 4 pl of 0.1 M NaCl were applied in the sample well. (B) 2 p1 of human serum t 4 pl of the pool from the gel filtration experiment on Ultrogel ACA 44 were applied in the sample well. Antibody: l O O O p l of anti-human serum.

Fig. 16. Crossed imrnunoelectrophoresis. (A) 2 pl of human serum were applied in the sample well. Antibody: 1000 pl of anti-human serum. (B) 2 p1 of human serum were applied in the sample well. Antibody: 1OOOpl of anti-human serum absorbed with loop1 of the pool from the gel filtration experiment on Ultrogel ACA 44.

Procedure for crossed-line immunoelectrophoresis The first-dimension agarose gel electrophoresis and transfer of the agarose gel slabs t o the coated second-dimension glass plates is the same as for crossed immunoelectrophoresis. A 1 mm thick glass plate is placed 1.2 cm from the first-dimension gel slab (see Fig. 7.1 7) and 1.8 ml of agarose solution mixed with 20 pl of the pool containing the pure transferrin is cast into the open space between the first-dimension gel slab and the glass plate. A reference plate is made in parallel, replacing the transferrin with 20 jd of 0.1 M sodium chloride solution. The glass plate is cut free and the gel is trimmed along the solid

PRACTICAL APPLICATION

151

Fig. 7.17. Template for crossed-line immunoelectrophoresis.

Fig, 7.18. Crossed-line immunoelectrophoresis. (A) 2 pl of human serum were applied in the sample well; 20 p1 of 0.1 M NaCl were added to the gel slab between the first-dimension gel slab and the antibody-containing gel. (B) 2 pl of human serum were applied in the sample well; 20111 of the pool from the gel filtration experiment on Ultrogel ACA 44 were added to the gel between the first-dimension gel slab and the antibody-containing gel. Antibody: 900 pl of anti-human serum.

h e below the dotted line in Fig. 7.17. The thin gel strip is discarded. A 9.5-ml volume of agarose solution is mixed with 900 pl of anti-human serum and cast on to the remaining part of the plate; the same is done with the reference plate. After immunoelectrophoresis, the plate is pressed, washed, dried and stained as for crossed immunoelectrophoresis. The result is shown in Fig. 7.18B. The transferrin peak has been elevated by the transferrin line. If approximately 320jd of the pure transferrin had been included in the antigencontaining gel, the transferrin line would have absorbed all the anti-transferrin and no precipitate would have formed on the plate, i.e., the transferrin peak would have been missing. This method is therefore recommended for removing antibodies to cross-reacting antigens, e.g., when tissue-specific antigens are investigated [ 171.

152

QUANTITATIVE IMMUNOELECTROPHORESIS

Fig. 7.19. Tandem-crossed immunoelectrophoresis. (A) Template. (B) 2 pl of human serum were applied in the sample well to the left; 4 pl of the pool from the gel fitration experiment on Ultrogel ACA 44 were applied in the sample well to the right. The plate was left for diffusion for 60min before electrophoresis in the fmst direction. Antibody: 1000 pl of anti-human serum.

The last method to be described is tandem-crossed immunoelectrophoresis [ 181, which is essentially the same as crossed immunoelectrophoresis, but two samples are applied in the first-dimension gel, the sample wells having a centre-to-centre distance of 10 mm (see Fig. 7.19A). Human serum (2 pl) is applied in the rear sample well and 4 pl of the pool containing the pure human transferrin is applied in the other well in front. Before electrophoresis in the first dimension is started, the plate is left for diffusion for 60 min on the cooled surface of the electrophoresis apparatus at 15'C. For more complex mixtures, a reference plate should be made in parallel, replacing in this instance transferrin with 4 p1 of 0.1 Msodium chloride solution. The diffusion causes the contents of the sample wells to mix partially in the gel. If identical antigens are present in both wells, the corresponding peaks will show reaction of identity, i.e., fuse into a double peak, as shown in Fig. 7.19B. The above methods can be used to analyse and characterize both antigen and antibody mixtures, including the study of partially identical antigens, e.g., comparing antigens from different species or tissues (see Chapter 11 of ref. 6). A prerequisite for employing the methods is that precipitating antibodies must be available or raised and that the antigens have mobilities different from that of the antibodies. If the mobilities of the antigens coincide with that of the antibodies, the former can be modified by carbamylation to change the pZ(see Chapter 20 of ref. 6 ) , and by carbamylation antibodies can be modified to exhibit zero net mobility in agarose at pH 5 , retaining 80%of their antibody titre. The sensitivity of the methods by normal staining with Coomassie Brilliant Blue R 250 is in the range 5-10 ng of antigen applied in the sample well. The sensitivity can be increased 15-30-fold by employing radioactively labelled reagents or enzyme-conjugated antibodies in a sandwich technique, but if the amount of antigen applied in the well is less than 0.3 ng the lower limit for forming a precipitate may be reached [ 191.

REFERENCES

153

REFERENCES

7 8 9 10 11 12 13 14 15 16 17 18 19

P. Grabar and C. A. Williams, Biochirn. Biophvs. Acta, 10 (1953) 193. C. -B. Laurell, hotides Biol. Fluids, 14 (1967) 499. C. -B. Laurcll, Anal. Biochem., 10 (1965) 3 5 8 . t1. G. M. Clarke and T. Freeman, Protides Biol. I;luids, 14 (1967) 503. J . KrQli, Scand. J . Clin. Lab. Invest., 22 (1968) 11 2. N. H. Axelsen, J . KrQll and B. Weeke, A Manual of Quantitative ImmunoelectrophoresiscMethods and Applications, Universitetsforlaget , Oslo, 1973. N. H. Axelsen, Quantitative lmrnunoelectrophoresis New Developments and Applications, Universitetsforlaget. Oslo, 1975. R . Verbruggen, Clin. Chem., 21 (1975) 5 . C. -B. Laurell, Scand. J. Clin. Lab. Invest., Suppl., 29 (124) (1972). B. G. Johansson, Scand. J. Clin. Lab. Invest., Suppl., 29 (124) (1972) 7. B. Weeke and J . P. Thomsen, Scand. J. Clin. Lab. Invest., 22 (1968) 165. P. J . Svendsen and N. H . Axelsen,J. lmmunol. Methods, 1 (1972) 169. N. M. G . Harboe and P. J . Svendsen, unpublished work (1969). P. J. Svendsen and C. Rose, Sci. Tools, 17 (1970) 13. P. I . Svendsen, in 2. Deyl (Editor), Electrophoresis. A Survey of Techniques and Applications. Part A : Techniques, Elsevier, Amsterdam, Oxford, New York, 1979, Chapter 16. J . Krgll, Scand. J. Clin. Lab. Invest., 24 (1969) 5 5 . E. Bock,J. Neurochem., 19 (1972) 1731. J. KrQll, Scand. J. Clin. Lab. Invest., 22 (1968) 79. A. Ingild, personal communication.

This Page Intentionally Left Blank

Chapter 8

Moving boundary electrophoresis in narrow-bore tubes F. M. EVERAERTS and J . L. BECKERS

CONTENTS Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principles of moving boundary electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mathematical model of moving boundary electrophoresis . . . . . . . . . . . . . . . . . . . . . Electroneutrality equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modified Ohm's law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mass balances for all cationic species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Procedure of computation, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . State of the art and comparison with other techniques. . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

155 156 158 158 158 158 159 160 162 165

INTRODUCTION Moving boundary electrophoresis is a collective noun for various electrophoretic procedures, and can be performed in various ways. Commonly the species to be separated determine whether a background conductance of ionic species is wanted or not. This background conductance is particularly needed if the separation of high-molecular-weight substances is to be performed, This co-flow of ionic species gives a good stabilization of e g , proteins, so that a higher load of sample species can be permitted and, moreover, the temperature of the zone does not increase too much. For low-molecular-weight substances, this co-flow of ionic species is not always necessary and therefore both the qualitative and the quantitative evaluation of the information derived is simpler. Also, the choice of detection systems to be used is mainly determined by the ionic species to be separated. Moving boundary electrophoresis has been especially applied in the separation of high-molecular-weight substances (proteins). Because nowadays more sensitive electrophoretic separation techniques are available for protein separation, e.g., disc electrophoresis (p. 1 13) and immunoelectrophoresis (p. 133), mainly workers concerned with isotachophoresis still show interest in moving boundary electrophoresis, because the separation procedure in isotachophoresis is a moving boundary method. It is beyond the scope of this book to deal with all varieties of moving boundary electrophoresis or to go into great detail. Somewhat arbitrarily, therefore, varieties of moving boundary electrophoresis have been selected that enable one to work with the equipment especially developed for isotachophoretic analysis (Chapter lo), in which thermometric, conductimetric and photometric detectors are applied. The model presented can be used for a better understanding of the separation procedure in isotachophoresis,

156

MOVING BOUNDARY ELECTROPHORESIS IN NARROW-BORE TUBES

which is why the separation of low-molecular-weightsubstances is described. References are given, however, to moving boundary electrophoresis in which the separation of proteins is considered.

PRINCIPLES OF MOVING BOUNDARY ELECTROPHORESIS At the end of the nineteenth century, Kohlrausch [ 11 demonstrated that the passage of electric current through a homogeneous conducting solution could not produce any changes in composition, and this was subsequently confirmed experimentally. The concentrations are all determined by a regulation function, the so-called “beharrliche Funktion”:

i

z

A change could be expected only in regions in which discontinuities ( e g , concentration differences or gradients) existed initially. For these regions, similar regulation functions (eqn. 8.1) apply. In its original derivation [2] moving boundary electrophoresis is characterized by the fact that both electrode vessels are filled with a buffering electrolyte, while the middle part is filled with the sample. The dimensions of this middle part (volume, length) are comparable with those of the electrode vessels. The choice of the buffers is normally determined by whether a separation of anions or cations is required. An unavoidable discontinuity in the conducting circuit is due to the boundary between the sample to be separated and the electrode solutions. As soon as the electric current is switched on, the sample ions migrate in the appropriate direction and form separating boundaries both at the front side and at the rear. Again, all of the concentrations inside the zones formed are determined by the Kohlrausch [ l ] regulation function (eqn. 8.1). Suppose that three sample anions, A, B, and C, are sandwiched between two electrode buffer solutions (Fig. 8.1). In moving boundary electrophoresis, at the front side will be obtained three zones, consisting of the ions A, AB and ABC, respectively. The velocity of the first boundary (q)is determined by the choice of the conditions of the buffer solution in the anode compartment, the current density applied and some minor factors (e.g., viscosity, solvent and temperature). If the anion of this buffer (I) has a higher

0

I

Buffer

0 Buffer1 A

I

I

A A‘

0

A

Bi A B C

c,

C

B

B

I

Buffer Il

10

U

0

I

C

I

Buffer

I

Fig. 8.1. Separation of three anionic species, A, B and C, according to moving boundary electrophoresis (schematic). This example is only one of many, because the separation depends on the various

operating conditions, included the sample to be separated.

PRINCIPLES OF MOVING BOUNDARY ELECTROPHORESIS

157

anion A, this boundary is a zone electrophoretic boundary. The velocity of the second separating boundary (v2) and the third separating boundary (v3) is lower than that of the first boundary (v,). The velocities v2 and v3 are determined by the chosen conditions of the anodic buffer solution, the current density applied, the conditions (concentration, effective mobilities) of the sample constituent A, B and C and some other minor factors. At the rear, three zone boundaries can be recognized: ABC/BC, BC/C and C/buffer 11. Again, the velocities of these boundaries (v4,v5 and v6, respectively) are determined by the concentrations and effective mobilities of the sample constituents, the choice of the cathodic buffer and some other minor factors. If the anion of buffer I1 has a lower effective mobility than that of anion C, the last boundary is again an isotachophoretic one. If its mobility is equal to or greater than that of anion C, it is a zone electrophoretic boundary. The concentrations are determined by the Kohlrausch [l ] regulation function (eqn. 8.1). This means that the regulating functions for the zones at the front need not match those at the rear*. Various means can be used to study the migration of the zone boundaries: optical methods for recording the refractive index, recording of the temperature of the various zones, the pH, the potential gradient or the conductivity, or light absorption**. Several workers [3-91 have given mathematical models for moving boundary systems, but it is very difficult to work in practice with the exact models and therefore some further simplifications need to be made. One of the simplifications is the application of the sample to be separated to one of the electrode compartments. This means that we make no use of the buffer solution I1 (Fig. 8.1), i.e., the sample ions A, B and C fill the cathode compartment***. This can be justified because the separating boundaries at the rear do not, in real experiments, give much more information than the zones at the front side [lo]. With this simplification, we need to work with only one regulation function. Of course, after this simplification, all zones still do not contain one ionic species of the sample and the counter ion originating from the buffer solution I, as in isotachophoretic analyses: the number of ionic species in the zones increases to the rear side. All zones have correlations with both their preceding and following zones, which explains the difficulties involved in computation. A simple model was used by Brouwer and Postema [l 11, who described a model for the separation procedure during isotachophoresis. Concentration effects, the influence of pH and differences in temperature were neglected. Although this is not a general model, it can be used for unbuffered systems of monovalent, fully ionized ionic species, for the calculation of effective mobilities. Here a model is described that is comparable to that of Brouwer and Postema [l 11. With the equations presented, a computer program was developed and the calculations carried out with it are compared with the results of experiments. Further, a procedure for the determination of the effective mobilities of strong electrolytes, using the moving boundary principle, is described. Also, the effective mobilities of, for example. weak fatty It will be clear that the choice of the buffer solution 11 determines whether a co-flow of ionic species is obtained or not.

** For an extensive discussion, see Chapter 11. *** Also, no additional substance is added to this sample to create a co-flow of ionic species.

158

MOVING BOUNDARY ELECTROPHORESIS IN NARROW-BORE TUBES

acids can be determined by working at high pH, where they are fully ionized, for it is assumed that the influence of pH or pK values can be neglected. For average calculations these values can be used as mobilities at infinite dilution. Simpler information on absolute and effective mobilities can usually be obtained from isotachophoretic analysis, however. Some analyses performed in the equipment developed for isotachophoretic analyses in narrow-bore tubes are also described, i.e., the detectors used were the thermometric, the conductivity (or potential gradient) and the photometric detector. The equipment used by Tiselius and Longsworth is described in ref. 10. Mathematical model of moving boundary electrophoresis When carrying out experiments on moving boundary electrophoresis, the narrow-bore tube can be f d e d with an electrolyte of a strong acid if the separation of, for example, cations is desired. The cation present has a mobility that is higher than that of any other cation in the sample. The sample is situated at one end of the narrow-bore tube, in the anode compartment. For the derivation of the equations, the following assumptions are made: fully ionized cations and anions are considered; the contribution of the background ions to the conductance of a zone is negligible; the influence of differences in pH and concentrations are neglected; the electric current is stabilized; diffusion, hydrodynamic flow and electroendosmosis are neglected; and the solution initially present in the capillary tube and anode compartment has a known, constant composition. The equations that need to be considered are the electroneutrality equations, the modified Ohm’slaw and the mass balances for all cationic species.

Electroneutrality equations If the influence of the background ions can be neglected and all species are fully ionized, the concentration of the counter ions will always be identical with the concentrations of the cations present in a zone, if monovalent ions are considered.

Modified Ohm’s law

For the modified Ohm’s law we can write the equation (8.2)* if the influence of the presence of the hydrogen and hydroxyl ions is neglected.

Mass balances for all cationic species In the stationary state, the amount of each ionic species passing a separating boundary is equal to the amount reaching the separating boundary. For each ionic species and all separating boundaries we can write The subscript u refers to the uth zone, which contains u ionic species of the sample.

PRINCIPLES OF MOVING BOUNDARY ELECTROPHORESIS +

cA, ,u-i ( ~ u - 1u A r -

C)=

159

-f

c A r . u ( ~ uA,. u

-

(8.3)

Substituting for Gu : -

+

-

vu = EuuAu

gives -

cAr,u-l

’)

I?,,

EuU A , - U A cAr,u(’ Eu-l U A , - Eu u A u +

Assume - - * + Qlu-1.u = E u - l I E u

then cAr,u-l

-

c A r , ~

Procedure of computation

For a separating boundary between the zones u - 1 and u , eqn. 8.2 gives u -1 zu-i

c

r=1

U

(uAr

+ uB)cAr,U-1 -

z u

C (u.,+ u B ) c A , , u r=l

(8-8)

Combining eqns. 8.7 and 8.8 gives

In fact, this is a modification of the Dole polynomals [3,9] and solutions for the equations are valid if (8.10)

If the compositions of the leading electrolyte and the sample solution are known, all parameters can be computed from the equations given above. The velocity of the concentration boundaries can be neglected. In the first instance, the concentrations of the ionic species in the last sample zone are taken to be identical with the original concentrations in the sample. Although this assumption is not correct, the ratio between the concentrations in the sample remains constant in the zones, according to eqn. 8.3 which gives cAr,u-l

-

ZU

cA,,u 7 Eu-1

(8.11)

Using eqns. 8.9 and 8.10, au-l,ucan be calculated, if all mobilities are known. From and eqn. 8.7, all concentrations of the zone u - 1 can be calculated, and thus all concentrations and a-values can be calculated for all zones. The concentration of the ionic species in the first sample zone can be calculated in two ways, either with the equations given above o r by using the isotachophoretic condition [12].

(Y,,-~,~

160

MOVING BOUNDARY ELECTROPHORESIS IN NARROW-BORE TUBES

In the first computation, we chose arbitrarily as concentrations for the ionic species in the last sample zone those concentrations present in the original sample solution, and all quantities could be obtained. If the parameters of the first zone obtained in this way did not agree with those obtained by the isotachophoretic method, we re-computed from the first to the last zone with the quantities obtained with the isotachophoretic condition, using the LY values form the moving boundary procedure. By this calculation, new concentrations for the sample zones can be proposed and new 01 values can be calculated. This iteration procedure must be repeated until the 01 values and concentrations fit. PRACTICAL APPLICATION With the equations presented above, a computer program was developed and experiments were carried out in order to check this model. To this end, all concentrations should be determined in all zones. However, as this is difficult, another possibility is to measure the velocities of the zones by means of a defector. Each zone has a specific constant velocity, Zu = EuUAu ;for practical reasons we use relative velocities instead of absolute velocities*:

Z:

+

= Zu/$L = (~uu~,,)/(~Lu~L)

(8.12)

If the distance between the point of injection and the point of detection is x, the time needed for a particular ionic species to be detected will be -+

--f

tu =

X / v u = x/(uAuEu)

or -+

x = vutu

(8.13)

-

The correlation between U,, and U, will be -+

x = vutu = VLtL

and hence -+

v,' =

-+-+

Vu/V=

= tJtu

(8.14)

The times of detection can be measured from the time of the starting point of the analysis up to the time of the appearance of the zone of a particular ionic species. Because the velocity of the leading electrolyte zone is equal to the velocity of the first sample zone (isotachophoretic condition), we use the relationship -+

v:

= tJtu = tJt,

In this way, the measured ratio t J t , from the electropherograms can be used to check the computed ratio gu/gL In order to check the model, some experiments were carried out. The values of t l / t ,

(z).

* The relative velocity of a moving boundary zone is related to the isotachophoretic velocity of the leading zone (isotachophoretic condition).

161

PRACTICAL APPLICATION TABLE 8.1 THEORETICAL AND EXPERIMENTAL RELATIVE TIMES OF DETECTION FOR SOME CATIONS 1N A MOVING BOUNDARY ELECTROPHORETIC SYSTEM ~

System A

0.02 1.000 1.oo

0.01 0.848 0.84

0.01 0.805 0.79

0.01

0.736 0.73

0.01 0.664 0.65

Concentration (N) t L / t , , theore tical t L / t , , measured

0.02 1.000 1.oo

0.02 0.889 0.88

0.02 0.845 0.83

0.01 0.753 0.76

0.01 0.671 0.66

Concentration ( N ) t L / t , , theoretical t L l t , , measured

0.02 1.000 1.oo

0.01 0.872 0.87

0.01

0.837 0.83

0.02 0.793 0.78

0.02 0.723 0.71

Concentration (N) tLIt,, theoretical t L / t , , measured

0.02 1.000 1.00

0.01 0.877 0.87

0.02 0.845 0.84

0.01 0.767 0.76

0.02 0.708 0.70

Concentration (N) t L / f , , theoretical tLlt,, measured

C

D

E

1

2

(C ,H 5)4N+ 0.01 0.717 0.70

0.90

K' 0.01 1.000

B

~~~~

(CH,),N' 0.01 0.863 0.85

Value Concentration (N) t L / t , , theoretical tLlt,, measured

3

4 S-

5 1

2

1.oo

3

4

Na + 0.01 0.904

Li'

0.01 0.793 0.79

5

5-

Fig. 8.2. Graphical representation of the theoretical and experimental values for the times of detection for some cations in a moving boundary system (see Table 8.1). r = t , / t u .S = ionic species: 1 = K'; 2 = Na'; 3 = (CH,),N';4 = Li'; 5 = (C,H,),N'. A, B, C, D and E refer to A, B, C, D and E in Table 8.1.

were measured from the electropherograms for different mixtures of K'. Na', Li+, (CH,)d+ and (C2H5)4N+.The leading electrolyte was 0.01 N hydrochloric acid and the current was stabilized at 0.04 Acm-*. Experimental and theoretical values are given in Table 8.1. In Fig. 8 . 2 , these values are represented graphically (the broken lines represent the

162

MOVING BOUNDARY ELECTROPHORESIS IN NARROW-BORE TUBES

experimental values). The experimental values agree well with the calculated values and it can be concluded that this model is suitable in many instances. Because the relative time of detection for a mixture of two ions of known concentrations is constant in a given system and depends only on the mobilities, it can be used for the determination of the effective mobility of an ion. In order to demonstrate this, the theoretical values of all relative detection times as a function of the mobility of a cation are shown in Fig. 8.3. The concentration of the cations varied from 0.01 Nto infinite dilution and the leading electrolyte was kept as 0.0 1N hydrochloric acid. From the experimental values for the relative time of detection given in Fig. 8.3, the mobility can be derived. Measurements were carried out with samples of Na', (CH3)4N' and (C2H5)4N+and the results are shown in Table 8.2. It can be seen that the theoretical and measured values agree. In Fig. 8.3, a linear relationship is obtained for a zero concentration of a cation mixed with 0.01 N potassium chloride in a sample. This corresponds with the theory, because now elution phenomena prevail (a uniform voltage gradient is present over the whole of the capillary tube and consequently the relative times of detection show a linear relationship with the mobility). TABLE 8.2 THEORETICAL AND EXPERIMENTAL MOBILITIES OF SOME CATIONS

t,ltlI

Sample Ion

Concentration

Ion

Concentration

0.01 0.01

Na'

0.01 0.005

0.01

(CH,),N'

"1 K'

K'

0.01 0.01

- s)

Measured

0.8375 0.7930

50.5

51.25 51.5

0.01 0.005

0.7900 0.7200

45.0

45.7 45.5

0.01

0.6770 0.6000

30.0

32.2 33.2

")

0.01

K'

U *10' (cm2/V Theoretical

(C,H,),N+

0.005

STATE OF THE ART AND COMPARISON WITH OTHER TECHNIQUES As shown previously, a theoretical relationship between the relative time of detection and mobility can be used for the determination of effective mobilities, although of course an experimentally obtained relationship between mobility and relative times of detection can also be used. Sometimes, disturbances during isotachophoretic analyses can be understood better by using a moving boundary model, instead of an isotachophoretic model. Moving boundary electrophoresis can hardly be used as an analytical technique. Of course, separations can be carried out in this way and Fig. 8.4 shows an electropherogram for the separation of a mixture o f (CH3)&', NI$, K+,Na', Ca2*, Li*, Co", Mn" and Cu2+in narrow-bore tubes, with H+ as the leading ion and methanol as solvent. The separation is reasonable but an interpretation will be very difficult if the sample is unknown, as both the retention

STATE OF THE ART AND COMPARISON WITH OTHER TECHNIQUES

163

-

t,/L

Fig. 8.3. Graphical representation of the calculated relative times of detection as a function of the mobilities for different concentrations of the cations, mixed with 0.01 N KCI, after the leading electrolyte (0.01 N HCI). U = mobility (lO-’cm’/V * s).

t

2 .I

times (time of appearance of a “moving boundary” zone) and step heights depend on both the mobilities and the concentrations of the ionic species in the sample and the cationic buffer. In comparison to other electrophoretic separation techniques, the moving boundary method can be used successfully only if the sample has a simple composition. This method can probably be used for simple, routine analyses, however, because much simpler apparatus (compared with “isotachophoretic” equipment) can be constructed. With moving boundary electrophoresis, simple chemicals can be tested for their purity. In Fig. 8 . 5 , a separation of anionic species by moving boundary electrophoresis,

164

MOVING BOUNDARY ELECTROPHORESIS IN NARROW-BORE TUBES

I8

I

-

--t

Fig. 8.5. Separation of a mixture of anions according to moving boundary electrophoresis. A solution of histidine (0.02N) and HCl(O.01 N ) was used in the narrow-bore tube (0.2 mm I.D.), applied as the separation compartment, and the anode compartment. The pH was 6.0. The cathode compartment was filled with a solution of chlorate (Ch), naphthalene-2-sulphonate(N2S) and morpholinoethanesulphonate (MES). The concentration ratios are: A, 0.005N (Ch); 0.005N (N2S); 0.005N(MES); B, 0.005N (Ch); 0.005N (N2S); 0.01 N(MES); C, 0.005 N (Ch); 0.01 N (N2S); 0.005N (MES); D, 0.01 N (Ch); 0.005 N (N2S); 0.005 N (MES). A' = increasing UV absorption; R = increasing electric resistance; t = increasing time; e = enlargement of the naphthalene-2-sulphonic acid zone due to the fact that the zones pass first the photometric detector, then the conductivity detector. The conductivity of the first zone (chlorate) is constant because it is an isotachophoretic zone; its zone length is not constant; also, the appearance of the first zone is constant ( t * ) . Both the zone lengths and the step heights of the other zones vary as function of the sample composition. The current density was 0.06 A cm" in all instances. 1 = Chloride; 2 = chlorate; 3 = chlorate + N2S; 4 = chlorate + N2S + MES.

performed in the narrow-bore tube of isotachophoretic equipment (Fig. 10.8), is shown. The zone boundaries were detected with a conductivity and a photometric detector. Because a separation of anions was being studied, the sample originally filled the cathode compartment. The anodic buffer, comparable with the leading electrolyte in isotachophoretic analyses, was kept constant in all four analyses.

REFERENCES

165

From the electropherograms shown, it can easily be seen that the linear traces of both the conductivity and the photometric detector are influenced by the concentration ratios of the ionic species in the sample. The zone lengths, of course, also vary according to the differences in the concentrations of the sample constituents. For theoretical approaches, Fig. 8.5 gives optimal information, especially if the separation procedure in isotachophoretic analyses is being studied. Again, in practice the information deduced from these analyses is useless, because normally neither the concentrations nor the mobilities of the sample constituents are known.

REFERENCES F. Kohlrausch, Ann. Phys. (Leipzig), 62 (1897) 208. A. Tiselius, Trans. Faraday Soc., 33 (1937) 524. V. P. Dole,J. Amer. Chem SOC.,67 (1945) 1119. R. A. Alberty, J. Amer. Chem, SOC.,7 2 (1950) 2361. J . C. Nichol, J. Amer. Chem. SOC.,12 (1950) 2367. J . C. Nichol, E. B . Dismukes and R . A. Alberty, J. Amer. Chem. SOC.,80 (1958) 2160. E. B. Dismukes and R. A. Alberty, J. Amer. Chem. Soc., 76 (1954) 191. D. Tondeur and J . A. Dodds, J. Chim. Phys., 3 (1972) 441. L. G. Longsworth, in M. Bier (Editor), Electrophoresis (Theory, Methods and Applications), Academic Press, New York, 1959, p. 91. 10 L. G. Longsworth, in M. Bier (Editor), Electrophoresis (Theory, Methods and Applications), Academic Press, New York, 1959, p. 137. 1 1 G. Brouwer and G. A. Postema, J. Electrochem. Soc., 1 1 7 (1970) 7 and 874. 12 F. M. Everaerts, J. L. Beckers and Th. P. E. M. Verheggen, Isotachophoresis. Theory, Instrumentation and Applications, Journal of Chromatography Library, Vol. 6, Elsevier, Amsterdam, 1976, Ch. 4 .

This Page Intentionally Left Blank

Chapter 9

Isoelectric focusing N . CATSIMPOOLAS

CONTENTS Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carrier ampholytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Support media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Geltubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . T ~ n h y e r.s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Polyacrylamide gel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sephadex gel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Density gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Free solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two-dimensional methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Im m u no is0electric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Polyacrylamide gel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agarosegel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transient state isoelectric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Instrumental aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluation of the steady state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Minimal focusing time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution, resolving power, segmental pH gradient and isoelectric point . . . . . . . . . . . . Defocusing and refocusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

167 168 170 171 172 173 174 174 175 176 176 176 177 178 179 182 184 184 187 188 190 190

PRINCIPLE Isoelectric focusing is an electrophoretic method that utilizes the migration behaviour of amphoteric molecules in a pH gradient to achieve their condensation into narrow isoelectric zones that are stationary in the electric field . The steady-state position of each zone in the pH gradient depends on the isoelectric point (PI)of a particular amphoteric molecule. and therefore isoelectric focusing can be used as a separation technique . The method involves mainly two processes. which can be carried out either simultaneously or separately . These are (a) the formation of a stable pH gradient that increases from the anode to the cathode and (b) the electrophoretic migration of the amphoteric molecules under study (e.g., proteins) towards their respective p l positions with subsequent attainment of the steady state . At present. the stable pH gradient is formed by the use of carrier ampholyte mixtures with specific properties which contain components with pls within a defined pH range .

168

ISOELECTRIC FOCUSING

The pH at which the protein has zero net charge is called the isoelectric point (pl). It is conceivable that if a protein is placed in an electric field with a superimposed stable pH gradient, it will migrate electrophoretically away from the positive electrode if it is positively charged (electrostatic repulsion) or away from the negative electrode if it is negatively charged. In order to ensure that a protein will be negatively charged at the negative electrode and positively charged at the positive electrode, the former electrode is placed in a strong base solution and the latter in a strong acid. Consequently, if the protein is placed in the separation field at any position, it will acquire a certain charge (positive or negative) and it will start to move towards the opposite electrode. However, during the electrophoretic migration in the pH gradient it will reach a pH corresponding to its isoelectric point where its net charge is zero and, therefore, it will be concentrated in that region. The concentration or focusing of a protein zone at the isoelectric point is a steady-state process between electric mass transport and zonal diffusion [ 1-31 . As the protein diffuses away from the isoelectric pH region it again becomes charged and it is forced by the electric field to its steady-state position. As proteins exhibit isoelectric points at various pH values, these can be focused at different regions of the pH gradient, which results in their separation. In addition, measurement of the pH at the position of focusing provides a direct determination of the isoelectric point of the particular species [ 4 , 5 ] . The pH gradient is formed by applying an electric field to a mixture of amphoteric compounds - called carrier ampholytes - which have isoelectric points in close proximity to each other [ 6 , 7 ] . Thus, the carrier ampholytes themselves can be focused by obeying the principle described above for proteins. The ampholyte which has the lowest p l (most acidic) will migrate closer to the anode, where it will condense in its isoelectric state at some distance from the anode because of repulsion. The same process will condense the most basic ampholyte close to the cathode. Carrier ampholytes with intermediate p l values will be focused at different positions along the electric field. The pH gradient is formed because of the overlapping distributions of a series of ampholytes with increasing pls from the anode towards the cathode, which also provide good buffering capacity and conductance. The addition of an anticonvection medium (e.g., density gradient, gel matrix) and a zone detection system (e.g., U V absorbance, staining) is necessary in the application of the method for either preparative or analytical purposes.

THEORETICAL ASPECTS The buffering capacity of an ampholyte is expressed [6,7] as a function of p l - pK, :

-dQ d(pH)

-

2 log 10

2

+ 10@r-PK+)

where Q is the mean electric charge of the ampholyte, p l is the isoelectric point and pK, is the dissociation constant of the more acidic of the two ionizable groups. The buffering capacity of the ampholytes decreases linearly at first and then exhibits an exponential decline. Thus, ampholytes with good buffering capacity should exhibit p l - pK, values as small as possible. Therefore, they should be isoelectric between two ionizable groups

THEORETICAL ASPECTS

1 69

with closely spaced pK values because PI = (PK+ + P K - W

(9.2)

Half of the limiting buffering capacity is still retained at p/ - pK, values of 1.5 pH units. Therefore, values less than 1.5 pH units are desirable. At PI- pK, = 3.5 pH units, only one tenth of the buffering capacity remains and the compound can be considered to be a poor carrier. The extent of protolysis at the isoionic point affects the conductivity of an ampholyte:

a, = 2 / [ 2

+ lO(pI-pK+) 1

(9.3)

As p l - pK, h:is a lower limit of log 2 , the highest value that can be obtained for a, (which is associated with maximal conductivity) is 1/2. Good conductivity and buffering

capacity are associated with small values of p l - pK,, as shown by eqn. 9.3. Thus, the parameter pZ - pK, becomes the most important factor in the choice of carrier ampholytes possessing both good conductivity and buffering capacity. The concentration distribution of an electrolyte has been described [ l ] as an equilibrium between electric mass transport and diffusional mass flow. The equation for this relationship is

cUJ/S,K

=

D(dc/dx)

(9.4)

where c is the concentration of component and ion constituent in arbitrary mass units per arbitrary volume unit; Ui is the electric mobility in ~ ' V - ' S - of ~ ion constituent (except H' and OH-) with + indicating cationic migration and - indicating anionic migration; I is the electric current in amperes; s, is the cross-sectional area in m2 of electrolytic medium measured perpendicular to the direction of the current; K is the conductivity of the medium in S m-l ; D is the diffusion coefficient in ni2 s-' of the component corresponding to the ion constituent with mobility Ui; and x is the coordinate along the direction of current (x increases from zero at the anode towards the cathode). The mass flow per second per square metre is expressed by each term in eqn. 9.4, the left-hand side being the electric and the right-hand side the diffusional mass flow. Given a constant pH gradient and a constant conductivity throughout a focused protein zone, the concentration distribution of the ampholyte can be expressed as a bell-shaped Gaussian curve with inflexion points at

where

and uG denotes the width of the Gaussian distribution of the focused zone measured from the top of the distribution of the focused arnpholyte to the inflection point (one standard deviation). The nature of the carrier ainpholytes determines the course of the pH gradient. The mobility at the isoelectric point, dUi/d(pH), and the diffusion coefficient, D,are characteristic properties of a protein. Low diffusion coefficients (small values of D) and

170

ISOELECTRIC FOCUSING

steep electrophoretic mobility curves at pZ [large values of - dUi/d(pH)] lead to successful isoelectric focusing of proteins. The degree of focusing is directly proportional to the square root of the field strength, E = Z/(sm K) as small values of UG indicate sharply. focused bands. However, owing to undesirable heating effects the field strength cannot exceed certain limits. The Joule heat per cubic centimetre of the medium is given by qj = I*/(&K)

(9.7)

The importance of the conductivity course in isoelectric focusing is shown by eqns. 9.5 and 9.7. A uniform and hgh conductivity course is desirable for the application of a high field strength and avoidance of local overheating. The molecular weight of the protein has very little effect on the degree of focusing as there is a sixth-root dependence of uG on molecular size. This conclusion derives from the third-root dependence of uG on D. The pH difference needed for minimal separation of two protein zones by isoelectric focusing is given by [S]

The numerical value of ApH can be estimated if the pH gradient, d(pH)/dx, is known and the value of UG is obtained from e-1’2 = 0.61 of the peak height. Such calculations have shown that the resolving power of isoelectric focusing can be as little as 0.010.02 pH unit.

CARRIER AMPHOLYTES These amphoteric compounds [ 6 , 7 ] are necessary for the formation of a relatively stable pH gradient. Some of their ideally expected properties are good buffering capacity and uniform conductance throughout the pH gradient, good solubility in water, lack of interaction with proteins, low molecular weight and low UV absorbance, especially at 280 nm. The buffering capacity of these compounds should be high in order to direct the pH of the gradient. The presence of proteins at high concentrations (even up to 1%) at the isoelectric point should not cause a pH disturbance and therefore deviation from the pH gradient ‘monotony’ which should increase consistently from the positive to the negative electrode. Uniform conductance in the pH gradient is also important for several reasons. Regions of low conductance exhibit higher electric field strengths and therefore produce more Joule heat which may cause convective disturbances. As the width of a focused protein zone depends on the electric field strength, among other factors, a nonuniform field strength results in a variable resolution along the column. In addition, the electrophoretic migration of proteins toward their pZ suffers from changes in velocity and therefore it is difficult to predict their minimal focusing time. Carrier arnpholytes of low molecular weight are desirable because they can be removed from the proteins by dialysis or gel filtration after fractionation. A low UV absorbance also facilitates the detection of proteins at 280 nm.

SUPPORT MEDIA

171

Carrier ampholytes available commercially (Ampholine, LKB; Biolytes, Bio-Rad Labs; Phisiolytes, Brinkmann Instruments) do not meet all of the ideal properties as expressed above, but they can nevertheless be used to achleve successful isoelectric focusing separations [8-121. Ampholytes represent a complex multicomponent mixture of homologues of aliphatic polyaminopolycarboxylic (sulphonic) acids with closely spaced p l (inter-species) and pK (intra-species) values. They cover several pH ranges as follows: 3.5-10.0; 2.5-4.0; 3.5-5.0; 4.0-6.0; 5.0-7.0; 5.0-8.0; 6.0-8.0; 7.0-9.0; 8.0-9.5; and 9.0-1 1.O. Their water solubility is good at the concentrations at which they are usually employed for isoelectric focusing. Their absorbance at 280 nm is relatively low, and thus there is only a small interference in the W detection of proteins which exhibit high concentrations when focused at their respective pls. The buffering capacity and conductance depend on the pH range of the particular ampholyte species; they are higher at the more acidic and basic ranges than in the pH range 5-8. The most disturbing effects are the presence of a non-uniform conductance course in the separation column, variability in the concentration of the various isoelectric species, non-uniform viscosity and buffering capacity, and non-linearity of the pH gradient. Further, the pH gradient is usually not stable with time, leading to a ‘flattening’ and ‘drift’ of the pH gradient curve [13-161. The ‘plateau’ phenomenon, as it is called, may cause migration of proteins towards the anodic end of the column and therefore contribute to irreproducibility of isoelectric focusing patterns. This occurs because the focused zones are moving in the electric field instead of being stationary as expected from the state-state theory. However, the flattening of the pH gradient may result in an increased resolution in certain parts of the column away from the negative electrode. Gel filtration studies with ‘‘C-labelled Ampholine [ 171 have shown that the molecular weight distribution of carrier ampholytes is in the range 300-1,000 with a minor part of the species reaching an apparent molecular weight as high as 5,000. It is, therefore, safe to conclude that Sephadex G-50 filtration should remove the Ampholine species from any polypeptide or protein with a molecular weight higher than 10,000 after isoelectric focusing. This depends on the assumption that complex formation of Ampholine and protein is readily reversible, which is usually the case. It should also be mentioned that as the carrier ampholytes consist of amino and carboxylic and sulplionic acid groups linked to aliphatic structures, no chemical reaction with proteins can be normally expected.

SUPPORT MEDIA There are several ways of classifying isoelectric focusing experiments in some useful operational fashion [8,9] . The distinction between preparative and analytical isoelectric focusing pertains to the end result of either collecting the separated material a t the conclusion of the experiment, or being interested only in the analytical aspects without consideration of the fate of the sample. Usually preparative isoelectric focusing is carried out in large columns (volume 100-400 ml) using a sucrose density gradient as support medium. Alternatively, blocks of granular gels (e.g., Sephadex) or zig-zag horizontal electrolysis cells can be used. Proteins are isolated in milligram or even gram amounts. On the other hand, analytical techniques can be performed with microgram or even nanogram

172

ISOELECTRIC FOCUSING

amounts of material in small columns (1-10ml) of density gradient or gel, and in flat gel slabs. With regard to the type of support medium, or the absence of it, we can distinguish several types of isoelectric focusing experiments. These include density gradient, gel, zone convection, and free solution isoelectric focusing. The density gradient is formed by using two different concentrations of sucrose or other neutral substances such as Ficoll and ethylene glycol with special devices, or by layering fractions of different density. Gel isoelectric focusing is carried out either in homogeneous (continuous phase) or granular (beads) polyacrylamide, agarose or cross-linked dextran gels. Zone convection and free solution isoelectric focusing require specially constructed apparatus. The application of specific detection methods involving antigen-antibody precipitin reactions in gels after isoelectric separation gave rise to the technique of immunoisoelectric focusing. Another detection method utilizing U V absorption optics for in situ continuous electro-optical monitoring of the separation field has been called scanning isoelectric focusing. Finally, if the kinetic aspects of pH gradient electrophoresis focusing and isoelectric diffusion can be measured by electro-optical methods, the technique is called transient state isoelectric focusing.

GEL TUBES Isoelectric focusing in polyacrylamide gels (PAGIF) is a high-resolution analytical technique which, in certain respects, offers considerable advantages over the density gradient method [8, 11-13, 18-28]. Some of these desirable features include: (a) resistance to convective mixing; (b) dramatic shortening of the focusing time; (c) employment of a simple apparatus; (d) simultaneous separation of several mixtures; (e) employment of specific stains; (f) combination with electrophoresis in two-dimensional separations; and (g) detection of separated proteins by use of immunodiffusion techniques. Another major advantage of the gel isoelectric focusing technique is that the gel matrix is able to support protein precipitates which may be formed at the isoelectric point. However, most gel isoelectric focusing methods have the inherent disadvantage of restricting the migration of large molecules because of sieving effects. An additional drawback of the utilization of gels as a support medium is that only discrete sample analysis is possible, which imposes limitations in possible automated operations. The essential components for the preparation of the gels include acrylamide and N,N‘-methylenebisacrylamide(BIS), which are polymerized either by light or chemically (TEMED), in the presence of catalysts such as N,N,N’,N’-tetramethylethylenediamine riboflavin and potassium persulphate. The carrier ampholytes (2% w/w of the total gel volume) are incorporated into the gel formulation before polymerization. The protein sample (free of salts) can be mixed with the pre-polymerization solution, or it can be loaded, in the presence of sucrose, on top of the gel and underneath a protective layer of ampholyte. Several methods have been described for the preparation of polyacrylamide gels for isoelectric focusing. The critical factor is usually the gel concentration. In principle, PAGIF should be carried out in gels that exhibit minimal sieving to the proteins under analysis. Separation in isoelectric focusing depends only on charge and not on size.

THIN LAYERS

173

Thus, any restriction in the migration of proteins toward their PIresults either in an increased time to reach the steady state or in an inability of certain species to be focused at all. The increased focusing time coupled with the pH gradient instability can cause serious problems of reproducibility, artifact band formation and incorrect p1 determination. Although a non-restrictive gel in absolute terms cannot be produced, a compromise can be reached by focusing proteins (molecular weight up to 500,000) in 3.75%T, 3.33%C gels, where % T is the per cent total gel concentration (acrylamide and Bis) and %C denotes the per cent ratio of Bis to total gel concentration. For proteins with molecular weights in the range 15,000-60,000, gels of 5%T can be used. The advantages of high % T gels are mechanical stability and formation of sharper bands because of restriction of the diffusion of focused proteins. The pH gradient at the end of focusing can be determined either by slicing the gel into sections followed by elution of ampholytes with water and pH measurement, or directly by touching the surface of the gel with microelectrodes [8, 121. Protein staining is more difficult in PAGIF than in polyacrylamide gel electrophoresis (PAGE) because the Ampholine binds to commonly used stains. Although a variety of procedures have been described for staining and destaining in PAGIF, none is entirely satisfactory [8, 121 . The problem of protein detection can be facilitated by histochemical staining with enzymes, specific staining for glycoproteins and lipoproteins, optical scanning, radioactivity detection and employment of immunoelectric focusing methods' [29,30] utilizing antigenantibody reactions in gels (Fig. 9.1). PAGIF can be carried out in both cylindrical gels and gel slabs with suitable apparatus being available from several manufacturers (e.g., LKB, ISCO,Hoeffer, Ortec, E-C Medical Research Apparatus, Buclller, Brinkman-Desaga). There are certain specific characteristics associated with the tube or slab techniques that may be desirable for a certain type of analysis. The slabs are better for direct comparison of multiple samples. The cylindrical gels are more convenient for sectioning techniques (e.g., immunodiffusion, radioactivity detection in scintillation vials), as the first separation in two-dimensional methods, and for methodological experiments requiring the withdrawal of individual tubes from a single electrolyte reservoir during a run. Other claimed differences are rather trivial.

THIN LAYERS The use of thin-layer plates of polyacrylamide or Sephadex gel as the stabilization matrix for isoelectric focusing [18,22, 31-33] was described at the same time that gel column techniques became available. One of the advantages offered by the thin-layer method is that several samples can be compared simultaneously on the same plate under identical conditions. Many similarities exist between thin-layer polyacrylamide gel and disc isoelectric focusing with respect to the mode of polymerization of the gel and the procedures for staining and pH measurement. However, the apparatus used is different [8, 18,22,31-331. Sephadex (3-75 gels have also been used successfully for isoelectric focusing. Details of the method can be found in the original paper by Radola [31].

174

ISOELECTRIC FOCUSING

Gel slice number

Fig. 9.1. Disc electrofocusing of soybean whey proteins in the pH range 3-10. The densitometer tracing of the stained electrofocused bands was obtained with a Canalco Model F rnicrodensitorneter. Open circles represent the pH gradient along the polyacrylamide gel column after electrofocusing [ 8 ) .

Polyacrylamide gel Many similarities exist between thin-layer polyacrylamide gel and disc electrofocusing. The mode of polymerization of the gel, the nature of the electrolytes and the procedures for washing, staining, destaining and pH measurement can be applied interchangeably. However, the apparatus used is constructed in an entirely different manner. Thin-layer electrofocusing is performed in a horizontal fashion whereas the disc electrofocusing technique requires a vertical arrangement. Two types of apparatus have been devised for thin-layer electrofocusing. Leaback and Rutter [22] used a glass base to carry the gel slab and the electrode wells. A Perspex lid for the glass base has terminals connected to horizontal carbon or platinum electrodes which fit into the trenches. Awdeh et al. [ 181 formed the gel on a glass plate which was inverted and rested on two carbon electrodes spaced 20 cm apart. Catsimpoolas has devised an apparatus based on commercially available equipment [8] . Sephadex gel Thin-layer electrofocusing on Sephadex gels is performed on 40 x 20 or 20 x 20 cm plates coated with a suspension of Sephadex G-75 (Pharmacia) in 1%carrier ampholyte solution (LKB) [31] . The coated plates (0.75-1 .O mm thickness) are mounted horizontally on a metal cooling block. Contact with the electrode vessels, containing solutions of sulphuric acid and ethylenediamine, is made by Whatman 3MM paper strips shielded with a dialysis bag to prevent liquid flow from the electrode vessels to the gel. The protein sample is applied by means of cover slips as described for thin-layer gel filtration [31].

DENSITY GRADIENT

175

Electrofocusing can be completed in 6-8 h at 10-20 V/cm on the 20-cm plates and in 18-24 h at 5-1OV/cm on the 40-cm plates. Protein bands are detected by taking a print of the gel with Whatmann 3MM paper, drying the paper, removing the ampholytes by extensive washing with sulphosalicylic acid or trichloroacetic acid, and staining with Cooniassie Blue R25 or Lissamine Green. Densitometry can be performed in a JoyceLoebl densitometer equipped with a thin-layer attachment. The pH gradient can be measured directly from the gel with microelectrodes.

DENSITY GRADIENT The density gradient in isoelectric focusing serves only as a stabilization medium against mixing of separated components by convection. The solute most commonly employed for the formation of the gradient is sucrose (0-50%), although other compounds such as ethylene glycol and Ficoll can be used if necessary [9, 101 . Desirable properties of materials for the preparation of gradients are: (a) the compound should be non-ionic; (b) it should have a high solubility and low viscosity in water; (c) it should produce density differences greater than 0.12 g/cm3 ;(d) it should not abolish the biological activity of the separated material or interfere with its assay; and (e) it should be readily removed from the components of interest. When using Ficoll, for example, one should bear in mind that its molecular weight is higher than that of many proteins and cannot be easily removed. The preparationof the density gradient can be carried out either by using commercially available gradient mixers or by manually mixing dense and light solutions in certain proportions in a series of test tubes [9, l o ] . The resulting solutions - exhibiting different densities - are then stratified, one on top of the other, by slowly flowing them into the column. In isoelectric focusing, the carrier ampholytes are incorporated in the solutions employed to make the density gradient at a recommended concentration of 2 g of Ampholine per 100 ml of gradient volume. The protein sample can be either added to the total volume of the density gradient or it can be applied as a zone of specified density at a position in the column close to the expected focusing region. In the latter instance. the minimal focusing time can be shortened because the protein does not have to migrate through the entire column to reach its isoelectric point. If dissociating agents such as urea. non-ionic detergents (Triton X-100, Brij 35. etc.) o r reducing agents such as dithiothreitol (DTT) and ascorbic acid have to be employed, these are also incorporated directly into the gradient and the electrode electrolytes. The anode electrolyte is usually placed at the bottom of the density gradient column and consists of a strongly acidic solution e g . , 5% orthophosphoric acid. The cathode electrolyte is a strongly basic solution such as 5% ethanolaniine. The main disadvantages of density gradient isoelectric focusing are the following. The separation (resolution) in the column is much better than in the collected fractions. Proteins diffuse after interruption of the electric field and consequently zone spreading increases proportionally to the time required for collecting the fractions. However, too rapid emptying of the column can cause mixing by swirling of the density gradient. Proteins insoluble at their isoelectric point can precipitate in the column and contaminate

176

ISOELECTRIC FOCUSING

other proteins which are soluble at their pZ. Additionally, there is an element of uncertainty regarding the focusing time required for optimal resolution as proteins and also carrier ampholytes reach their respective PIS at different times. Prolonged focusing, to ensure that all components are at the steady state, is not free of penalty because of the instability of the pH gradient with time and intensity of the electric field, which may result in a decrease or increase in resolution depending on the position of a protein in the pH gradient. The great advantage of the method is that proteins can be obtained in relatively large amounts (i.e., milligram amounts) at a very high purity. Often no further purification is required. Furthermore, dissociating media such as urea and dithiothreitol allow the isolation of proteins subunits and the separation of protein components of membranes, chromatin and other biological materials requiring disruption of their native structure.

FREE SOLUTION Isoelectric focusing without support media has been performed in an apparatus originally designed for free zone electrophoresis [34]. The apparatus contains a watercooled horizontal quartz tube (3 mm I.D.), which rotates around its longitudinal axis at 40 r.p.m. to counteract convective disturbances [35] . In situ W scanning of separations has been demonstrated. In addition, focusing has been performed in pH gradients generated thermally and electrophoretically from buffer mixtures without the presence of ampholytes. In practice, however, these gradients were relatively unstable and exhibited a slope that was either too steep or too shallow for effective focusing. More recently, free zone isoelectric focusing with Ampholine in coils of polyethylene tubing has been described [36].

TWO-DIMENSIONAL METHODS Of considerable promise for resolving complex protein mixtures into their individual components are two-dimensional separation techniques [ 11,20,37-391. Usually PAGIF is the method of choice for the first separation, followed by either PAGE or SDS-PAGE. The combination of PAGIF and SDS-PAGE produces high-resolution maps, the coordinates of which correspond to the pZ and molecular weight of the separated species, if appropriate protein markers are used for calibration. Alternatively, PAGIF can be employed in combination with a number of immunological techniques [29,30] to provide additional separation and sensitive criteria of purity and identity.

IMMUNOISOELECTRIC FOCUSING Immunoisoelectric focusing is a technique that combines isoelectric separation of protein antigens in gels and detection by immunochemical specificity reactions [23,29, 40-431 . Usually, the immunochemical reaction is performed in gels by the process of

IMMUNOISOELEaRIC FOCUSING

177

diffusion. Both qualitative and quantitative information can be obtained, depending on the variation of the technique used. lmmunoisoelectric focusing is a useful adjunct to other forms of immunochemical analysis because of the added dimension of isoelectric separation that offers very high resolution in the fractionation of proteins, which is often not possible with other techniques. The combined features of high resolution and sensitivity of detection render this method suitable for the microanalysis of protein antigens. In addition, certain enzymatic reactions can be performed directly on the immunoprecipitin lines in gels, thus giving supplementary information and definition of a different nature. Two main variations of the technique can be distinguished, involving the isoelectric separation stage prior to immunodiffusion that can be performed in either polyacrylamide or agarose gels. Polyacrylamide gel Isoelectric focusing in polyacrylamide gel columns is first carried out by any of the methods reviewed [8, 10, 111. Subsequently, the gel column is either used intact or is sectioned for the immunodiffusion experiments as illustrated in Fig. 9.2. The intact gel column can be processed immunochemically in at least three modes. It can be embedded in buffered agar gel with trenches cut parallel to the column [40] and filled with antiserum. Immunoprecipitin arcs are formed by immunodiffusion. This technique is useful when an overall qualitative immunodiffusion pattern of a mixture of proteins is desired. A maximum of two different antisera can be used with the same column, However, if the column is sliced once longitudinally, four different trenches can be employed for analysis. An obvious modification of the above method is that the longitudinal strips can be laid directly on an agar layer instead of being embedded. An alternative method [41] that also employs intact gel columns involves the technique of electrophoresis in agarose-gel-containing antibodies [42] . The first step in this procedure is the isoelectric separation in polyacrylamide gel followed by a second electrophoresis at right angles to the first, forcing the separated proteins into the agarose bed containing an antiserum. Migrating zones of antigen-antibody complexes appear very rapidly. At equilibrium, a stable precipitate is formed at the leading edge of the antigenantibody complex that remains stationary. The distance travelled by the peak of a certain antigen can be used to quantitate the amount of antigen present in the mixture. A third variation of the immunoisoelectric focusing technique in polyacrylamide gel using intact columns involves the specific absorption of antisera [43] . After isoelectric focusing, the gel is incubated with absorbed antiserum and stained after elution of nonprecipitated proteins. Other methods [44] that have been described for disc immunoelectrophoresis can also be adapted for immunoisoelectric focusing experiments. A microgel immunoisoelectric focusing method in capillaries has been described [45] . At least three variations of the sectional [46-481 technique can be used. The electrofocused gel column is sliced into approximately 40 sections using any of the commercially available lateral gel slicers. Each section is fragmented by extrusion through a glass microsyringe and placed in sample wells punched in an agar gel plate. When all the sections are extruded into the sample wells, one drop of agar solution is placed in each hole to fill the

ISOELECTRIC FOCUSING

178

00000000 Polyacrylarnide gel

I!& I

Crossed electrophoresis

1

I

Irnrnunoisoelectric focusing

I

Agarose gel

Radial diffusion

- + Elect rophoresis

Specific absorbent

Fig. 9.2. Various forms of immunoisoelectric focusing analysis in gels [30).

vacant space and ensure contact between sample and gel. The sections can be analysed by double-gel diffusion, radial immunodiffusion or immunoelectrophoresis. In double-gel diffusion experiments, a trench is cut parallel to the sample well row and filled with antiserum. Precipitin arcs are formed by ininnmodiffusion. Reactions of immunochemical identity or non-identity between sections are easily recognizable by loop or spur formation of the arcs. If a specific monovalent antiserum is available, the sections can be analysed by the technique of radial immunodiffusion [49] , The antiserum is incorporated into the agar. After diffusion, ring-shaped precipitation bands are formed and migrate concentrically around the holes. A straight-line relationship usually exists between the antigen concentration and the area or diameter of the immunoprecipitate. Standard sections with known amounts of antigen should be included in each plate. A useful extension of the sectional techniques involves inimunoelectrophoresis [50] of the sections obtained from an electrofocused gel column [48]. The additional separation afforded by using electrophoresis after electrofocusing is of definite advantage in the examination of a complex mixture of antigenic proteins. Agarose gel The immunoisoelectric focusing technique in agarose gels formed in a microscope slide technically resembles immunoelectrophoresis [50]with the exception that the buffer is replaced with the carrier ampholytes and electrode electrolytes. An important aspect of the technique is the neutralization of the pH gradient, developed in the agarose during isoelectric focusing, by brief immersion in buffers that will allow the antigen-antibody reactions to occur in a neutral or slightly alkaline environment. The disadvantage of this

TRANSIENT STATE ISOELECTRIC FOCUSING

179

method is the various degrees of electroosmotic flow that are observed with commercial preparations of agarose. This problem has been examined thoroughly by Quast [51], who recommended the use of anion-exchange resin-treated agarose. The ampholyte concentration should be high t o reduce electroosmotic flow, and the viscosity of the liquid in the gel should be increased by adding neutral, chemically inert substances such as sucrose. Also, the anodic solution should not be a strong acid when using agarose gel. Widespread use of the method cannot be predicted until all of the conditions that contribute t o the electroosmotic flow can be alleviated by establishing Htandard methods of agarose purification and isoelectric focusing procedures.

TRANSIENT STATE ISOELECTRIC FOCUSING Transient state isoelectric focusing is a relatively new. but well documented [52-741 technique involving the kinetic aspects of pH gradient electrophoresis in the presence of carrier ampholytes. Before the advent of transient state isoelectric focusing, the most commonly used plot in isoelectric focusing was one in which protein concentration (C) is plotted versus distance (x) (or fraction number) in the separation path with a superimposed relationship between pH and distance (Fig. 9.3, c). At the ‘steady state’, the statistical moments of the protein peak (or a carrier ampholyte peak), i.e., zeroth (rno), reduced first ( m i ) and second ( r n z ) , corresponding to the peak area, position and variance, respectively, should remain constant with time [66,67]. If any of these parameters changes with time, one has n o other choice but t o assume that the system is in the ‘transient state’, n o matter what the cause of this instability may be. Thus, one aspect of the transient state may just reflect a ‘deformation’ of the steady state. An additional aspect of the transient state involves the examination of the kinetic behaviour of a protein (or carrier ampholyte) during its electrophoretic journey towards the steady state. The migration of a protein towards its isoelectric point is a function of its pH-mobility curve, dUi/d(pH) (Fig. 9.3, a) and the pH gradient curve, d(pH)/dx (Fig. 9.3, c). The observed mobility, dUi/dx (Fig. 9.3, b) is the product of the above two parameters. As the mobility, Ui, cannot be measured directly but can be obtained only as a peak velocity the electric field strength, E , is involved, which in turn depends on the conductance of the medium. Non-linear viscosity (in density gradients) and sieving (in gels) effects add t o the complexity of the system. Therefore, one can ask the question: is there any useful parameter that can be measured during the journey towards the steady state? The answer is affirmative, as the minimal focusing time, r M F . (as will be shown below) can be determined at this stage of the experiment [65] . Additional transient states can be intentionally imposed upon the system by interruption of the electric field followed by reapplication after a certain time has been allowed for the free diffusion of both protein and ampholytes (60, 66, 671 . The advantages of this experimental set-up are that only a narrow segment of the pH gradient is utilized. where d(pH)/dx and dUi/d(pH) can be assumed t o be approximately linear, the protein exhibits a Gaussian concentration distribution, the pH gradient is already formed and non-linear viscosity and conductance effects are minimal. The purpose of inducing the defocusing and refocusing states is the

180

ISOELECTRIC FOCUSING

x =

u,

I

PI = C =

02

i

Fig. 9.3. Schematic diagram of the relationships among pH, mobility, distance and protein concentration in isoelectric focusing [74]. TABLE 9.1 TRANSIENT AND STEADY STATES Staee

State

Remarks

Focusing

Trans-1 Trans-2 ss-1 ss-2 Trans4 Trans4 Trans-5 Trans-6 Trans-7 Trans-8

Focusing of ampholytes Focusing of proteins Steady state of arnpholytes Steady state of proteins Deformation of the SS1 state Deformation of the SS-2 state Diffusional defocusing of ampholytes* Diffusional defocusing of proteins* Refocusing of ampholytes** Refocusing of proteins** Steady state of arnpholytes Steady state of proteins

Steady state Deformation Defocusing Refocusing Refocusing steady state

ss-3 ss-4

* In the absence of the electric field.

** After reapplication

of the electric field.

kinetic measurement of apparent diffusion coefficients (0) and the electrofocusing parameter pE, where p = - [d(pH)/dx] [dUJd(pH)] and E is the field strength. These studies have revealed a very significant effect of both carrier ampholyte concentration and protein load on D and p E and therefore on resolution, as this is a function of peak variance, and in turn peak variance is a function of the ratio of D to p E [67] .

181

TRANSIENT STATE ISOELECTRIC FOCUSING

I

Resolutton (R

-E

280-

0

e

*

240-

D

200-

i al

4

Ovolbumin

L

P

4

(pH 3-10 range)

160-

( 2% arnpholytes)

C -

1

50

103

150 200 250 Elapsed time ( m in)

300

350

Fig. 9.8. Plot of integrated peak area versus time of focusing for ovalbumin in pH 3-10 (2%) Ampholine. Arrow B indicates the SS-2 state [ 6 0 ] .

186

ISOELECTRIC FOCUSING

03

02

E, 0 d

I

A

01

C

7,5?

Peak position,

X (cm)

Fig. 9.9. Scanning electrophoretic patterns of histidyltyrosine peaks migrating towards the p l position. Peak 1 migrates away from the negative electrode and peak 2 from the positive electrode. The pI position is arbitrarily set at R = 0. The original sample distribution was uniform. Ampholine concentration, 2% [ 6 5 ] .

Time

x

loF3(set)

Fig. 9.10. Plot of peak position difference ( M )versus time. Histidyltyrosine, 3%; Ampholine concentration, 2% [ 6 5 ] .

migrate from the two ends of the column (positive and negative) until they merge into one at pl. In the pulse mode of loading, only one peak can be distinguished. The time at which constancy of plposition is attained is designated as t M F . A typical example concerning the evaluation of the effect of carrier ampholyte concentration on the f M F

187

TRANSIENT STATE ISOELECTRIC FOCUSING

I

I

I

1

2

1

3

4

I

I

5

6

7

I

8

Time, t (set x o 3 )

Fig. 9.11. Plot of peak position versus time. Soybean trypsin inhibotor, Ampholine concentration 2%. Closed circles represent uniform sample loading, and open triangles pulse negative loading [ 741.

1

05

10

15

20

25

30

% AmDhollne concentration

Fig. 9.12. Effect of Ampholine concentration on minimal focusing time (fMF) of histidyltyrosine [ 741.

of soybean trypsin inhibitor in a standardized polyacrylamide gel system is shown in Fig, 9.12. Such experiments are needed in order to assess objectively the effect of other factors on the f M F of ampholytes and proteins in typical isoelectric focusing systems. Resolution, resolving power, segmental pH gradient and isoelectric point The resolution (i.e., degree of separation) between two adjacent zones in isoelectric focusing can be expressed [ 6 2 ] as

R, = A X / I . ~ ( U A+ U B )

(9.9)

188

ISOELECTRIC FOCUSING

where AT (metres) is the peak separation between two zones A and B with standard deviations uA and U B (metres). A resolution of unity signifies two just-resolved zones [75]. The resolving power [S] is defined as APr =

[dbH)/dxl 06

(9.10)

where ApI is the minimal p l difference for complete resolution of two focused zones and d(pH)/dx is the segmental pH gradient. The latter parameter can be measured by using two p1 markers of closely spaced isoelectric points [60] from (9.1 1) A(~H)/AX= (PZA - P ~ B ) / @ A -ZB) where p l is the isoelectric point, I? is the peak position and subscripts A and B denote two p l markers. The assumption is made that species A and B have reached their isoelectric R A - R B represents points and that A(pH)/Ax is constant between p l and ~ p l ~where , a small segment of the separation path. As X and (iG can be obtained by TRANS-IF, the resolution, segmental pH gradient and resolving power can be obtained directly as a function of time, electric field strength and other factors [62] . The apparent isoelectric point of an ‘unknown’ protein (C) can also be determined by TRANS-IF. If the protein is focused in the region of an estimated ‘segmental pH gradient’ as described above, its p l can be calculated by linear interpolation

PIC

= P ~ A+ (A(PH)/A~)(ZA-Xc)

(9.12)

All three species, A, B and C, should be at pH equilibrium, i.e., at the steady state. Defocusing and refocusing The parameters that can be determined from these stages of the experiment are the apparent diffusion coefficient (0) and p E (Fig. 9.13). If E and d(pH)/dx are known, the electrofocusing coefficient [dUi/d(pH)] could be estimated. D can be determined from defocusing data and pE from refocusing. Experimentally, the peak variance (0); is measured as a function of time (Fig. 9.14). The apparent diffusion coefficient (0) is calculated from the defocusing stage as (9.13) Data from the refocusing stage of the experiment can be analysed by unweighted, nonlinear least-squares curve-fitting methods utilizing the Gauss-Newton algorithm for the equation ’

U&

= (u& - u&) exp (- 2pE t )

+ a%

(9.14)

which provides estimates of pE, u’Gp.= D/pE and D , where u& , a&- and u & ~denote peak variance during refocusing, refocusing steady state and zero focusing time (end of defocusing), respectively. Alternatively, p E can be estimated from In (0; - U & - . / U & ~ ) = - 2pE. t

(9.15)

189

TRANSIENT STATE ISOELECTRIC FOCUSING lsoelectric defocusing

I soelectr i c refocus1ng

x.0

x=Q

I

x

X

2 = ( u i - u z ) exp ( - 2 p ~ t ) u:

D:Ld(r2

+

2 dt

5:

= D/pE

Fig. 9.1 3. Schematic diagram of the concentration proffle of a focused zone during defocusing and refocusing, and the corresponding equations allowing estimation of D and p E [ 741.

:

5

(t-tq)

Time,

sec x I O - ~

Fig. 9.14. Plot of corrected second moment (I&) for histidyltyrosine versus elapsed time during defocusing, refocusing and the refocusing steady state [ 741.

The refocusing steady state variance (u&) is estimated from a graph of u2 versus time (Fig. 9.14) as the statistical mean of all of the points appearing to be at the steady state Subsequently, the parameter log,,(u& - u&-/&,) is plotted against 2t, where t is the duration of refocusing. Least-squares linear regression is utilized to calculate the slope, which is an estimate of -pE/2.303.

190

ISOELECTRIC FOCUSING

Utilizing the defocusing and refocusing data, recent studies [67]have shown that the diffusion coefficient depends on protein zone load in approximately linear fashion. Accordingly, it is necessary to measure D at several zone loads, and then extrapolate to zero load by linear regression techniques. The same applies for the estimation of the parameter p E . Most disturbing is the marked effect of ampholyte concentration on both D and p E . Although the effect is approximately linear (i.e., increasing ampholyte concentration causes a decrease in D and an increase in p E ) there is no a priori reason to extrapolate to ‘zero ampholyte’ concentration to obtain corrected values of D and p E . Furthermore, these values have to be corrected for temperature, viscosity and field strength Q. Reliable measurement of E presents serious problems because of the nonuniform conductance of ampholytes. Another problem arises from the possible nonuniform concentration distribution of ampholytes which may affect the local (around the p l ) viscosity. Because of all of these non-ideal ampholyte effects which affect peak variance directly, by way of influencing D and E , it is doubtful that the predicted resolution from Svensson’s equation [ l ] is ever achieved in practice.

CONCLUSIONS The analytical isoelectric focusing technique offers extremely high resolution for the separation of proteins. It also provides methodological flexibility in terms of equipment, supporting media and detection methods. When coupled to other electrophoretic techniques, such as SDS-polyacrylamide gel electrophoresis, it can be used to display a separation pattern composed of hundreds or even thousands of proteins. Such powerful methods were not available a few years ago. It is hoped that the technology available will augment the solution of biological and biomedical problems in the near future.

REFERENCES 1 H. Svensson, Acta Chem. Scand., 15 (1961) 325. 2 H. Rilbe, Ann. ff. Y.Acad. Sci.,209 (1973) 1 1 . 3 G. H. Weiss, N. Catsimpoolas and D. Rodbard, Arch. Biochem. Eiophys., 163 (1974) 106. 4 H. Svensson, Arch. Biochem. Eiophys., Suppl., 1 (1962) 132. 5 0. Vesterberg and H. Svensson, Acta Chem. Scand., 20 (1 966) 820. 6 H . Svensson, Acta Chem. Scand., 16 (1962) 456. 7 0. Vesterberg, Ann. N. Y. Acad. Sci., 209 (1973) 23. 8 N. Catsimpoolas,Separ. Sci., 5 (1970) 523. 9 N. Catsimpoolas,Separ. Sci., 8 (1973) 71. 10 H. Haglund,Methods Biochern. Anal., 19 (1971) 64. 11 C. W. Wrigley, in A. Niederwieser and G . Pataki (Editors), New Techniques in Amino Acid, Pepiide and Protein Analysis, Ann Arbor Sci. Publ., Ann Arbor, Mich., 1971, p. 291. 12 P. G. Righetti and J. W. Drysdale, J. Chromatogr., 98 (1974) 271. 13 G. R. Finlayson and A. Chrambach,Anal. Biochem., 40 (1971) 292. 14 L. S. Bates and C. W. Degoe, J. Chromatogr., 73 (1972) 296. 15 A. Chrambach, P. Doerr, G. R . Finlayson, L. E. M. Miles, R. Sherins and D. Rodbard, Ann. N. Y. Acad. Sci., 209 (1973) 44. 16 N. Catsimpoolas,Anal. Biochem., 54 (1973) 66.

REFERENCES 17

18 19 20 21 22 23 24 25 26 21 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

191

V. Gasparic and A. Rosengren, in J . P. Arbuthnott and J. A. Beeley (Editors), Isoelectric Focusing, Butterworths, London, 1975, p. 178. Z. L. Awdeh, A. R. Williamson and B. A. Askonas,Nature (London), 219 (1968) 66. N. Catsimpoolas, Anal. Biochem., 26 (1 968) 480. G. Dale and A. L. Latner, Lancer, (1968) 847. J. S. Fawcett, FEBS Lett., 7 (1968) 81. D. H. Leaback and A , C. Rutter, Biochem. Biophys. Res. Commun., 32 (1968) 447. R. F. Riley and M. K. Coleman,J. Lab. Clin. Med., 72 (1968) 714. C. W. Wrigley, J. Chromatogr., 36 (1968) 362. P. Righetti and J . W. Drysdale, Biochim. Biophys. Acta, 236 (1971) 17. 0. Vesterberg, Biochim. Biophys. Acta, 257 (1972) 11. 0. Vesterberg, Sci. Tools, 20 (1973) 22. H. Davies, in J. P. Arbuthnott and J. A. Beeley (Editors), Isoelectric Focusing, Butterworths, London, 1975, p. 97. N. Catsimpoolas, Sci. Tools, 16 (1969) 1. N. Catsirnpoolas, Ann. N. Y . Acad. Sci., 209 (1973) 144. B. J . Radola, Biochim. Biophys. Acta, 194 (1969) 335. B. J. Radola, Ann. N. Y. Acad. Sci., 209 (1973) 127. B. 5. Radola, Biochim. Biophys. Acta, 235 (1973) 412. P. Lundahl and S. H j e r t h , Ann. N. Y. Acad. Sci., 209 (1973) 94. S. Hjerttn, Chromatogr. Rev., 9 (1967) 122. J. Bours, in N. Catsimpoolas (Editor), Isoelectric Focusing, Academic Press, New York, 1976, pp. 209-228. N. Catsimpoolas, Biochim. Biophys. Acta, 27 (1969) 365. K. C. Kenrick and J. Margoles, Anal. Biochem., 33 (1970) 204. P. H. O’Farrell, J. Biol. Chem., 250 (1975) 4007. N. Catsimpoolas, lmmunochemistry, 6 (1969) 501. L. Rotbdl, Clin. Chim. Acta, 29 (1970) 101. L. B. Laurel, Anal. Biochem., 1 0 (1 965) 358. S. Carrel, L. Theilkaes, S. Skvaril and S. Barandun, J. Chromatogr., 45 (1969) 483. H. R. Maurer, DiscElectrophoresis, Walter de Gruyter, New York, 1971, p. 106. M. Jirka and P. Blanicky, Clin. Chim. Acta, 31 (1971) 329. N. Catsimpoolas, Biochim. Biophys. Acta, 175 (1969) 214. N. Catsimpoolas, Clin. Chim. Acta, 23 (1969) 237. N. Catsimpoolas, Clin. a i m . Acta, 27 (1970) 365. G. Mancini, A. 0. Carbonara and J. P. Heremans, Immunochemistry, 2 (1965) 235. P. Grabar and P. Burtin, Immuno-electrophoretic Analysis, Elsevier, New York, 1964. R. Quast, J. Chromatogr., 54 (1971) 405. N. Catsimpoolas and J. Wang, Anal. Biochem., 39 (1971) 141. N. Catsimpoolas, Separ, Sci., 6 (1971) 435. N. Catsimpoolas, Anal. Biochem., 44 (1971) 41 1. N. Catsimpoolas, Anal. Biochem., 44 (1971) 427. N. Catsimpoolas and J. Wang., Anal. Biochem., 44 (1971) 436. N. Catsimpoolas and B. E. Campbell, Anal. Biochem., 46 (1972) 674. N. Catsimpoolas, Ann. N. Y. Acad. Sci., 209 (1973) 65. N. Catsimpoolas, Fed. Proc., Fed. Amcr. SOC.Exp. Biol., 32 (1973) 625. N. Catsirnpoolas, Anal. Biochem., 54 (1973) 66. N. Catsimpoolas, Anal. Biochem., 54 (1973) 79. N. Catsimpoolas, Anal. Biochem., 54 (1973) 88. N. Catsimpoolas and A. L. Griffith, Anal. Biochem., 56 (1973) 100. N. Catsimpoolas, in R. C. Allen and H. R. Maurer (Editors), Electrophoresis and Isoelectric Focusing in Polyacrylamide Gel, Walter d e Gruyter, Berlin, 1974, pp. 174-188. N. Catsimpoolas, B. E. Campbell and A. L. Griffith, Biochim. Biophys. Acta, 351 (1974) 196. G. H. Weiss, N. Catsimpoolas and D. Rodbard, Arch. Biochem. Biophys., 163 (1974) 106.

192

ISOELECTRIC FOCUSING

N. Catsimpoolas, W. W. Yotis, A. L. Criffith and D. Rodbard, Arch. Biochem. Biophys., 163 (1974) 113. 68 W. W. Yotis, N. Catsimpoolas, M. S. Bergdoll and E. J. Schantz, Infect. Immun., 3 (1974) 974. 69 N. Catsimpoolas, in I. P. Arbuthnott and J. A. Beeley (Editors), lsoelectric Focusing, Butterworths, London, 1975, pp. 58-73. 70 N. Catsimpoolas, in E. Grushka (Editor), New Developments in Separation Methods, Marcel Dekker, New York, 1976, pp. 79-100. 71 N. Catsimpoolas, in N. Catsimpoolas (Editor), Methods ofprotein Separation, Vol. 1, Plenum, New York, 1975, pp. 27-67. 72 N. Catsimpoolas, A. L. Griffith, J. M. Williams, A. Chrambach and D. Rodbard, Anal. Biochem., 69 (1975) 372. 73 N. Catsimpoolas, in N. Catsimpoolas (Editor), Isoelectnc Focusing, Academic Press, New York, 1976, pp. 229-258. 74 N. Catsimpoolas, in P. G. Righetti (Editor), Isoelectnc Focusing and Isotachophoresis, Elsevier, Amsterdam, 1975, pp. 79-92. 75 H. Svensson,J. Chromatogr., 25 (1966) 266. 67

Chapter 10

Analytical is0tachophoresis

.

J . VACfK and F M . EVERAERTS

CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The regulation function (“die beharrliche Funktion”) . . . . . . . . . . . . . . . . . . . . . . . Dynamics of the isotachophoretic separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . The steady state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adaptation of concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The terminating electrolyte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adaptation of sample concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Separation of an ionic species with a very small (large) effective mobility . . . . . . . . Influence of impurities in the terminating and the leading electrolyte . . . . . . . . . . Application of spacers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Distribution of potential gradient and temperature . . . . . . . . . . . . . . . . . . . . . . . Effective mobility. actual mobility and absolute mobility . . . . . . . . . . . . . . . . . Influence of ionic strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of the pH of the operational system . . . . . . . . . . . . . . . . . . . . . . . Influence of complexation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of the solvent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of the temperature in the zones . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of electric current. electroendosmosis and hydrodynamic flow . . . . . . . . Quality and quantity in isotachophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Qualitative characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quantitative characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Separation compartment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrode compartments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Detectors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Universal detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Specificdetectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Counter flow of electrolyte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermostating of the equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Current-stabilized power supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operational systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Some examples of anionic and cationic separations . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

194 195 196 197 199 200 201 201 201 201 202 202 202 204 204 204 205 205 205 206 206 207 208 211 213 213 214 215 216 217 217 218 218 218 219 224

194

ANALYTICAL ISOTACHOPHORESIS

INTRODUCTION Even in the nineteenth century papers were published that dealt with some theoretical and practical aspects that are still used today in isotachophoretic separation. From those in which theoretical aspects are discussed, the practical papers can easily be understood. These papers deal with moving boundaries of ionic species in electrophoretic experiments and the concentration distribution of these ionic species [ 1,2], the effect of a hydrodynamic counter flow of electrolyte [3-71, the sharpness of zone boundaries [8] and the relationship between the composition of zones and the electric field applied [9]. In other papers the authors used the theories described to introduce new separation techniques. For instance, Ornstein 1101 and Davis [l 11 described the so-called “disc electrophoresis’’ (p. 33), in which the concentrating effect of the isotachophoretic procedure in the upper gel is made use of, followed by zone electrophoresis in the lower gel. In other papers [12-141 the use of so-called spacers to improve the detection or even to make this detection possible is considered. For protein analyses, sometimes spacers [15] are applied or moving pH gradients [I61 are used. Also, experiments have been described that made use of the isotachophoretic principle. The separation is carried out in classical Tiselius tubes [ 171 in which the “moving boundary method” is carried out with optical means of detection (gradient differences). Sometimes the separation is performed on strips [13, 14,181 and UV detection or autoradiographic detection is used. Sometimes the separation is performed in columns making use of different stabilizing agents, or in U-tubes making use of conductivity or high-frequency [ 191 detection. For the measurement of transport numbers the “moving boundary method” has been used, performed in glass capillaries (I.D. ca. 0.1 mm). For the detection of the boundaries, an optical method has been used, and differences in refractive index have been measured [4,6]. About 50 years ago it was suggested that the migration of zones could be detected by means of thermometric, conductivity and refractive index detectors [20-231. Analytical evaluation [24] is started when the isotachophoretic separation principle was applied in narrow-bore tubes and a sensitive detector was developed that is applicable to analyses in narrow-bore tubes. First a thermometric method for detection [24], then a photometric detector [25] and finally a conductimetric (potentiometric) detector [26] have been used. In some equipment a combination of these detectors is applied. Since 1964 [27], narrow-bore tubes have been widely used in analytical isotachophoresis [28-841 in work dealing with both applications and instrument development. Prior to 1970, several names were used for similar electrophoretic techniques, including the ion migration method [Kendall [85] (1928)], the moving boundary method [MacInnes and Longsworth [86] (1932)], displacement electrophoresis [Martin [87] (1942) and Everaerts [27] (1964)], steady-state stacking [Ornstein [ l o ] and Davis [ l 11 (1964)], cons electrophoresis [Vestermark [I81 (1966)] and ionophoresis [Preetz and Pfeifer [7] (1966)l. Together with Haglund [88], a group of research workers introduced a ‘new name, based on an important phenomenon of this electrophoretic technique, namely the identical velocities of the sample zones in the steady state: iso-tachoelectrophoresis (from the Greek, this means ‘‘equal velocity electrophoresis”), shortened to isotachophoresis. Later, two further names, transphoresis [89] and omegaphoresis were introduced, but they are seldom used.

THEORETICAL

195

t 0

Fig. 10.1. Separation of a mixture of anions according to the isotachophoretic principle. The sample ABC (s) is introduced between the leading anionic species (L) and the terminating anionic species (T). A suitable cationic species is choscn as the buffering counter ion. (a) Original conditions. After some time (b), some mixed zones are obtained according t o the moving boundary principle. Finally (c), all anionic species of the sample are separated and all zones contain only one anionic species of the sample (“ideal case”). (UL)eff > (uA)eff (uB)eff > (Uc)eff (UT)eff.

’

THEORETICAL The principles of isotachophoretic separations have already been discussed. The difference from other electrophoretic separation techniques lies mainly in the experimental performance. An essential condition is the use of two different electrolytes, the so-called leading and terminating electrolytes. The composition of these electrolytes must be chosen such that the following condition is fulfilled:

> (UiIeff> (UT)eff

(U~)eff

(10.1)

The effective mobility, Uerr,is defined by eqn. 1.28, subscripts L, i and T refer to the leading ions, the sample ions and the terminating ions, respectively. The sample is commonly introduced between the leading electrolyte and the terminating electrolyte, or at least at the boundary between them. The leading electrolyte initially fills the separation compartment and the counter electrode compartment. We shall consider the separation of anionic species, because for isotachophoretic

196

ANALYTICAL ISOTACHOPHORESIS

analyses a choice must be made as to whether a separation of anionic or cationic species is sought. When an electric current is passed through a system (see Fig. lO.la), a uniform electric field strength over the sample zone occurs and hence each anionic species in the sample will have a different migration velocity according to eqn. 1.22. The sample anionic species with the highest effective mobility will run first and those with lower effective mobilities will remain. Hence, both in front of and behind the original sample zone, the moving boundary procedure results in two series of mixed zones. In the series of mixed zones, the sample ionic species are arranged in order of decreasing effective mobility (see Fig. 10.lb). The anionic species of the leading electrolyte can never be passed by sample anions, because its effective mobility is chosen so as to be higher. Similarly, the terminating anions can never pass the anionic species of the sample. In this way, the sample zones are sandwiched between the leading and terminating electrolyte. In the mixed zones of the sample (see Fig. lO.lb), the separation continues and, after some time, when the separation is complete, no further changes to the system occur and a steady state has been reached (Fig. 10.1~). The regulation function (“die beharrliche Funktion”) In analytical isotachophoresis, the separations are almost always performed in narrowbore tubes, because no stabilizing media need to be applied. The transport of various substances in an arbitrarily chosen point inside the separation compartment is described by the transport equation (eqn. 1 S9). Because the separation is performed in an inert narrow-bore tube, we can neglect the interaction with the carrier, i.e., the narrow-bore tube itself. Eqn. 1.59 for the i , zth components (if z # 0) can be given as (10.2) In Chapter 1 the sign is explained. If the total concentration of all substances in a given volume of the separation compartment is c = ? Z: ci,z (see p. 2), the transport equation (eqn. 10.2) for all components through that volbge will be (10.3) Because the electroneutrality principle is always valid, this means that 7 Z: Z C ~ = , ~0 , i.e., the right-hand side of eqn. 10.3 is zero. During an electrophoretic ;e;aration, in the separation chamber, zones, and thus zone boundaries, of the species to be separated can always be distinguished. In isotachophoretic analyses, characteristic sharp zone boundaries are obtained and the zones themselves are characterized by constant values of c ~ , K~, , T, p and E for each zone. This holds for zones consisting of pure substances (real zones) and for those consisting of two or more substances (mixed zones). The constant values of the above parameters with respect to the coordinates in the separation chamber means that eqn. 10.3 can only be zero if ? ( Z I C ~ , ~ ( Uis~a ,constant ~ ) - ~ and does not depend on time or on position inside the skparation chamber for all zones. This equation has already been given in Chapter 2 (eqn. 2.2):

5

197

THEORETICAL

11 z

= constant ui,z

Eqn. 2.2 is known as the Kohlrausch regulation function or, as he called it, “die beharrliche Funktion”. This equation is valid for all electrophoretic separation processes, not only for isotachophoresis. This means that both during the separation (moving boundary procedure) and in the steady state in isotachophoretic analyses this equation holds. In the steady state the situation is simple, because only one sample species is present in a zone, assuming that no steady-state mixed zones are present. Let us first consider a boundary between two adjacent zones, containing the anionic species L and A, with UL > U A .The influence of diffusion, etc., will be neglected. Suppose the counter-ionic species Q are similar in both zones and have a constant mobility U Q ,all zones are monovalent and fully ionized and the influence of H+and OH- can be neglected, then from the regulating function (eqn. 2.2) we can write (10.4)

Using the electroneutrality principle ( c ~ =, C~ Q , ~and cA,? = c & , ~we ) can write eqn. 10.4 in the form cA,2

= cL.IcIA(uA

+ uQ)-l(uL+

uQ>(uL)-’

(10.5)

The Kohlrausch regulating function is often used in this form. If the ionic species are not monovalent, the Kohlrausch regulating function must be written in the form as given in Chapter 2 (eqn. 2.4). The complexity of eqn. 10.5 increases if the dissociation of weak acids (and/or bases) is taken into consideration. Dynamics of the isotachophoretic separation The solution of the dynamics of the total separation process, from the start to completion, provides the concentrations of all substances that take part in the separation process as a function of position and time, and thus gives complete information with respect to the separation process. From the results obtained, we can establish the route a pure component travels as a function of time from the original sample, via a stationary-state mixed zone, to its final pure (steady-state) zone. This is not only important from the theoretical point of view, but also has practical value. No great difference exists between the zone boundaries of mixed zones and pure zones, especially if thermometric detection is used (Fig. 10.2). With high-resolution detectors, the “pure” zone boundaries are registered much sharper. The dynamics of the isotachophoretic separation can be obtained, if we solve the system of partial differential equations (Chapter 1) that describe the separation. This has been done on an EAI hybrid computer [9 1,921, the result of which is shown in Fig. 10.3, which was produced by the computer. A schematic diagram showing the isotachophoretic separation of two components, A and B, is given in Fig. 10.4. It is assumed that UL > U, > U, > U,. Fig. 10.4a shows the initial condition, the sample ions A and B (with concentrations A. and Bo) introduced between the leading electrolyte L and the

198

ANALYTICAL ISOTACHOPHORESIS

T

TlUE

Fig. 10.2. Isotachophoretic separation of formate (A) and acetate (B), with@and w i t h o u t a a counter flow of electrolyte. The leading electrolyte was histidine (0.02 mole)-chloride (L) (0.01 mole) ( H 6). The terminating electrolyte was glutamic acid (operational system listed in Table 10.1). A shows the steady state, whereas B shows the stationary state. A thermocouple, mounted on the outside of the PTFE narrow-bore tube, was used as the detector. Qualitative information is found in step heights (temperature differences); quantitative information is obtained from measuring step lengths (distances between the peaks, i.e. dT/dt of zones passing the detector). The difference in baseline in the differential signal is due to the imbalance of the differential thermocouple 1931.

6

6

terminating electrolyte T. The solid line at the top of each section in Fig. 10.4 gives the total concentration of the ionic species present in that zone. The initial concentrations of the sampIe ions (A, B) and the terminating ion (T) are arbitrarily chosen, i.e., not yet adjusted to the leading electrolyte L according to eqn. 10.5. In Fig. 10.4b the situation is shown after a certain time, the leading zone L having migrated over a certain distance but its concentration remaining constant. According to all moving boundary procedures, a zone containing the anionic species A is formed and the concentration in this zone, Al ,is adapted t o zone L. The mixed zone A + B that has passed the original boundary is also adapted. However, behind the original boundary, the original mixture A, B, is present, still not adapted. Behind that zone Ao.+ Bo, a zone B1 is formed that contains only the anionic species B and this zone now is adapted to the zone A. Bo. In Fig. 10.4d, the terminator has passed the original boundary and from this time also a zone T3 exists, already adapted to the leading zone L. At this moment three concentrations of T can be found: a zone T3adapted to the leading zone L, a zone T2 adapted t o the no longer existing zone A. Bo,and the original concentration T1. It is assumed that the volume of the

+

+

+

THEORETICAL

199

X

Fig. 10.3. Dynamics of an isotachophoretic separation process, calculated with the hybrid computer (Charles University, Prague). c = concentration; K = specific conductance; L = leading ion; A , B = sample ions; T = terminating ions.

terminating compartment is infinite in comparison with the volume of the separation compartment. In Fig. 10.4e, the same situation is shown, the mixed zone A + B being much smaller. In Fig. 10.4f, the mixed zone A + B has disappeared, i.e., the anionic species A and B are fully separated. The steady state has been reached and no further changes can be expected. The steady state The steady state in isotachophoresis can be defined as the state in which all ionic species, except the counter ions (s), with a sufficient difference in effective mobility, are separated according to their effective mobilities in different consecutive zones. Between these zones exist so-called separation boundaries, characterized by the presence of a given substance only on one side of this boundary. All separated substances (zones and zone boundaries) migrate with equal speed, v , described by the equation

Eqn. 2.3 is called the "isotachophoretic condition" and is characteristic for isotachophoretic separations. For the isotachophoretic state, further characteristics are the adaptation of the concentration, the stepwise change of the electric field strength, the electric conductivity and the temperature and actual effective mobility of the ionic species present.

ANALYTICAL ISOTACHOPHORESIS

200

ct

ct

ct

ct

"'"":y

.......... ....:...,...,......,,

, Usot-, but at 10-3M concentration Ucl- < Use;-. Influence of the p H of the operational system The dissociation of weak electrolytes varies as a function of the pH of the solution and hence (eqn. 1.29) the effective mobility changes. Two points are very important here: firstly, the choice of the pH itself, and secondly, the choice of the type of buffering counter ion, which defines the pH at which the analysis is to be performed. The most important function of the buffering counter ion is its buffering capacity for the stabilization and regulation of the pH in the different zones. By this means the effective mobilities are fixed and thus the steady state can be maintained. In general, the pH is chosen in such a way that maximal differences in effective mobilities can be obtained according to eqn. 1.29, but a limitation is that if a zone differs by more than 2 pH units from the pK values of the ionic species in that zone, such low effective mobilities can be obtained that the potentials required rise above the maximal potential of the current-stabilized power supply. The pH of the operational system, further, must be chosen such that the counter ion keeps its buffering capacity, not only in the zone of the leading ion but also in the sample zones and the zone of the terminating ion. In anionic separations the pH shows a tendency to increase (and to decrease in cationic separations) from the leading electrolyte zone towards the terminating electrolyte zone [93]. A counter ion will have a low buffering capacity if its pK value differs by more than about 2 pH units from the pH in a zone. As a general rule, we can say that the pH of an operational system (pH,.,) must always be chosen such that (pK + I ) > pH,., > (pK - 1).

Influence of complexation The effective mobility of ionic species can easily be varied by using complexation reactions [96,97]. Even neutral molecules can be analysed by using complexation to give these molecules a net charge. The ionic species to be separated may form irreversible (e.g., Ni2++ EDTA4-+ NiEDTA2-) or reversible complexes. With irreversible complexes, the complexation reaction commonly takes place before the isotachophoretic separation. If reversible complexes are obtained, the complexation agents are commonly used as additives in the electrolytes of the operational system (i.e., leading electrolyte and/or terminating electrolyte). The complexation agents used can be substances that show co-migration, countermigration or no migration.

205

THEORETICAL

Influence of’the solvent One of tlie most important features of a solvent is the solvation of the ionic species dissolved in it. Because the mobility (eqn. 1.24) is not directly proportional t o the hydrodynamic shape of the ionic species involved, but with its dimensions as it is solvated, the influence of the solvent chosen is alniost always far from negligible. The second important role of the solvent is t o decrease the electrostatic interactions between the oppositely charged particles of the ionic substances as a result of its perniittivity. The permittivity also influences tlie ionic equilibria. In solvents with high permittivity and in very dilute solutions. only free ions are present. whereas a t lower permittivity, the equilibrium lies towards the left-hand side, which means molecules with n o charge are present. For these reasons, solvents (or mixtures of solvents) that have a high permittivity can be recommended. In particular, methanol can be recommended as an electrophoretic solvent if water cannot be uscd 1931. Influence o,f the terriperuture in the zoiies If the temperature of an electrolyte increases, the viscosity decreases and thus the mobilities increase (eqn. 1.24). Not only the viscosity, but also the dissociation constants and instability constants are influenced by the temperature. It needs no further explanation that the effects of temperature on the dissociation and instability constants are different for different substances. Further details on this aspect can bc found in the literature [98]. Influence of electric current, electroetirlosrriosis and hytlrodynatnic ,flow Another frlctor that influences the isotachophoretic separation is the electrophoretic driving current. The heat produced is a square function of the electric current applied and is proportional to the electric resistance (eqn. 1.36). Because the terminator zone has the lowest conductivity, the temperature is maxinial. If‘ the difference is too great. a teniperature gradient will result that is not negligible. The temperature gradient coincides with a gradient in effective mobility of the ionic species present. This may diffuse the zone boundary, as is shown schematically in Fig. 10.6. Because a potential gradient is applied, an electroendosmotic flow may be present. This flow conirnonly will also have a negative influence o n the sharpness of the zone boundary. This disturbing effect is obtained especially if high-resolution detectors are used. To reduce this electroendosmotic flow, surface-active compounds are added t o the leading electrolyte (931. A hydrodynamic flow of electrolyte forms a radial velocity gradient in the narrowbore tube (Fig. I .?). This also disturbs the zone profile, especially its sharpness. An unwanted hydrodynamic flow (e.g., due to gas production at the electrodes) is prevented by a good construction of the equipment. Commonly the gradient built up by the

T

c

B

A

L

I:ip. 10.6. Influence of radial temperature gradient on the shape of the zone boundary. (UL)eff (UA)eff

> ((iB)eff > (uC)eff

(uT)eff,

>

206

ANALYTICAL ISOTACHOPHORESIS

hydrodynamic flow is directed opposite to that built up by the electroendosmosis and the temperature. This means that the influence on the sharpness of the zone boundaries may be positively influenced if a counter flow of electrolyte is present [93].

QUALITY AND QUANTITY IN ISOTACHOPHORESIS Qualitative characteristics The basic qualitative characteristic of the ionic species, which can be deduced from the signals of universal detectors, is a complex numerical value of the absolute mobility of the ionic species, separated in consecutive zones according to their effective mobilities. The mathematical model can be used for a computer program [93] for the calculation of the effective mobilities. With the calculated effective mobilities, the signals obtained from universal detectors can be evaluated. In practice, first a series of relative qualitative characteristics, in a well defined operational system, need to be collected. These values are proportional to the effective mobilities (or proportional to values related to these effective mobilities). Commonly the conductivity of zones (Fig. 10.7a) is measured, as soon as they pass the sensing electrodes of the conductivity probe. Because of their step-wise character, we call these value step heights (h). Instead of the absolute step height, h , relative step heights, h,., are almost always used (eqn. 3.13, p. 41); h,.

= (hi --h,)(h,

-hJl

(3.13)

Analogously, the signals derived from specific photometric detectors can be handled and transformed into “photometric relative step heights” [99] (fA), as shown in Fig. 10.7b, given by the equation f,l

=

(6-fL)(.f, - f X

(10.6)

Analogous to the definition of hrel,f;:,fl and .f, refer to signals obtained with a photometric device and refer to zones of sample ionic species (i), leading ion (L) and reference species (s) passing the detector. As reference species almost always a strong UV-absorbing substance must be chosen, especially if the leading electrolyte shows no or only slight UV absorption. The main difference between hRl andfEl is that hRl always is positive, whereasf,,] may have both a positive and a negative value“. For the characterization of a complex mixture of UV-absorbing substances, migrating in a mixed zone (equal effective mobilities), it can be advantageous to use spectrophotometric detection 1991. Particularly if the zones are so long that the total spectrum of the substances in the mixed zone can be measured, the qualitative information is much more reliable.

*hrel will only have a negative value if, owing to a pH shift, an ionic species with a higher effective mobility cannot pass the ionic species in front of it, with a lower effective mobility. This is found especially if ti+is applied as a terminator ion (operational system for cationic separation at low pH).

QUALITY AND QUANTITY IN ISOTACHOPHORESIS

207

a

Fig. 10.7. Isotachophoretic separation of a mixture of cations in the operational system listed in Table 10.2. A conductimetric (linear and differential traces) (a) and a U V absorption photometric detector (b) were used. Qualitative information is derived from measured step heights in R and A (hi and fi); quantitative information is obtained from measured distances between peaks, A R (L2). I = K (lead5 = Pb"; 6 = CSH,N+ Anf

(1 1.3)

233

GENERAL ASPECTS

Introducing the angle of deflection, a d , into eqn. 1 1.3, this condition can be written as B(tan

ad,

-tan ad,?) - (al + a z )

A >-

(1 1.4)

nf

An analysis of eqns. 11.3 and 11.4 shows that it is desirable to obtain the greatest difference in deflections at the given length of the apparatus. It follows from eqns. 1 1.1 and 11.2 that both velocity components can be affected. The rate of electrophoretic migration can be increased by increasing the current and the intensity of the electric field. Within certain limits, the same effect is obtained if the flow-rate of the electrolyte is decreased. However, the values of a l and a2 increase in both instances [25]. The result is therefore a compromise between minimal flow-rate of electrolyte and maximal current or electric field intensity in the separation cell. Cells represented by a cylindrical annulus with radial orientation [14, 19,20,26] of the electric field vector are characterized by eqn. 11 .l in differential form: the momentary orientation of the resulting velocity gradient is tan a d . The magnitude of the velocity vector of electrophoretic migration in this instance depends on the position of the particle, i.e., on its distance from the inner wall of the annulus in a radial direction. Making the simplification that the physical properties of the electrolyte are constant, the velocity of the radial motion of a particle of substance 1, drl/dt, is given exclusively by the product of its mobility, U , ,and the potential gradient (intensity of electric field), dU/dr, where r characterizes the distance between the point considered and the cylinder axis. The values of r can vary between r A (outer radius) and rg (inner radius of the cylindrical annulus of the liquid). The differential equation describing the velocity is (1 1.5) Integrating [13,14] the potential gradient dU/dr across the annulus diameter on which the terminal voltage UAB from r B to r A has been imparted, we obtain (1 1.6)

The radial velocity, i.e., the rate of electrophoretic migration, can be expressed by (1 1.7)

Therefore, (1 1.8) where vf is the linear flow-rate of the carrier electrolyte along the axis of the cylinder. An analysis of eqn. 11.8 shows that tan ad in this instance varies inversely with the radius, r. With increasing deflection of the particle towards the outer radius (i.e., when the sample is applied close to the inner radius), the velocity of the electrophoretic motion decreases and the angle of deflection also decreases gradually. The resulting path travelled by the particle is represented by the surface of a paraboloid cap. Fig. 11.3 shows a segment of

234

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

1

E

Fig. 11.3. Sectional diagram of annular arrangement of cylindrical chamber for continuous zone electrophoresis according to Verrneulen et al. [ 141. rA, Outer wall of the annular chamber; rg, inner wall of the annular chamber;rb, radius of sample feed ring;r, and r l , points of outflow of the migrants 1 and 2; OE, outer electrode; CE, central inner electrode; DP, ceramic diaphragm for separation of electrode electrolyte-coolant and of carrier electrolyte; vf, velocity vector of carrier electrolyte flow;EE, coolant (electrode electrolyte); r, radius (the arrow shows the direction of the electrophoretic migration); drldt, velocity vector of electrophoretic migration (for further details, see text).

the cylindrical annulus, the individual force components and the resulting path of the particle. If the sample is applied to an annulus of radius 1 6 , then at this single point the ratio of tan ad, to tan (Yd,2 for two substances [ 121, differingin electrophoretic mobilities depends exclusively on the ratio of these mobilities: tanad _ -U1 (1 1.9) tan ad,2 r/, Both substances subsequently arrive in an electric field of varying intensity. Vermeulen and co-workers [13,14] integrated eqn. 11.7 and obtained eqn. 1 1 .lo, defining the

-*-

ELECTROPHORETIC CELLS OF OTHER SHAPES

235

distance between particles of substances 1 and 2 at time t after application of the sample: (1 1.lo)

ELECTROPHORETIC CELLS OF OTHER SHAPES Separations carried out in serpentine cells according to Kolin and Cox [ 121 are governed in principle by equations derived for the rectangular cell. An entirely specific motion is that of the particle in the annulus made to rotate by the radial component of the magnetic field under the simultaneous influence of the electric field [27,28], acting along the cylinder axis, and of the flow of carrier electrolyte oriented in the same direction. The particles travel in Kolin’s instrument along a helix whose pitch in the direction of the longitudinal axis of the cylinder is directly proportional to the electrophoretic mobility. A detailed definition of the path of the particle in an electromagnetically rotating annulus is given in Chapter 12. The individual types of arrangement of flow-through electrophoretic cells are summarized in Table 11.l. The relationships between vectors of electrophoretic migration and gravitational force are given in Table 1 1.2.

FACTORS AFFECTING ZONE WIDTH IN FLOW-THROUGH CELLS Zone broadening in flow-through cells is caused especially by the following four factors: (1) Poisseuille velocity profile of the liquid curtain, (2) velocity profile caused by electroendoosmosis at the walls of the electrophoretic chamber, (3) temperature gradient in the liquid curtain and (4) temperature diffusion. The first factor was studied by Hannig [52], who suggested that part of this hydrodynamic could be compensated by electroosmotic flow. The factors were studied from the mathematical point of view by Kolin [IS] and Strickler and Sacks [ 161. These phenomena were examined experimentally in thin-layer rectangular cells by Hannig et al. [29] and in a cylindrical annulus by Vermeulen et al. [ 141, Aitchison et al. [ 191 and Mattock et al. [20]. Poisseuille velocity profile of the liquid curtain The time, t , for which the substances separated and the carrier electrolyte are exposed to the electric field is given by the ratio of the effective cell length, B (see Fig. 11.1 and Table 11.l), to the velocity of the liquid curtain, v f :

B

t = --

(1 1.1 1)

Vf

A symmetrical Poisseuille velocity profile is formed during the flow of the liquid through the rectangular cell. Hence

radial

Orientation of vector of imparted magnetic field

rotary

parallel to longer axis of cylinder

serpentine; effective length B

parallel to B axis of A X B walls

Resulting direction of carrier electrolyte flow

deflected at angle 'Yd fIom longitudinal axis B of serpentine surface of cap of rotational paraboloid (or coil tracing this surface) helical

ffd

deflected at angle from direction parallel to B axis

Resulting direction of zone motion

O F FORCE COMPONENTS AFFECTING THE FINAL ZONE MOTION IN FUNDAMENTAL TYPES OF Direction of vector of imparted electric field paralllel to A axis of A X B walls

ell parallel to A axis of A X B walls

radial

parallel to longitudinal axis of cylinder

h)

-w v\

0

2

2 2

s8 crl

0

r

Lc

z

?s X

C 0

E

2 s 0 30!

E

R

References (selected examples)

parallel

perpendicular parallel parallel

13,14,26 15,28

31,SO

5,6

9-1 1

perpendicular perpendicular (effect of magnetic field -helix)

% v m

n

2z

x z

-I

CI

r

t;

r

m

c)

X

n

0 C

ga

4

2

12

X

19,20

21,22

7, 8, 36,40,41, 49,51

2, 48, 49

rectangular cell rectangular cell rectangular cell packed with gritted material paper, filter cardboard of rectangular shape cylindrical annulus cylindrical annulus or endless belt

ATION OF FLOW-THROUGH CELLS IN SEPARATION COMPARTMENT Separation compartment Orientation of velocity vector of electrophoretic migration with respect to direction of gravitational force

nt

changing from perpendicular to parallel perpendicular or parallel

perpendicular

not defined

rectangular serpentine-shaped Cell rectangular cell

rectangular cell

cylindrical annulus

4

w

t 4

238

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

B

or

(1 1.12)

-

1

(1 1.13)

where V f is the mean velocity at which the liquid curtain moves. The value of 5f is calculated from the ratio of the flow-rate of buffer through the cell to the effective cell volume, i i s the mean residence time of the particle in the chamber and d is a variable distance between the centre of the liquid curtain and the walls of the electrophoretic cell A x B . The distance between walls A x B is equal to the total thickness, C,of the liquid curtain. Particles with increasing distance d from the centre to the margin of the curtain are exposed to the electric field for a longer period and are therefore more deflected.

Effect of electroosmotic transmission The electroosmotic profile depends on the intensity of the electric field, E , and the magnitude of the electroosmotic mobility, Uos.The profile of the electroosmotic velocity, vo,, in a rectangular cell is given by v,,

=

(

-$Uo . E * 1--

(1 1.14)

It follows from the studies made by Strickler and Sacks [18] and Hannig et al. [29] that optimal resolution, i.e., maximal suppression of effects causing zone broadening, occurs if the zone potential of the inner walls is identical with that of the material used for the separation. The zeta potential of the walls can be adjusted by appropriate treatment of the walls and by proper selection of the material. Strickler and Sacks [ 171 divided the separation chamber into two compartments. The compartment that is nearer to the sample inlet has a marked electroosmotic profile, whereas that which is nearer t o the outlet has a low zeta potential with a prevailing effect of the Poisseuille profile. Optimal compensation is achieved by an independent change of electric field intensity in both compartments of the chamber. Hannig et al. [29] prefer a stable adjustment of the zeta potential of the walls because of the long-term character of the separation. The zeta potential of the silicate walls of the chamber is adjusted best by treating the walls with a 1% solution of (3-aminopropy1)triethoxysilane [30]; the originally negatively charged surface is thus converted into a surface that is electrically neutral or even positively charged. Separations of living cells can be carried out to advantage [29] in chambers with walls coated with albumin. The effect of zeta potential on the electrophoretic distribution profile is shown in Fig. 11.4.

239

FACTORS AFFECTING ZONE WIDTH IN FLOW-THROUGH CELLS 30

b

C

d

Fig. 11.4. Influence of zeta potential on electrophoretic distribution profiles of (a) albumin and (b) rat erythrocytesina thin-layer rectangular electrophoretic cell [29]. Potential gradient E = 1 X 1 0 4 V m - ' . a, Ultrasil glass; b, silicate glass; c, silicate glass pre-treated with 5% albumin solution; d, silicate glass pre-treated with 1% solution of (3-aminopropyl)triethoxysilane; e, acrylic glass. (a) Distribution of albumin in Tris-borate buffer (0.3M), conductivity K = 1.1 1 0 - 2 S m - ' ,liquid curtain velocity 2.35 lO-'m s-'. 5"C, sample dose 0.4 mlh-' of 1%albumin. T,,,,transmittance at wavelength 225 nm. T,,,, transmittance at 260 nm. (b) Distribution of rat erythrocytes in triethanolarnineacetate buffer (0.0212M triethanolamine, 0.26Mglycine, adjusted to pH 7.2 with acetic acid). K = 8 Sm-'.

-

Temperature gradient in liquid curtain The parabolic profile of the temperature gradient in the liquid curtain causes a change in the actual electrophoretic mobility, Ui,=,as viscosity decreases with the decreasing value of d , i.e., towards the centre of the liquid curtain. The relative temperature coefficient of viscosity, a,, ,is approximately 0.03 (in aqueous media). The relation [29] determining the actual mobility of the particle with respect to maximal particle mobility,

(11.15)

involves a factor k,which depends on the magnitude of the total electric power,p, converted into heat in a chamber whose height is B and width A , with a liquid curtain of thickness C: k(Wm-') G 55*a,,*@+--P (1 1.16) A *B The construction of the cell and the material used have a marked effect on the resulting temperature profile. Likewise, the efficiency of the removal of Joule heat by a

240

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

TABLE 11.3 MODES OF REMOVAL OF JOULE HEAT FROM FLOW-THROUGH CELLS Method of removal of

Medium

pre-cooled carrier electrolyte

liquid

cooling medium in contact with outer walls of electrophoretic cell

Cell type (see Table 11.1)

References (selected examoles)

temperature increase of carrier electrolyte

1, 3

19

gas

air Freon

1 3

8,36

liquid

water electrode electrolyte

1,4 3

I,4 0

Peltier thermoelements

1

24

solid body

14

Commercial type FFS.

heat-transfer medium has a considerable effect on the final temperature profile. To prevent the formation of the temperature profile in certain instruments, the cell walls are not cooled from the outside and the Joule heat is cumulated in the rapidly flowing precooled carrier electrolyte. A summary of procedures used for the removal of Joule heat from flow-through cells is given in Table 11.3. The shape of the temperature gradient in the cylindrical annulus and in the cylinder has been calculated by Porath [31]. The relationship derived experimentally by Porath [3 11 for a cylinder and a temperature difference at a distance r from the cylinder axis, where the highest temperature is T,,, ,is in simplified form a quadratic function of the radius, r , and a linear function of the Joule heat produced. Peniston et al. [32] arrived at similar conclusions. A relationship expressing the relative change of velocity of electrophoretic motion as a function of the Joule heat produced in a cylindrical cell has been derived by HjertCn [33].

Temperature diffusion The influence of temperature diffusion on modern carrier-free methods is relatively small. This influence can often be neglected when rapidly sedimenting particles and proteins are separated by zonal electrophoresis. The residence time of these instruments is from several tens of hundreds of seconds. The effect of temperature diffusion cannot be disregarded if low-molecular-weight compounds are separated under non-steady-state conditions. This holds especially for flow-through instruments using carriers in which the residence time is of the order of 103-104s. The residence time in flow-through cells cannot be prolonged without limit because of temperature diffusion and also because of the decreasing stabilizing effect of the too slowly flowing electrolyte (see Fig. 11 .S). For these reasons, the effect of temperature diffusion is usually masked by zone broadening caused by thermal convection.

24 1

FACTORS AFFECTING ZONE WIDTH IN FLOW-THROUGH CELLS C CmmJ 1.0r

Fig. 11.5. Approximate limiting conditions of liquid curtain stability in free-flow electrophoresis according to Hannig 1251. Deduced from experiments carried out at a potential gradient of E = 1 l o 4 V m - l by means of carrier electrolyte with specific conductivity K = 9.5 S m" ; dependence between liquid curtain velocity, vf,and cell thickness, C. Area to the left of the lines, unstable; area to the right of the lines, sufficient stability. One-sided cooling.

-

-

Application of sample

The final shape of the zones obtained under non-steady-state conditions, i.e., by zonal electrophoresis, is considerably affected by the manner in which the sample is applied. It follows from an analysis of the Poisseuille velocity profile that the negative effect of the velocity profile is suppressed best if the volume of the sample injected occupies only the middle part of the layer, i.e., if d = 0 . If the cross-section of the solution injected is circular, the diameter being A o , the zone width increases owing to the velocity profile at the outlet of the flow-through cell according to eqn. 11.17, even if temperature diffusion is negligible. If the path travelled by an idealized particle in the direction of electrophoretic migration is s,, then the resulting path assumes the value s, according to the following equation [29]:

s,

=

+"&$

s -2c

(11.17)

The initial zone width depends on the ratio of the linear sample flow velocity, v,,to the mean velocity at which the liquid curtain moves, i j f . The zone width, A , , at the point of injection is given by 7

A o ( m ) = 0.0123-C

J J 32ijf 1-

1 --

(1 1.18)

This equation does not hold for large A . values. Preparative applications [34] often require a larger A . and thus its circular shape becomes rectangular. The effect of the velocity profile is naturally more marked. The theoretical magnitude ofA, can also be

242

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

increased by the geometry and orientation of the injection hole of the sample applicator. The shape of the applicator has been examined by Kolin and Luner [28] and is especially significant in instruments for analytical continuous zonal electrophoresis [35]. Collection of fractions The manner in which fractions are collected in instruments designed for preparative operations, which represent the most important application of continuous electrophoresis, has a considerable effect on the resolving power of the apparatus. In addition to passive fraction collection, used in instruments in which the flow-rate and collection are controlled by the soaking effect of paper wicks and by capillary tubes sucking off fractions according to differences in hydrostatic pressure, active methods of fraction collection have also been developed. The fundamental prerequisite is stability of the flow in individual channels with respect to time and site. Whereas at high flow-rates the outflow can be regulated successfully by hydrodynamic resistance, at low flow-rates, and especially in carrier-free, thin-layer rectangular instruments, the liquid must be sucked off with a vacuum-operated device automatically controlling [8,36] the pressure in the individual channels of the collector. Multi-channel peristaltic pumps with low pulsation have been used to advantage in instruments used for the separation of rapidly sedimenting particles. If only one component is to be separated, a special apparatus with a moving collection channel can be used; its position can be changed with respect to the deflection of the component to be collected. Carrier types of flow-through electrophoresis

Only a few of the types [3,37,38] described above are used in practice. One of these is a rectangular-cell continuous apparatus using as the separation medium filter cardboard located vertically in a chamber saturated with vapour. As in analogous instruments [24] with suspended paper, saturation with the electrolyte from above maintains a uniform flow. The cardboard is not in contact with a cooling plate and the Joule heat is removed with a stream of cooled air. The losses of the solvent due to evaporation on the cardborad surface are compensated for by saturation secured by another cardboard sheet, the temperature of which is ca. 5°C higher than that of the carrier. The capacity of this type is about 20 times greater than that of the older versions [37] of continuous paper electrophoreses. The sharpness of the separation and the stability of the angle of deflection are satisfactory and do not vary over several days. These types of instruments permit up t o 8 g of protein hydrolysate of 40 ml of serum to be separated daily, whereas the capacity of simple uncooled types is limited by the low electrolyte flow-rate through the paper and by a maximal output of 20-40 W of Joule heat removed passively via the moist chamber. The principles of the carrier version of continuous electrophoresis are shown in Fig. 11.1. As an example of circular carrier cells, the high-capacity apparatus of Vermeulen et al. [14] will be considered. This apparatus is a cylindrical annulus. The electrode compartment walls, i.e., the proximate walls of the annulus, are made of ceramic, semipermeable material. The cell is refrigerated from both the outside and inside as a

FACTORS AFFECTING ZONE WIDTH IN FLOW-THROUGH CELLS

243

considerable amount of Joule heat (up to 2 kW) is formed at 25-IOOV and 50 A between the electrodes, which are close to each other. The annulus is packed either with glass beads (particle diameter ca. lo-' mm) or cross-linked polystyrene in similar form. When the sample is circularly applied on to the packed layer, placed in a cell of total length ca. 1.2 m and outer diameter ca. 0.25 m, the capacity of separation of simple mixtures of amino acids which have major differences in electrophoretic mobility is 100-1 000 g/h, according to Vermeulen et al. The arrangement of the cell is shown schematically in Fig. 1 1.3. The resolving power of this type of apparatus is negatively affected by the long time required for the separation and by the high electroosmotic transfer through the semipermeable membranes and in the carrier layer. Large volumes of oxygen and hydrogen form at the electrodes as a result of electrolysis. The apparatus must therefore be equipped with compartments packed with Raschig rings, where the gases are liberated from the electrolyte and hydrogen diluted with air beyond the explosive level. The anodic and cathodic electrolytes are mixed and the changes in pH caused by electrolysis are thus compensated. The electrolyte is subsequently cooled and pumped to the electrode compartments. Carrier-free pow-through apparatus for continuous zonal electrophoresis Apart from thck-layer highcapacity rectangular cells with a small number of collecting channels and a small angle of deflection [ S , 61, which were constructed as the first carrier-free cells, the widest application to date is enjoyed by thin-layer flat-bed square cells. Rectangular cells, oriented vertically and designed for separations of rapidly sedimenting particles, and square-shaped cells arranged vertically or horizontally [8,36, 39-41], which are suitable for separations of soluble substances, are still being developed. Some of these types of cells are cooled from both sides. The layer thickness in such apparatus is ca. lo-' mm. Higher separating capacities are given by cells of greater thickness, but at the cost of the resolving power. There is an empirical relationship between the cell thckness and the minimal flow-rate of the electrolyte [25]. The lower limit of the flow-rate as a function of cell thickness under conditions of zonal electrophoresis is shown in Fig. 11.5. The most common thickness is 0.4-0.7 mm. Fig. 11.6 shows a commercial apparatus (VaPS) of the vertical type with one-sided cooling designed for the separation of cells and of other rapidly sedimenting particles. The apparatus is equipped with 90 collecting channels; collection is effected by a peristaltic pump. The electrode vessels are separated from the separation compartment by ionexchange membranes. The membrane placed at the anodic side must be a cationexchange membrane and an anionexchange membrane must separate the cathodic side. If the poles are interchanged, the resistance of the membranes increase periodically and can be damaged by the large potential difference. The separation capacity is of the order of lo8 erythrocyte cells per hour. Discrete cell zones can be obtained in electrophoretic chambers with silicate walls if the zeta potential of the latter is properly adjusted [16, 171 to the value nearest to the zeta potential of the cells, e.g., by coating with albumin or silane. Whereas the usual conductivity of carrier electrolytes employed in zonal electrophoresis is about 0.1 S m-l, a better separation of the cells is sometimes obtained when the ionic strength is decreased to the limit [42]

244

CONTINUOUS FLOW-THROUGH ELECTROPHORESI

Fig. 11.6. Elphor VaPS flow-through apparatus in a vertical arrangement suitable for separation of sedimenting particles 1291. Thin layer of carrier buffer with gap width C = 7 lO-'m, height B = 0.5 m, width A = 0.1 m, cooling capability 200 W at 5"C, maximum voltage 1300 V. fractionating pump 90 channels, peristaltic system, regulated current range 0.05-0.3 A, residence time 240-2400 s, temperature range 4-25°C. This apparatus has a similar arrangement to the commercial FF5 type (Labomed, Munich).

where the integrity of the cell walls of the particles separated is still intact. Thus, for example, erythrocyte ghosts and outside vesicles [29] were separated in electrolytes of conductivity 9 x lo4 S me'. A marked influence of the shape of particles with similar specific charge is observed at low ionic strengths. The potential gradient used in vertical apparatus is about 1.2 104Vm-'. As the width of such apparatus is ca. 10-'m, the required voltage is lower than that necessary for square-shaped apparatus, All-glass square-shaped cells have been used for carrier-free electrophoresis since 1958

-

FACTORS AFFECTING ZONE WIDTH IN FLOWTHROUGH CELLS

245

Fig. 11.7. Flow-through carrier-free continuous electrophoresis in a horizontal arrangement suitable for separation of solutes (according to Prusfk and Stdphnek) [ 3 6 , 391.Thin layer of carrier buffer with gap width C = 5 * 10-4m. length B = 0.44 m, width A = 0.5 m, cooling capability 950 W at 0°C. maximum voltage 4000 V,maximum current 0 A, vacuum fractionating system 48 channels, residence time 600-4000 s, temperature range - 3 to + 25"C, continuous pneumatic temperatureregulating system, carrier buffer pumped by six-channel allglass pump. This type was produced in the Institute of Organic Chemistry and Biochemistry, Czechoslovak Academy of Sciences, Prague.

The dimensions almost exclusively used are 0.5 x 0.5 m, the layer thickness being 0.0004-0.0006m. The number of fractions collected varies between 48 and 100. This type of apparatus, wluch was designed for the separation of medium- to low-molecularweight substances and also for high-molecular-weight substances of appropriately high electrophoretic mobility, are operated at a potential gradient of up to 6000 Vm" .The refrigeration system must be efficient as heat of the order of 400 and 1000W must be removed from instruments with one- and two-sided cooling, respectively. The procedures for the removal of Joule heat are summarized in Table 1 1.3. The capacity of this type of apparatuses depends on the difference in angles of deflection and on the complexity of the mixture. A capacity commonly achieved is 100-200mg h-' of a mixture of lowmolecular-weight substances, e.g., of peptide mixtures, provided that the resolving power of the apparatus is adequately high. These types are used for the preparation of standards of fine chemicals [24,43]. A detailed view of the apparatus is shown in Fig. 1 1.7. The separation of basic polypeptides [43] is shown by way of example in Fig. 11.8. The flow-rates in these apparatuses are lower than in vertical rectangular apparatuses. The flow time is of the order of 10' to 103s.

246

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

F r o c t . NO.

Fig. 11.8. Separation of Arg-vasopressin from pituitary extract by continuous flow-through electrophoresis according to Prusfk et al. [43]. Pyridine (0.08M)-acetate carrier buffer (pH 5.6), K = 0.195 S m-I, 1 ml h-' of gel fdtrate fraction NE5 (4.7%)applied in the position between fractions 24 and 25, residence time 2910s, flow velocityvf= 1.48 lO"ms-', currentZ= 0.192A,potential gradient E = 4500 V m-I ,E,,, = extinction at 720 nm after Folin-Lowry reaction, left peak = Arg-vasopressin.

The analytical version [22,35,44] of the thin-layer apparatus has been used to demonstrate the separation power of carrier-free electrophoresis in a low-gravity environment during the space fight. The marked improvement in the resolving power [21,22, 3.51 and the substantial increase in the separation capacity indicate the potential of continuous carrier-free zonal electrophoresis. This makes the method promising from the viewpoint of separation of clinically important proteins and cell suspensions under conditions of weightlessness. Analytical cells can be operated without a fraction collector. The zones of the substances separated are detected by measurement of W absorbance at 225 nm. The dispersed light ray, having passed through the quartz walls of the cell, is evaluated in a multiplier and the signal is manipulated electronically, i.e., the blank value of the signal is subtracted and the record is processed in the control computer. In calculating electrophoretic mobility in a cell with layer thickness 0.00035 m, effective length B = 0.18 m and interelectrode distance A = 0.03 m, a relatively high flow-rate (vf = 0.0017-0.006 m s-l) must be considered. The residence time, t , of the zone with respect to the Hagen-Poisseuille law is expressed by t = 2/3 ( V ~ / W )

(I 1.19)

where Vk is the volume of the separation compartment and w the volume flow-rate of electrolyte through the chamber. Orientation of the ionexchange membranes parallel to the plane of the chamber walls permits the apparatus to be operated at extremely high potential gradients (14000 Vm-I). The residence time varies between 20 and 70 s. The instrument is suitable for rapid clinical assays of serum proteins. Velocity gradient electrophoresis

A novel annular electrophoretic separator which can be operated without refrigeration has been developed [19,20]. The stabilization of a laminar flow against thermoconvective turbulence is maintained by carrying out the electrophoresis in an annulus 0.005 m wide, bounded on its inner diameter by a stationary cylinder (stator) and on its outer wall by a cylinder rotating at constant speed (ca. 150 r.p.m.).

FACTORS AFFECTING ZONE WIDTH IN FLOW-THROUGH CELLS

247

A gradient of angular velocity, w , is thus imparted to the carrier solution flowing upwards through the annulus (ca. 7-25 ml s-’). The stabilizing effect is due to inertial forces and is an example of the principle of the Taylor-Proudman [45,46] theorem: “All steady slow motions in a rotating inviscid fluid are necessarily two-dimensional”. Raleigh’s criterion [47] that a stratification of angular momentum about an axis is stable if, and only if, it increases monotonically outwards: (1 1.20)

(where r z u is the angular momentum) is here fulfilled by rotation of the outer wall of the annulus. The arrangement is slightly similar to Vermeulen’s annular apparatus. The sample solution is introduced into the annulus as a fine stream through a narrow circumferential slit near the bottom of the stator and is carried up as a thin curtain travelling next to the stator wall. The surfaces of the rotor and stator above the migrant slit are semipermeable and separate the annulus from the electrode compartments. A continuous flow of electrolyte thus removes the electrolysis products generated at the stainless-steel electrodes without disturbing the laminar flow in the annulus. The components of the migrant exposed to the radial electric field move in a radial direction away from the stator wall towards the rotor. At the same time, they are carried spirally upwards through the annulus by a combination of the carried flow and rotation. At the top of the electrode section (0.3 m long), each migrant species occupies a discrete zone around the whole of the annulus and a cross-section of the annulus represents the migrant species as a series of concentric rings. In order to collect these migrants from the annulus, a series of 30 discs are positioned on the stator as a stack above the electrode section. Each disc has a number of channels cut into its surface connecting with a groove that passes round the circumference of the annulus. The lowest plate thus removes a layer of electrolyte nearest to the stator, the next plate removes the layer nearest to the first plate, and so on. Each disc is connected to a separate outlet and continuous collection of 30 fractions across the annulus is thus obtained. Fig. 1 1.9 summarizes the principle of operation and Fig. 1 1.I0 shows an actual working prototype. The outer stainless-steel electrode is clearly shown, being in contact with the power supply through carbon brushes. The separator is mounted on the drive shaft of a variable-speed motor. Individual off-takes are on the top of the stator. All solutions are fed through a cooling tank before electrophoresis to reduce the temperature t o ca. 2 O C and then introduced at the top of the stator. A temperature increase of up to 20°C occurs as a result of Joule heating during the residence time through the whole annulus. This time is usually very short (ca. 30-60 s). Immediately after collection, the temperature can be reduced. Even with labile compounds recoveries are virtually quantitative. The resolution of the separator is demonstrated on bovine serum albumin distribution in Fig. 11.1 1 . The optimal rotation rate is between 100 and 200 r.p.m. (Fig. 11-12). Sample loading is a compromise between throughput and resolution (Fig. 11.1 3). It has been found that satisfactory results are obtained with 0.1-0.2 g of albumin in 60 min. The location of the migrant in the annulus is shown in Fig. 11.1 1 as the relationship between the distances across the annulus to which the albumin migrates versus the standard deviation of the distribution.

248

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

Fig. 11.9. Diagrammatic representation of the velocity gradient stabilized annular separator according to Mattock et al. 1201 and Aitchison et al. 1191.

Fig. 11.10. View of the electrophoretic annular separator showing the rotor and stainless-steel electrode according to Mattock et al. [20] and Aitchison et al. [19].

Also in this type of electrophoresis the essentially parabolic velocity profile of flow influences the width of distribution. Assuming a constant specific conductivity, K ,a will migrate at a current I ( A ) and move se,, (m) towards species of actual mobility U,,, the rotor according to the relationship K

(1 1.21)

where w (rn’s-l) is the carrier electrolyte volume flow-rate. During the radial movement from the stator to the rotor the migrant moves into a regime of increasing axial velocity. The leading edge of the zone moves axially more quickly than the trailing edge remaining shorter in the electric field. It is therefore electrophoresed radially for a shorter distance than the trailing edge, the effect of which is to reduce the width of the migrant zone. Past the centre of annulus the opposite effect occurs and the zone gradually increases in width. An example of the separation achieved by this equipment is shown in Fig. 11.14.

249

FACTORS AFFECTING ZONE WIDTH IN FLOW-THROUGH CELLS

Fig. 11.1 1. Relationship between the distance of migration across the annulus and the distribution of bovine serum albumin (presented as standard deviation in off-take units) for two sample loads and comparison with the theoretical distribution in the annular separator according to Mattock et al. [ 201 and Aitchison et al. [ 191.

I

6

R o t o r speed w

g Alburninlrnin

Fig. 11.12. Effect of rotation on the distribution of bovine serum albumin (presented as standard deviation in off-take units) during electrophoresis in the annular separator [ 19, 201. Peak width, standard deviation; rotor speed, w (rotations per 60 s ) . Fig. 11.13. Effect of sample loading on the distribution of bovine serum albumin (presented as standard deviation in off-take units) in an annular separator according to Aitchison et al. [ 191 and Mattock et al. [20].

250

CONTINUOUS FLOW-THROUGH ELECTROPHORESIS

3 Fraction outlet n a

Fig, 11.14.Fractionation of desalted human plasma at a throughput of 0.25 g of protein in 60 s in Tris-citrate buffer (pH 8.5), conductivity K = 0.1 S m-', showing the distributions of immunoglobulin (IgC), transferrin and albumin; annular separator according to Mattock et al. [ 201 and Aitchison e t al. [ 191.

Human plasma was electrophoresed at pH 8.5 and the throughput was maintained at 0.25 g in 60 s. Recoveries of almost all of the components were quantitative. When using the separation equipment in large-scale industrial applications, for the preparation of pharmaceutical materials, etc., there is no great problem in running the system continuously under sterile conditions for several tens of hours.

REFERENCES 1 2 3 4 5 6 7 8 9 10 11

12 13 14 15 16 17

J. S. Philpot, Trans. Faraduy SOC.,36 (1940)38. H. Svensson and 1. Brattsten,Ark. Kemi, 1 (1949)401. W.Grassmann and K. Hannig, Naturwissenschaften, 37 (1950)397. W.Grassmann and K. Hannig, Ger. Pat., 805399,May 24th, 1949. R. Dobry and R. K. Finn, Chem. Eng. Prop.,54 (1958)59. R. Dobry and R. K. Finn, Science, 127 (1958)697. J. E. Barrollier, E. Watzke and H. Gibian. 2. Nafurforsch. B. 13 (1958)754. K. Hannig.2. Anal. Chem., 181 (1961)244. H.C. Mel,Science, 132 (1960) 1255. H.C. Me1.J. Theor. Biol., 6 (1964) 159. H. C. Me1,J. Theor. Biol., 6 (1964) 181. A. Kolin and P. Cox, Proc. Nat. Acad. Sci. US.,52 (1964) 19. R. Hybarger, C.W. Tobias and T. Vermeulen, Ind. Eng. Chem., Roc. Des. Develop., 2 (1963)65. T. Vermeulen, L. Nady, J. M. Krochta, E. Ravoo and D. Howery, Ind. Eng. Chem., Proc. Des. Develop., 10 (1971)91. A. Kolin,J. Chromatogr., 26 (1967)180. A. Strickler, Separ. Sci., 2 (1967)335. A. Strickler and T. Sacks, Prep. Biochem., 3 (1973)269.

REFERENCES

25 I

18 A. Strickler and T. Sacks, Ann. N. Y.Acad. Sci., 209 (1973) 497. 19 G. F. Aitchison, P. Mattock, D. Steptoe, A. R. Thomson and W. J. C. White, Tenth International Congress of Biochemistiy, Hamburg, July 1976, Abstracts, p. 677. 20 P. Mattock, G. F. Aitchison and A. R. Thommn, personal communication, 1977. 21 L. R. Creight, R. N. Griffii and R. J. Locker, ESA Special Publication No. 114, Continuous-flow Separation for Biologicals, (Soyuz-Apollo Flight), 1975. 22 K. Hannig and H. Wirth, ESA Special Publication No. 114, Detailed Results of A S P Experiment MA-014-Continuous-flowElectrophoresis, (Soyur-Apollo Flight), 1975. 23 M. Bier,Science, 125 (1957) 1084. 24 K. Hannig, Preparative Electrophoresis, in M. Bier (Editor), Electrophoresis, Academic Press, London, 1967, p. 423. 25 K. Hannig, personal communication, 1977. 26 E. Ravoo, P. J. G e U i s a n d T. Vermeulen, Anal. Chim Acta, 38 (1967) 219. 27 A. Kolin, Proc. Nat. Acad. Sci. US.,46 (1960) 509. 28 A. Kolin and St. Luner, Anal. Biochem., 30 (1969) 111. 29 K. Hannig, H. Wirth, B. H. Meyer and K. Zeiller, Z. Physiol. Chem., 356 (1975) 1209. 30 W. Haller, U.S. Pat., 3,758,284 (1965). 31 J. Porath, Ark. Kemi, 11 (1957) 161. 32 Q. P. Peniston, H. D. Agar and J. L. McCarthy, Anal. Chem., 23 (1951) 994. 33 S. Hjertkn, Free Zone Electrophoresis, Dissertation, Uppsala, 1967, pp. 64-66. 34 K. Zeiler, R. Loser, G. Pascher and K . Hannig, Z. Physiol. Chem., 356 (1975) 1225. 35 K. Hannig, H. Wirth, R. K. Schindler and K. Spiegel, Z. Physiol. Chem., 358 (1977) 753. 36 Z. Prusfk, J. Chromatogr., 91 (1974) 867. 37 E . L. Durrum,J. Amer. Chem. Soc., 73 (1951) 4875. 38 W. Grassmann and K. Hannig, Z. Physiol. Chem., 292 (1953) 32. 39 F. Sorm and B. Meloun, in M. Bier (Editor), Primary Protein Structure in Electrophoresis, Academic Press, London, 1967, p. 90. 40 H. Wagner, D. Neupert and K. Schlick, J. Chromatogr., 115 (1975) 357. 4 1 K.Wiek,J. Chromatogr., 13 (1964) 111. 42 K. Hannig and H. Heidrich,Methods Enzymd., 21 (1974) 746. 43 Z. Prusik,E . Sedlhkovd and T. Barth, 2. Physiol. Chem., 353 (1972) 1837. 44 K. Hannig and H. Wirth, Z. Anal. Chem., 243 (1968) 522. 45 S. Chandresekhar, Hydrodynamic and Hydromagnetic Stability, Clarendon Press, Oxford, 197 1, pp. 83 and 273. 46 G. I. Taylor, in G. K. Batchelor (Editor),Scientijic Papers, Vol. IV, Cambridge University Press, Cambridge, 1971, pp. 17-23. 47 L. Raleigh, Scientific Papers, Vol. VI, Dover, New York, 1964, p. 452. 48 W. Grassmann, Angew. Chem., 62 (1950) 170. 49 G. D. Howery, T. Vermeulen and L. Nady, J. Chromatogr. Sci., 10 (1972) 557. 50 W. Grassmann, Naturwissenschaften, 38 (1951) 200. 51 N. Seiler, J. Thobe and G. Werner,Z. Anal. C h e m , 252 (1970) 179. 52 K. Hannig,Z. Pyrysiol. Chem., 338 (1964) 211.

This Page Intentionally Left Blank

Chapter 12

Continuous flow deviation electrophoresis*

.

A KOLIN

CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical aspects of continuous flow deviation electrophoresis . . . . . . . . . . . . . . . . . . . Velocity distribution in fluid band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Distortion of the streak cross-section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolving power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroosmotic streaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abolition of streak-profile distortion by electroosmosis . . . . . . . . . . . . . . . . . . . . . . Modes of implementation of continuous flow deviation electrophoresis . . . . . . . . . . . . . . . Flat fluid band electrophoresis (“free-flow” electrophoresis) . . . . . . . . . . . . . . . . . . . . . The preparative instrument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modification of the fluid curtain apparatus for analytical work . . . . . . . . . . . . . . . . . . Endless fluid belt electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives of the system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The circular endless belt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle of suppression of thermal convection . . . . . . . . . . . . . . . . . . . . . . . . . Achievement of electromagnetic circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . A simple apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of magnetohydrodynamic driving force . . . . . . . . . . . . . . . . . . . . . . . . . Some illustrations of instrument performance . . . . . . . . . . . . . . . . . . . . . . . . . . Some unique properties of the endless belt system . . . . . . . . . . . . . . . . . . . . . . . Modes of operation of the endless fluid belt system . . . . . . . . . . . . . . . . . . . . . . The noncircular endless belt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transition to the vertical “racetrack” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The separation space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Serpentine fluid belt electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Examples of preparative separations by continuous flow deviation electrophoresis . . . . . . . Measurement of electrophoretic mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Symbols and units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

253 255 255 257 258 259 260 261 261 261 264 267 267 267 267 269 269 270 272 273 275 276 276 277 278 279 282 291 295 295 296 296

INTRODUCTION

In this chapter the frequently used term “free-flow’’ electrophoresis will not be used. as it is inappropriate . An example of fluid motion to which the term “free flow” could be

* The author of this chapter preferred not to use the recommended SI units.For the convenience of readers. a table of symbols may be found on p . 296 .

254

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

applied is the efflux of water from a tap. It flows without solid confines and in this sense it is “free flow”. In the electrophoretic methods that will be considered here, a continuous fluid flow is confined between walls in channels of different shapes and we shall classify the methods according to the shape of the confining channel. If the term “free flow” was meant to indicate an absence of solid obstacles that would retard the flow by viscous drag, it would be equally untenable because the walls of the channel exert such a retarding force which results in the laminar flow profile. The subject of this chapter is electrophoresis in electrically conductive fluid ribbons flowing between two solid walls, which may be flat or curved. An electric field is maintained at right angles to the direction of flow. The charged particles to be subjected to electrophoresis are injected into the fluid stream as a fine streak. Owing to superposition of the transverse electrophoretic velocity upon the fluid flow velocity, the direction of the streak is deviated by an angle ad, which depends on the electrophoretic mobility of the particle, the electric field intensity and the velocity of fluid flow. Fig. 12.1a1 shows the simplest configuration of this type. The fluid flows downwards with a velocity u between the plates P, and Pz (at first we shall make the unrealistic assumption of a uniform velocity distribution). E is the transverse electric field and is the angular deviation of the electrophoretically slower negative particles from the vertical dotted line, which is the path of electrically neutral particles; a2 is the angle of deviation of an electrophoretically faster component in the particle mixture emerging from the injector, IN. This figure illustrates the essence of deviation electrophoresis, which is the subject of this chapter.

U

”f

IN

2

a2

b

C

Fig. 12.1. Fluid paths used in deviation electrophoresis. Perspective view of the flatcurtain flow. P I , P,: plates between which the fluid flow is sandwiched. MI,M,: membranes forming the lateral confines of the curtain flow. IH: influx h+oIes for buffer. EH: exit holes for buffer. IN: injection capillary for sample. 2: fluid velocity vector. E: electric field vector. a,,a*:angles of deviation of two electrophoretic components. (a,) Lateral view of curtain. Symbols as in a,. (b) Serpentine flow path. S,, S,: solid serpentine surfaces between which the serpentine flow is sandwiched. M I: lateral membrane (m, on the opposite side of walls S,, S, is not shown). 2: fluid velocity vector. (c) Endless fluid belt. S , , S,: solid surfaces confiing the fluid belt. 3 fluid velocity vector. (There are no lateral membranes confining the flow.)

THEORETICAL ASPECTS

255

Fig. 12.1a2 is a side view of the configuration shown in Fig. 12.1al. The electric field vector is to be imagined as pointing away from the reader at right angles to the page. This is the scheme of what is referred to in the literature as “free-flow electrophoresis” [2]. Another configuration that has been used [3] employs a serpentine ribbon of fluid between suitably shaped solid confmes, as shown in Fig. 12.lb. Finally, the fluid band may have the shape of an endless belt of an arbitrary shape. Fig. 1 2 . 1 shows ~ a shape [4] that is used in a more recent version of the endless belt electrophoresis system [5]. We can thus see that the continuous fluid flow may be unidirectional, tortuous or cyclic. In one of the figures illustrating such motions (Fig. 12.1b and 1 2 . 1 ~ the ) electric field is perpendicular to the page and the sample to be subjected to electrophoretic analysis is injected in the direction of the velocity ti or vf. The resolving power of this electrophoretic scheme for the separation of molecular mixtures is considerably lower than that of methods in which chromatographic and/or molecular sieving effects are combined with electrophoresis. However, the absence of a porous matrix makes continuous flow deviation electrophoresis ideally suitable for the separation of mixtures of particles, such as cells and cell organelles. In molecular separations, the disadvantage of a lower resolving power is compensated for by the advantageous preparative capabilities of continuous collection of separated fractions.

THEORETICAL ASPECTS OF CONTINUOUS FLOW DEVIATION ELECTROPHORESIS The following theory is common to all modes of continuous flow deviation electrophoresis, namely to flat curtain (“free-flow”), serpentine flow and endless belt electrophoresis. It was developed in connection with endless belt electrophoresis [6,7], with reference t o its applicability to fluid curtain electrophoresis, to which it was eventually applied [2,8]. In fact, the treatment of curved fluid paths uses the approximation of a fluid band width of negligible thickness compared with the radius of curvature and treats the fluid ribbon as being flat. Velocity distribution in fluid band We shall now consider, as a prototype of fluid motion in a deviation electrophoresis apparatus, the flow between two parallel plates. Fig. 12.2 illustrates fluid circulation in an endless fluid belt around a noncircular core C. This fluid belt has two flat sections adjacent to the front and back vertical surfaces of the core C. As a result of viscous interaction with the walls, the velocity profile in the fluid curtain is given by [9] (1 2.1 a) (1 2.1 b)

Eqn. 12.la follows from two integrations of the differential equation [lo]

256

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fig. 12.2. Perspective view of the endless fluid %It (partial view, showing the front portion of the fluid belt) [5]. C: Inner core; M: outer mantle. u : fluid velocity; vo: fluid velocity a t centre of fluid belt. X, Y,Z coordinate axes centred at mid-point o:fluid belt. Z: distance o t a point from centre of belt. f,: force density of force propelling the fluid;& magnetic field vector; J : current density vector. h : thickness of endless belt; v: lateral flow velocity vector; F, R: front and rear surfaces confining the fluid belt, respectively; L , : section on level 1; L,: section on level 2. C,: original circular cross-section of injected sample streak;C,, C,: shapes of distorted sample streak downstream of C , . D: distance between levels C, and C,. a: angle of deviation of electrophoretic component. 1-1: points along line A-B, I*?*: final locations of points originating from line A B . A*B*: final position of points on line A B in ideal case.

(12.2) and consideration of the velocity u vanishing at the walls;f, is the vertical force per unit volume which maintains the flow of the fluid belt. Eqn. 12.lb is the familiar velocity distribution of laminar flow. It is immediately clear that only the centrally located particles injected from a well centred injector IN in Fig. 12.la

THEORETICAL ASPECTS

257

will move downwards with the maximal curtain velocity uo. The particles closer to the walls of the curtain will move more slowly. This would have an obvious effect upon the deviation of a charged particle in the transverse electric field E . It is also clear that the injector must be centred exactly in the fluid curtain and must be parallel to the flow. In the worst case, the injector would be perpendicular to the walls of the curtain confining the fluid band injecting its output in the direction of one of the walls. This would increase the excentric spread of the particles, making their velocity profile more inhomogeneous. Because of the flatness of the velocity distribution curve near the centre (Fig. 12.2), a good approximation to a uniform vertical particle velocity within the streak emanating from a well aligned injector can be achieved by using an injector with a small diameter compared with the thickness, h , of the fluid band (see also subsequent discussion of resolving power). Distortion of the streak cross-section Fig. 12.2 shows two perpendicular sections, L1 and L2, through the fluid band. Let us imagine that a species of negatively charged electromigrating particles is introduced on the upper level L1,not through a tubular injector, but rather through a fine slit AB extending almost from wall to wall (we avoid the immediate vicinity of the wall where the downward flow velocity u is zero); 1, m, n, 0 , p, q and r are representative particles (o being centrally located) at different distances z from the centre of the fluid band. The electrophoretic velocity of the particles, u, = - pE,is the same for all values of z . If the same were true of the vertical particle velocity, u , within the fluid band, the above row of particles would maintain its linear configuration on its way down while moving to the left due to electrophoresis. It would eventually arrive at the downstream level L in a linear arrangement between points A* and B* .We could imagine a collection system, COL, consisting of fine exit slits parallel to A*B* intercepting the arriving particles. The particle that we are considering would thus exit through slit A*B* . A particle species moving more rapidly in the same direction would have entered a slit to the left of A* B* . Let us now remember that the downward velocity, u , is not uniform but rather is represented by the parabolic velocity profile shown in Fig. 12.2. As a result, the centrally located particle o would move downwards faster than all of its neighbours and would thus require the shortest time to reach the level L2. Its displacement to the left will therefore be smaller than for other particles located on the line AB owing to the shorter electromigration time. The closer a particle is located to one of the walls confining the fluid flow, the slower will be the downward velocity u and the longer will the particle migrate t o the left on its way to the level L.As a result, particles 1, m, n, 0,p, q and r will no longer lie on a straight line by the time they have reached the level b.Their arrangement will have been distorted as indicated by the points 1*, m*, n*, o*, p*, q* and r* on level L2. These particles will no longer pass through the single linear slit A*B* as they did for a uniform distribution of the downward velocity u , but will now invade many neighbouring escape slits. The effect of this distortion of the slit profile on the resolving power in preparative separations is obvious. In practice, the injector is not shaped like a slit, but is rather a thin cylindrical tube. The larger is the desired throughput in preparative work, the larger will be the diameter

258

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

of the injected particle streak. Such a circular initial particle distribution is depicted as circle, C1, on level L1. We can imagine the particles within C1 to be arranged in lines parallel t o AB. These linear particle arrays will suffer the same distortion as was illustrated for the particle row AB. Thus, the initially circular cross-section C1 of the particle streak will have been distorted into the shape C1, vaguely resembling a crescent by the time the particles have reached the level L,. Particles issuing from position C1 with a higher electrophoretic mobility will form a “crescent”, C , , to the left of C2. If the collector at level L, consists of a row of adjacent circular openings, the particle streak of originally circular cross-section could no longer pass through one entrance circle. The “horns” may invade adjacent collector openings, thus decreasing the preparative resolving power. The simplest means of reducing this streak-broadening effect is to reduce the diameter of the circle C1 [6]. The particles will then lie close to the apex of the parabola of the velocity profile where the velocity varies relatively little with the distance z from the velocity maximum vo. One thus has to make a compromise between high resolution and a high throughput in preparative work, sacrificing high resolution in order to increase the yield. The above effect of impairment of resolving power is due to a mismatch between the parabolic velocity distribution in the flowing fluid belt and the uniform velocity distribution in the transverse particle motion due to electrophoresis. We have considered in a previous example the case where both the vertical and the horizontal velocity distributions were uniform. There was no distortion of the streak profile. Similarly, it is easy to show [7] that there would be no streak profile distortion if the vertical and horizontal velocity distributions were both parabolic. In general, both of these velocity distributions may be arbitrary. There will be no streak distortion if the vertical and horizontal velocity distributions are described by the same function [5, 71. Resolving power When, in Fig. 12.2 at the level b,we view the separated streak originating from C1 in the direction of the Z-axis, the “horns” of the crescent Cz may obstruct from vision the apex of the “crescent” C, . The electrophoretic fractions, although distinctly separated, would appear in a side view to form a single broad streak. We shall consider two electrophoretic components as resolved when the angle of deviation of the apex of the “crescent” (trailing edge) of the faster component ( C , in Fig. 12.2) is larger than the angle of deviation of the “horns” (leading edges) of the slower component (C, in Fig. 12.2). On the basis of the above criterion of resolution [7] and a definition of resolving power, R , as the ratio between the mean mobility, p, of two electrophoretic components and the difference, A p , between their mobilities [7]:

we can arrive at an expression for the resolving power in terms of the instrumental parameters d and h of the continuous flow electrophoresis apparatus [7]:

259

THEORETICAL ASPECTS

(12.4)

This expression shows that very high resolution can be obtained by making the streak diameter, d , very small in comparison with the thickness, h , of the fluid curtain. For instance, in order to resolve two components that differ by 1% in their mobilities ( A p = 10-2p)rwe could choose an injector of inner diameter 0.15 mm with a fluid curtain 1.5 mm thick. Electroosmotic streaming In all of the systems of implementation of continuous flow deviation electrophoresis, there is an unintended complex fluid motion, i.e., the electroosmotic flow, which has usually been considered as a source of a serious disturbance that would impair resolution. However, it was shown in 1960 [ 6 ] that one can actually take advantage of this effect in order to improve the electrophoretic resolving power by abolishing the streak broadening (“crescent” effect) caused by mismatching of the electrophoretic and hydrodynamic velocity profiles. As a result of the ionic double layer at the walls confining the fluid ribbon, electric forces set in motion the electrolyte adjacent to the cell walls. A flow in the opposite direction results in the central region between the fluid “curtain” walls to ensure zero net fluid transfer in a closed system. Such electroosmotic streaming between two parallel plates PI,P2 is shown in Fig. 12.3. The dashed lines depict the electroosmotic convection and the solid line shows the resulting parabolic velocity distribution [5]. This flow is unlike the parabolic flow maintained at right angles to the electric field

I

0

Fig. 12.3. Electroosmo_tic convection: infinitely extended parallel plates separated by distance h which confine the fluid 151. J electric current density. Z,, 2,: Srnoluchowski zones of zero velocity. The closed dashed lines depict qualitat$ely electroosmotic convection. The electroosmotic velocity distribution is shown by the parabola. v,: central velocity vector. v,: velocity at the walls (v, = - 2 ~ ~ ) .

260

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

(velocity ;in Fig. 12.2), where the velocity vanishes at the wall. The velocity is actually highest at the walls, v,, and is opposite to the central velocity, vo. Smoluchowski [l 11 showed that there are two parallel zones of zero velocity located at a distance of 0.21h from the walls of the cell. These zones are labeled Z, and Z2 in Fig. 12.3. The velocity distribution for the electroosmotic flow follows from considerations [121 similar to those which led to eqn. 12.1. The horizontal force density,fy ,which is responsible for the electroosmotic streaming is given by t

62V

fY

=7)s

(12.5)

Its integration yields V =

"[

:I'

- z+27)

[ I;]

+Az+-

+B

(12.6)

The condition that the velocity becomes % at the cell walls yields the integration constants A and B. We thus obtain the velocity distribution for the electroosmotic flow [6, 121: (12.7)

Abolition of streak-profile distortion by electroosmosis The horizontal velocity, v", of a charged particle in the electrophoresis apparatus will be a resultant of its electrophoretic velocity, v, = - $3, and the electroosmotic streaming velocity, given by eqn. 12.7: (12.8) We have introduced in this equation the symbol W for the electroosmotic mobility, i.e., for the electroosmotic velocity at the wall at unit electric field strength of E = 1 V/cm, and thus the electroosmotic velocity is v, = EW. The slope of the path of a particle will be given by the ratio of its horizontal velocity, v * , to its vertical velocity, u . We thus obtain from eqn. 12.la for u and eqn. 12.8 for v* (12.9a)

Thus, (1 2.9b)

This is the essence of an equation derived in 1960 [ 6 ] ,which showed that the angle a becomes independent of the coordinate z (i.e., from the distance of the particle from the wall) when the second term in eqn. 1 2 9 b vanishes. It was pointed out that this would happen only for one set of particles whose electrophoretic velocity was equal to the

MODES OF IMPLEMENTATION

26 1

electroosmotic wall velocity. For a streak of such particles, no matter how thick (it could be as thick as the fluid curtain), all particle trajectories would be parallel and the crosssection of the streak would remain constant without broadening. This would be most favourable for high-yield preparative electrophoresis. The disadvantage is that it is not simple to modify the wall zeta potential in order to make the second term in eqn. 12.9b vanish. A clever method for modification of the wall zeta potential in a flat continuous flow electrophoresis cell was suggested in 1973 by Strickler and Sacks [8]. However, this method, even when maximally successful, could achieve optimal collection for only one electrophoretic component of a given mixture.

MODES OF IMPLEMENTATION OF CONTINUOUS FLOW DEVIATION ELECTROPHORESIS FLAT FLUID BAND ELECTROPHORESIS (“FREE-FLOW’ ELECTROPHORESIS) The origins of flat fluid band continuous flow electrophoretic separators go back to prototypes of deviation electrophoresis which utilized a fluid flow through a porous medium such as a bed of sand or glass beads [13] or a filter-paper curtain [14]. The porous matrix was effective in suppressing thermal convection but proved troublesome in separations of particulate component mixtures, such as cells. This led to the development of fluid curtain electrophoresis apparatus in whch the porous matrix was omitted [ 15, 161. These early instruments consisted of parallel dielectric plates between which was sandwiched a thin layer (typically 0.5 mm in thickness) of a flowing buffer solution. The plates had a slight inclination against the horizontal plane which helped to suppress thermal convection. However, sedimentation of particles and of the injected streak of molecular solutes towards the lower plate led to a modification in which the plates were mounted vertically [ 171. An elaborate thermoelectric cooling system with Peltier elements was used to minimize thermal convection. A substantially vertical fluid path was later also adapted to the endless belt electrophoresis system [4] for the same reasons.

The preparative instrument Hannig’s first preparative vertical plate instrument [ 171 was designed as follows (Fig. 12.4). Fig. 12.4a shows the front view and Fig. 12.4b a section along the axis A-B indicated in Fig. 12.4a; 3 and 4 are glass plates of unequal thickness between which the fluid curtain is sandwiched (see legend for dimensions and other specifications). The buffer flow is delivered at a rate of 0.2-50 ml/h through six openings (8) from a peristaltic pump which acts upon six silicone rubber tubes. The buffer of the fluid curtain leaves at the bottom through 50 silicone rubber tubes (9) (0.5 mm I.D.) which are driven by a 50channel peristaltic pump. The output of these exits tubes is channelled to a fraction collector. The rate of buffer delivery through the channels (8) must, of course, be equal to the rate of buffer drainage through the channels (9). The achievement of this critical

262

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

A

2

9

a)

1 B

Fig. 12.4. Scheme of fluid curtain apparatus [ 171. (a) Front view; (b) section along A-B. 1 : Copper plate (48 X 48 cm); 2: Peltier thermoelectric cooling elements (Philips valve; Type PT 20/20); 3: glass plate (50 X 50 cm, 3 mm thick); 4: glass plate (50 X 50 cm 6 m m thick); 5: spacers (0.5 mm thick); 6: insulating mantle; 7: electrode chambers; 8: buffer inlets; 9: collector tubes; 10: water level tube; 1 1 : resistance thermometer; P and 12: sample inlets; 1 3 and 14: inlet and outlet for cooling water; 15: ion-exchange membrane.

adjustment for equality of buffer in- and out-flow through the tubes 8 and 9 which are operated b y two different peristaltic pumps [17] is achieved as follows. A capillary tube (1 0) shaped as shown in Fig. 12.4 communicates with the fluid curtain and is filled with buffer up to a point in its lowest horizontal section. When the rate of buffer delivery exceeds the rate of buffer drainage, the meniscus in the horizontal section of tube 10 moves to the right;in the converse case it moves to the left. One of the two peristaltic pumps can now be adjusted so that the meniscus in 10 does not move, which indicates equality of in- and out-flow of buffer. The location of the fluid curtain is indicated in Fig. 12.4b by arrows (top and bottom) which point in the direction of flow and indicate where the buffer enters the space between the glass plates (3 and 4) and where it leaves this separation chamber. Spacers (5) (0.5 mm thick) ensure parallelism of the glass plates and fix their distance apart. The sample mixture t o be separated is injected into the fluid curtain through channels (12), of which the channel P is shown in use. As the sample enters at right angles to the

FLAT FLUID BAND ELECTROPHORESIS

263

plane of the curtain, there is the following problem. Optimal pIacement of the sample streak would have been at the centre of the curtain at the maximum of the parabolic velocity profile. Even if the injector tube (IN) were to terminate in the mid-plane of the curtain, the momentum of the injected fluid would tend to carry it towards the glass plate facing it, i.e., away from the optimal mid-plane. It is thus clear that the rate of sample injection could affect the centering of the sample streak within the curtain. This problem can be avoided by orienting the injector parallel to the curtain flow [4]. The cooling is accomplished by 28 thermo-electric Peltier elements ( 2 ) ,the cool junctions of which face the copper plate (1) and whose hot junctions are cooled by water entering at 13 and leaving at 14. The copper plate serves to achieve a more uniform temperature distribution along the glass plate (3) that is being cooled by 28 discrete cooling elements spaced along the 50 x 50 cm surface area of the curtain surface. Uniform thermal contact between the copper plate (1) and glass plate (3) is secured by a thin layer of silicone oil. As only the rear glass plate is cooled, the temperature distribution normal to the glass plates cannot be symmetrical. There is also a vertical temperature gradient because the cool buffer entering the curtain at the t o p is heated by the electrophoretic current as it moves toward the collection tubes (9), which it reaches at a higher temperature. An electrical resistance thermometer permits measurement of the curtain temperature at location 1 1. Laterally adjacent to the 0.5 mm thick fluid curtain are the two electrode chambers(7), which contain the positive and negative electrodes shown in Fig. 12.4a. These chambers are perfused by a buffer with an electrical conductivity higher than that of the intermediate buffer (3-fold or higher) in order to reduce heating in the electrode chambersaswellas losses in available voltage, The ionic strength of the electrode compartment buffer is about 0.03-0.05 and the electrical conductivity is approximately 0.1-0.25 (ohm-cm)-’ [ 171. The electrode compartments (7) are separated from the fluid curtain by ionexchange membranes (1 5), which seal them hydraulically from the fluid in the curtain while permitting an electrical current t o flow between them. An anionexchange membrane is used on the cathode side and a cation-exchange membrane on the anode side. These membranes have the advantage of a low electrical resistance (and hence lower heating effects) compared with neutral membranes, such as acetylcellulose. The asymmetry in the properties of the cathodic and anodic ionexchange membranes results in changes in electrolyte concentration within the separation curtain with an enrichment of anion concentration near one membrane and cation concentration near the other [ 171. The resulting gradients in pH and electrical conductivity are illustrated in Fig. 12.5. It can be seen that the uniform range o f these parameters extends only about from fraction No. 8 to fraction No. 40, i.e., over approximately two thirds of the separation space. The data plotted in Fig. 12.5 were obtained with a typical buffer solution traversing the curtain and exiting from the terminals (9) in Fig. 12.4a. The plotted values of pH and conductivity are those of the collected fractions in the numbered collection tubes. These membrane-induced shifts in pH and electrical conductivity of the buffer in the separation space are one of the main distinguishing essential features between fluid curtain (“free-flow”) and endless belt electrophoresis. The former system requires the membranes because the curtain flow must be maintained across a rectangular channel by means of a pressure gradient, which could not be generated without the channel’s side

264

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fig. 12.5. Conductivity and pH distribution in the collected fractions after electrophoresis in the fluid curtain apparatus [ 171. Acetate buffer, pH 4.8; conductivity K = 2.6 10-3(ohm * ern)-'. Solid line, pH distribution, and dashed line, conductivity distribution in collected fractions.

-

walls (membranes). To transmit the transverse electrical current, these side walls must be electrically conductive. The mechanism of fluid propulsion in endless belt electrophoresis, on the other hand, requires no pressure gradient and is purely magneto-hydrodynamic. A force is exerted upon each fluid element of the endless fluid belt (which corresponds t o the buffer curtain of the former system) by the interaction of the electrophoretic current with a transverse magnetic field. As a result, the lateral boundaries of the endless fluid belt need not be closed and the membranes can be (and actually are) omitted altogether [18, 191. This omission results in elimination of above-mentioned temporal and spatial shifts in pH and electrical conductivity within the separation space. Fig. 12.6 shows a partial view of the instrument. The dimensions are apparent from comparison with the separation curtain whose glass plates (1) are about 50 x 50cm. On the right-hand side of the fraction collector (6) is the electrophoresis d.c. power supply (21, and on the left the d.c. power supply for the Peltier elements. The buffer supply vessels (4)are located in the cooling chamber ( 5 ) and the mixture to be analysed is located at 7 and delivered by a pump (8). This instrument is particularly suitable for preparative cell separations.

Modification of the fluid curtain apparatus for analytical work Recently, the preparative instrument described above has been modified [20] so as to incorporate some of the performance features of the endless belt apparatus [ 4 , 7 , 181. The length of the curtain is approximately equal to the circumference of the current endless belt apparatus, the cooling of the curtain is no longer accomplished by Peltier elements but rather, as in the endless belt apparatus, by cooling water circulating through cooling jackets adjacent to the curtain, The streaks can be seen and photographed or photoelectrically recorded through quartz windows and the exiting buffer can be fed back so as t o reenter the curtain at the t o p to achieve a buffer circulation as in the endless belt apparatus.

FLAT FLUID BAND ELECTROPHORESIS

265

Fig. 12.6. View of the main part of the vertical fluid curtain apparatus [ 171. 1 : Separation chamber; 2: d.c. power supply; 3: d.c. power supply for Peltier cooling elements;4: buffer supply vessels; 5: cooling chambers;6 : fraction collector; 7 : sample reservoir; 8: pump for sample supply.

We shall describe this instrument with comparative references to its flat-curtain predecessor and to the endless belt apparatus. Fig. 12.7 shows an exploded view of the separation chamber. It consists of two frames (1 and 2) that can be clasped and locked together, the fluid curtain being sandwiched in between them. Glass plates (3) form the walls of the fluid curtain. Quartz windows (5) permit the transillumination of the buffer curtain with W light for streak detection. A wavelength of 225 nm (range of the peptide absorption band) is used for detection of protein streaks. Cooling jackets (4) are provided inside the frames (1 and 2), and are supplied with cooling fluid through nipples (1 5 ) . Platinum electrodes (7) are located in electrode chambers (6) through which the electrode buffer solution is circulated via nipples (1 6). Ionexchange membranes (8) are placed on the sides of the curtain with their planes parallel to it and are sealed against leakage by sealing strips (9 and 10). A spacing step (1 1) fixes the distance between the chamber walls (3), i.e., the thickness of the buffer curtain. The buffer solution enters the curtain at 12 and exits at 15. Thus, unlike the previous preparative instrument, this one

266

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

@j

i

c 1

I

4 , i -

11

Fig. 12.7. Exploded view of analytical fluid curtain separation chamber 1201. I , 2: Frames; 3: chamber walls; 4: cooling jacket; 5: quartz window; 6: electrode chambers; 7 : Pt electrodes: 8: ionexchange membranes; 9, 10: sealing gaskets; 1 1 : spacing step; 12, 13: inlet and outlet of curtain buffer; 14: sample inlet; 15: inlets for cooling fluid; 16: inlet for electrode chamber buffer,

uses an upward flow of curtain buffer. The sample to be subjected to electrophoretic analysis through an inlet (14). The width of the windows, which limits the electrophoretic migration distance, is 3 cm, compared with 50 cm in the preparative curtain and 6 cm in the current endless belt apparatus. The height of the buffer curtain which determines the residence time of the sample in the curtain at a given rate of buffer flow, is 18 cm, compared with 50 cm in the preparative instrument and about 20 cm circumference of the endless belt instrument. Actually, I f turns is the minimal path length in the endless belt, which corresponds to a migration path of about 30 cm. Like the preparative instrument, this analytical device uses a much narrower thickness of the buffer curtain (0.35 mm) as compared to the endless belt (1.5 mm). Although this has the advantage of better cooling and greater stabilization against thermal convection, this narrowness of buffer gap has the disadvantage of a steeper velocity gradient in the flowing curtain for a given maximal buffer velocity in the buffer curtain. The result is a greater tendency for streak broadening (“crescent” effect) to occur, which impairs the sharpness of visual resolution. To counter this, exceedingly thin (0.05 mm) well centred streaks of sample must be used. The voltage of 140V/cm applied

ENDLESS FLUID BELT ELECTROPHORESIS

267

across the narrow width of the cell (3 cm) accounts for the rapid separations that can be achieved with this instrument (see Fig. 12.23). The volume of sample injected may be as small as 0.1-0.3 p1 (comparable with 0.1 pl reported for the endless belt apparatus [7]).

ENDLESS FLUID BELT ELECTROPHORESIS Objectives of the system The objective of this method [4-7, 211 is to utilize a combination of electric, magnetic and gravitational fields to achieve a stable system of deviation electrophoresis. The constancy of the gravitational field can be matched, for all practical purposes, by the constancy of the field of a permanent magnet. The constancy of the electric field, however, requires a current- or voltage-stabilized power supply like any other electrophoretic method. The gravitational and magnetic fields serve to propel the buffer solution without any discontinuities in operation such as those which can be minimized, but not entirely eliminated, in a peristaltic pump. While maintenance of the gravitational and magnetic field within a buffer solution produce no changes in it, the electric field can only be maintained with a concomitant electric current which heats the solution. This heat escapes by conduction via the solid boundaries of the flowing ribbon of buffer. As a result, the buffer temperature will be lowest at these boundaries and highest at the centre of the buffer curtain. The result will be thermal convection [21], such as the circulation pattern shown in Fig. 12.8. It is obvious that such convection could re-mix or at least blurr a streak separation pattern. Hence its effective suppression is the main objective in any similar separation method. Endless fluid belt electrophoresis achieves the suppression of thermal convection vortices by fluid circulation in a vertical plane about a horizontal axis as described below [6,21]. The circular endless belt Principle of suppression of thermal convection Fig. 12.9 illustrates the principle on which we shall base suppression of thermal convection in the buffer curtain. Fig. 12.9a shows a section through two concentric horizontal circular cylinders. Let us assume that the annular space between them is filled with a liquid and that we maintain the inner cylinder at a higher temperature than the outer cylinder. The vector g gives the direction of the gravitational field. As a result of the temperature gradient between the two cylinders, vortex formation will occur, such as the solid vortex shown on the right-hand side of the annulus. The specifically lighter warm fluid rises near the warm inner cylinder and the specifically heavier fluid, which is cooled by the outer cylinder, descends so that a clockwise circulation of fluid results. Let us now assume that we can manage to rotate the fluid in the annulus in the direction of the upper curved arrow so as to place this vortex in the diametrically opposed location indicated by the dashed circulation path. Owing to inertia, the fluid near the

268

‘i

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

a

a

b

b C

Fig. 12.8. Thermal convection pattern in a rectangular channel cooled at the ve$cal walls W, and w b (51. The temperature is highest at the centre (To) and lowest at the walls (TWJg: gravitationalfield vector.

Fig. 12.9. Thermal convection patterns in channels of d!ferent shapes extending perpendicular to the page [S]. The inner wall is warmer than the outer wall. g: gtavitational field vector. Solid closed line: original thermal convection vortex;dotted closed l i e : inverted vortex after relocation by circulation of the fluid in the channel. (a) Circular annular channel; (b) non-circular quasi-annular channel; (c) endless fluid belt “annulus”.

inner cylinder will move downwards and the fluid near the outer cylinder upwards, as indicated by the arrows. These directions of motion are, however, contrary to the directions of the forces now acting upon the fluid elements. The descending fluid near the inner cylinder, being warmer and specifically lighter than the outer fluid masses, will suffer a net upward force, whereas the ascending cooler, and thus specifically heavier, fluid near the outer cylinder will experience a net downward force. Thus, transplantation of the vortex in this new location exposes it to forces that retard its rotational motion. Eventually the vortical motion will come to a standstill and subsequently resume, but now in a counter-clockwise sense because the warm fluid near the inner cylinder will be rising and the cooler outer fluid will be sinking near the outer cylinder [2 11. As a certain amount of time is required in order to give a vortex its rotational energy, a slow, circular flow of fluid in the annulus in the direction of the upper curved arrow will expose nascent vortices to forces that will periodically tend to reverse, and thus stop, the vortical motion. The circulation path of the fluid need not be circular, as in Fig. 12.9a, but could have the shape of an arbitrary closed path as in Fig. 1 2 9 b and c [21]. It is indeed possible to suppress thermal convection by such imposed circulation of fluid about a horizontal axis so effectively that high electric field intensities can be maintained in the buffer belt (of the order of 100 V/cm) without detriment to the sharpness of the streak separation pattern in deviation electrophoresis [6,2 I]. It now remains t o be shown how one can maintain such a constant circulation of the circular buffer belt in the annulus.

ENDLESS FLUID BELT ELECTROPHORESIS

269

Achievement of electromagnetic circulation Let us assume that the horizontal annular space between the concentric cylinders shown in Fig. 12.9a is filled with a buffer solution and that it represents the electrophoretic separation column. The electrophoretic current will be perpendicular to the page. If we now introduce a cylindrically shaped magnetic pole, say a North pole, into the inner cylinder, its radial magnetic field will be perpendicular to the current. As a result, the same electromagnetic forces that drive an electric motor will be exerted upon the fluid elements in the annulus at right angles to the magnetic field and the electric current and the annular buffer column will be set in rotational motion with clock-like constancy [ 6 ] . If the current vector points away from the reader, the circulation about a magnetic North pole will be clockwise but could be reversed by reversing either the current or the magnetic field. Fig. 12-10 shows how such a radial magnetic field can be generated in practice. It is encountered close t o the surface of a soft-iron cylinder, m, between two cylindrical co-axial bar magnets, NS,facing each other with their like poles. A simple apparatus

Fig. 12.1 1 shows the configuration in the simplest type of instrument in which endless belt electrophoresis can be conducted [ 6 ] .The magnets, NS,with the intermediate softiron core, m (in which a “window”, W,is milled to permit passage of light through the core region), are placed inside an inner transparent plastic (lucite) cylinder, C1. The outer, larger cylinder, C2,leaves a gap of 1.5 mm between the two concentric cylinders, which is the annulus in which electrophoresis is to be conducted. The cylinders C, and C2 are cemented t o the electrode compartments, ECI and EC2, in such a way that the annulus serves as a communication channel between them. The electrodes El and E2 permit the passage of the electrophoretic current through the annulus, thus generating in it an electric field. As soon as the current is turned on, the buffer in the annulus begins to revolve about a horizontal axis. The mixture t o be analysed electrophoretically is in the reservoir R. It is conveyed into the separation space via a fine injection capillary, IN (made of electrically nonconductive material), which places it centrally between the cylinders C1 and C2. Neutral injected particles are entrained in the revolving buffer and are accumulated in a stationary orbit (assuming absence of electroosmosis and other lateral streaming). Negative particles are transported electrophoretically towards the anode while orbiting counter-clockwise, as seen in Fig. 12.8. The combination of these two motions is a left-handed helix indicated as the particle path in Fig. 12.1 1. Positive particles will move towards the cathode in a right-handed helical path. The solid and dashed helical paths in Fig. 12.8 illustrate the helical streaks corresponding to two electrophoretic components of different mobilities. These streaks can be intercepted at the end of their paths, as described below, and conveyed to a fraction collector. This figure indicates another characteristic feature of endless belt electrophoresis. If the separation between two components is x after one revolution, it will be nx after n revolutions as the separation will be increasing by equal increments with each revolution. There is a hazard, however, which is analogous to superposition of grating spectra of

270

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fig. 12.10. Magnetic field distribution about two cylindrical bar magnets, NS, with an intermediate soft-iron cylinder, m [ 71.

different orders encountered in optics. After a certain number of revolutions, the faster component of the nth revolution may merge with the slower component of the (n + 1)th turn. This can be easily avoided by superimposing a lateral flow of buffer in the direction of the electromigration. In the arrangement shown in Fig. 12.8 this will amount to adding buffer to compartment EC2 while removing it at the same rate from EC,.

Effect of magnetohydrodynamic driving force Instead of considering a curved buffer belt, we shall make an approximation in which the distance, h , between the cylindrical walls confining the buffer belt will be considered to be infmitely small in comparison with their radii of curvature. This amounts to a

ENDLESS FLUID BELT ELECTROPHORESIS

27 1

Fig. 12.1 1. Circular endless belt electrophoretic separator IS].NS: north and south poles of bar magnets; m: intermediate soft-iron cylinder; W: window in m; C,, C,: plastic cylinders, which enclose buffer-filed annulus; EC,, EC, : electrode compartments communicating through the annulus; IN: injector; R: sample reservoir.

consideration of flow between flat plates. A magnetic field of intensity B is maintained at right angles to these walls and a current of density ?flows through the buffer parallel to them. Under these conditions, each fluid element experiences an electromagnetic force per unit volume, say in the direction of the x-axis:

f7, =

lo-"Jxzi]

(1 2.10)

where ?is in A/cmz, B in gauss and f, in dynes/cm3. The result o f this force is a viscous flow of the buffer. The velocity distribution can be found from the differential equation for the force per unit volule [ 2 2 ] :

272

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

(12.11) where q is the fluid viscosity in poise. Remembering that the velocity vanishes at the walls and integrating twice, we obtain [23] (12.1a)

and, for the central maximal velocity, uo, (12.12) From eqns. 12.10 and 10.12 we obtain an expression for the central maximalvelocity, u o , in the fluid belt: (12.1 3) We can now calculate the streaming velocity for a typical set of parameters: h = 0.1 5 c m , i = 2 0 0 m A , J = 0 . 6 1 A/cmZ;B =340gaussand (at 16"C)q= 1.llpoise.Eqn. 12.13 then yields the value uo = 0.53 cm/s. The circumference of a typical instrument such as that shown in Fig. 12.1 1 is about L = 22 cm. From this value and uo,we can calculate the time, t o ,a centrally located particle takes to complete one revolution:

L

t o = - = 41.5s

(12.14)

UO

Some illutrations o f instrument p e r f o m n c e Fig. 12.12 illustrates a helical streak obtained with erythrocytes in an instrument filled with physiological saline [7]. The cells enter through an injector IN at the bottom center and perform 15 turns of undiminishing sharpness, which can be seen by dark-field illumination through a window, W, in the central core, m, in Fig. 12.1 1. Fig. 12.13 shows electrophoretic separation between two yeasts: Saccharomyces cerevisiue and Rhodotomla [7]. A streak of the mixture enters from injector, IN, and splits in two as the particles spiral in a helical path around the core to the left. The consecutive turns are designated by a, b, c and d. The descending streaks in front of the iron core can be seen most clearly, but some of the ascending streaks moving behind the core upwards and to the left can still be distinguished. The separating particles are seen to gain equal increments in separation with each consecutive turn. The double streaks are spread far apart from each other by superposition of lateral streaming from right t o left. A clearly visible separation has been obtained after one turn (or in about 30s). Fig. 12.14 shows a separation [7] of a mixture of small molecular components (dyes). The point of injection of the dye mixture is marked by an arrow. The two slower components are not resolved after one turn, but are clearly separated after two turns (in about

ENDLESS FLUID BELT ELECTROPHORESIS

273

Fig. 12.1 2. A 15-turn helix of erythrocytes in circular endless belt revealed by light scattering [ 71. IN: injector; W : window in central core cylinder.

1 min). Finally, after four turns the dye streaks reach the collector, C, from the openings of which plastic tubes convey the separated components to a fraction collector.

Some unique properties of the endless belt system Of the unique properties of the endless belt electrophoretic system, at this point we shall mention two that are both a consequence of the utilization of the same current for electrophoresis and for the propulsion of the buffer perpendicular to the electrophoretic path. The first is the stability of the streak deviation against variation in the voltage applied to the electrophoretic cell. Fig. 12.15a shows four helical turns of the dye Evans Blue obtained at a cell voltage of 75 V. Fig. 12.15b shows four helical turns of Evans Blue after the cell voltage has been increased to 150 V. The reason why the pitch of the helix (i.e., the electromigration distance traversed by a dye ion during one revolution) has not changed is as follows. Although doubling of the voltage did double the electrophoretic velocity, it also doubled the rate of revolution of the buffer, thus reducing the revolution time by half and with it, of course, the pitch of the helix [ 6 ] . The second unique and essential property of this system is based on the fact that the propulsion of the buffer does not require a pressure gradient. As an electromagnetic force is exerted upon each buffer element, it is not necessary to enclose the channel through which the buffer flows completely. The two surfaces between which the buffer is

274

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fig. 12.1 3. Separation in circular endless belt of two micro-organisms (Saccharomyces cereuisiae and Rhodotomla) revealed by light scattering 171. a, b, c and d represent consecutive helical turns. IN: injector. W: window in iron core.

Fig. 12.14, Separation and collection of dyes in circular endless belt [7].The sample enters at arrow. Five helical turns of separating dye components can be seen (from right to left: Evans Blue, “Brush” green recording ink and Rose Bengal). At the end of their paths, the helical turns are seen entering the collector.

ENDLESS FLUID BELT ELECTROPHORESIS

275

Fig. 12.15. Effect of cell voltage o n helical pitch in endless belt electrophoresis 161. (a) Helix of Evans Blue obtained at a voltage of 75 V across t h e cell and a current of 25 mA. (b) Ifelix of Evans Blue after incrcasing the voltage t o 150 V. There is n o perceptible change in helical pitch.

sandwiched are sufficient and the sides of the flowing belt can be left open to allow it to communicate with the electrode or buffer compartments without separation by membranes. The omission of membranes [18, 191 eliminates the drifts in pH and electrical conductivity in the separation space which they cause [ 171 in fluid curtain electrophoresis. The absolute smoothness of buffer propulsion, which cannot be achieved by a peristaltic pump, is also unique. Finally, the helical path and the cyclic nature of the buffer movement which imparts, without discontinuities, equal increments in separation to the components under analysis are unusual features characteristic of this system.

Modes of operation of the endless fluid belt system The endless fluid belt electrophoresis apparatus can be operated in several different modes 171.

Single-order collection This is the simplest normal mode of operation. The mixture to be separated is introduced via the injector and one then has to decide after how many helical turns the separated fractions are to be allowed to enter the collector. The output of the helical turn in which two or more fractions of interest are far enough apart to enter separate entry ports of the collector (which are about 1 mm apart) can be guided into the collector. Such guidance can be accomplished simply by modifying the imposed lateral flow. Split-order collection Sometimes it is advantageous not to collect all of the separated components after the same number of turns. For instance, suppose that we have a collector with 20 entrance holes and three fractions which, after four turns, have separated so that the two slowest fractions, A and B, enter after four turns the adjacent collector holes 1 and 2, while the fastest component, C, enters hole 5 . One could adjust the lateral flow so that the slowest component just misses the collector and proceeds to make a fifth turn, while the component of intermediate mobility is collected in collector tube 1. The slowest

276

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

component will then, after an additional turn, enter the collector ahead of the fastest component and may, for example, enter collector hole 12. Such a manoeuvre would have been particularly advantageous if components A and B had been too close after four turns to be collected in two separate tubes. An extra turn for both of these components could have separated them sufficiently for separate collection after five turns, while the fastest component would have been collected after four turns.

Orbital accumulation The imposition of an axial flow makes it possible to make the sum of the axial flow and electrophoretic velocities of a particle zero relative to the apparatus for a given particle species. In this instance, such particles would accumulate in a stationary orbit. Particles present at too low a concentration to be observed in the normal mode of operation could thus be detected by orbital enrichment after a sufficient number of buffer revolutions. Axial flow could then be used to shift the storage orbit so as to intercept it by the collector.

Zonal separation Microanalysis of suspension or solution volumes of the order of a few tenths of a microlitre can be accomplished by “orbiting zone” electrophoresis. One injects a fine streak (about 0.1 mm or less in diameter) about 1 cm in length, which forms a tiny zone moving on a helical path, Axial flow can be adjusted so that the helical pitch of the slowest component is zero, i.e., it moves in a stationary circular orbit, The other components of the mixture form other tiny orbiting zones that progress in the direction of the electric field. After the separation of the components has reached the desired value, axial flow is used to guide the separated fractions into the collector. The noncircular endless belt

Tkansition to the vertical “racetrack” The circular endless belt apparatus has the advantage of simplicity of construction, but this virtue is overshadowed by a drawback illustrated in Fig. 12.1 6. Particles entering the centre of the annulus via the injector, IN, will sediment if their density has not been adjusted so as to be equal to that of the buffer. This adjustment would be laborious and of limited usefulness, as equilibrium can thus be achieved for only one component of the mixture. Fig. 12.16A shows how the combination of circulation with sedimentation results in an eccentric particle orbit. If the disparity in density is great enough, the particles may precipitate on the walls of the annulus. A similar behaviour wiIl also be exhibited by a streak of molecular solutes emanating from the injector which will act like a large sedimenting object if it is denser than the surrounding buffer. This problem is solved effectively by distorting the circular annulus of Fig, 12.16A into the vertical “racetrack” [ 4 , 5 , 2 1 ] shown in Fig. 12.16B. The soft-iron core now has an elongated form with high vertical flat surfaces, C1 and Cz.The magnetic field is perpendicular to the surfaces of C1 in the immediate vicinity of the core, m, and the electrophoretic current, flowing at right angles to the page, is perpendicular to this

ENDLESS FLUID BELT ELECTROPHORESIS

277

Fig. 12.16. Particle circulation paths in annuli of different shapes [4]. m: Iron core inside the annulus; IN: injector;C, ,C, : walls confining the annulus; COL: collector; t : collector tubing;CC: collector compensator; T: test-tube.

magnetic field. We thus obtain, as with the circular annulus, a tangential electromagnetic force that moves the buffer as indicated by the arrows. A dense particle entering from the injector will not deviate from the central path until it reaches the short horizontal section where the flow turns around. There, the particle may sediment below the mid-line and continue its upward path to the left of the left vertical mid-line, This slight deviation is, however, corrected when the particle reaches the upper turning point of the flow, where sedimentation returns it to the mid-line, With this shape of the circulation path, one can use particles and solutions whose densities differ considerably from that of the surrounding buffer. Fig. 12.16C shows interception of a descending streak by the collector, COL. The fluid escaping from the collector can be replenished at an equal rate by injection via a compensator, CC.

The separation space The effective separation space of the non-circular endless belt apparatus is shown schematically in perspective in Fig. 12.17. The electrode and buffer chambers have been omitted. The central soft-iron core, C (corresponding to m in Fig. 12.16), is hollow and is cooled by a cooling liquid conveyed via the tubes CP. A quartz window, W,permits observation by visible and ultraviolet light (L in Fig. 12.17) transmitted normal to the core. Only the North poles, N, of the four magnets used are shown. The core, C,is surrounded with a plastic mantle, MA, in which hollow chambers (not shown) parallel to

278

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

CP

Fig. 12.17. Electrophoretic separation space of non-circular endless fluid belt apparatus [ S ] . N: North Boles of magnets (South poles are not shown); C: hollow cooled iron core; CP: tubes carrying cooling fluid to and from C; W: window in C; MA: mantle surrounding annulus (i.e.,&he fluid belt); IN: injector; Cap: ca@llary; COL: collector; t : c+ollector tubing; &? electric field vector;l: electric current density vector; B : magnetic field vector;F: electromagnetic force vector; 1*, 2*: first and second ascending streaks (behind the core C); 1, 2, 3: first, second and third descending streaks (in front of the core C); L: light beam passing at the bottom of the annulus.

the front and back surfaces of the core serve as outer cooling chambers by carrying a cooling solution. Thus, the fluid belt which is sandwiched between the core, C, and mantle, MA, is cooled from both the inside and outside. The sample to be analysed enters the annulus centrally from the injector, IN, and moves towards the intercepting collector, COL, in a non-circular helical path. The first, second and third descending streaks are labelled 1 , 2 and 3, respectively, and the corresponding ascending streaks (behind the core, C) I* and 2* (there is no third ascending streak). The dimensions of the core are roughly 10 cm in height, 6cm in width and about 1 cm in thickness.

The apparatus Fig. 12.1 8 shows the separation cell in a state of partial disassembly [5]. The central separation chamber includes the outer mantle, MA, and the outer cooling chambers, CC (incorporated in the mantle); the route of the cooling water, CW, through the mantle is shown. The tubes CP convey the cooling water to the iron core, C. WO is a window opening and COL (with attached plastic tubes, CT) is the collector which plugs into the gasketed opening, CO. The quartz window, W,, also plugs into a gasketed opening, WO, in the mantle. IN is the injector channel and Cap the injector capillary that conveys the sample solution to the separator (it corresponds to IN in Fig. 12.17).

ENDLESS FLUID BELT ELECTROPHORESIS

279

The left-hand side o f the drawing shows the sequence of chambers in the assembled instrument. In direct contact with the buffer in the annulus are the buffer chambers, BC, which possess three openings: OT, through which the core can be introduced, and O1 and 02, which create hydraulic and electrical communication with the adjacent electrode chambers, EC. Instead of membranes, these openings are closed by perforated plates, P. The perforations are in the lower portions of the plates and serve the following purpose. Bergrahm 1241 has shown that a centrifugal flow of buffer away from a separation column towards the electrodes can effectively prevent migration of ions of electrolysis products from the electrodes into the separation space. This flow must be sufficiently rapid, w h c h is achieved by the small perforations in the plates, P, which offer a small cross-section t o the flow. Bergrahm’s idea was adapted by Luner [19] t o endless belt electrophoresis. The buffer solution for the centrifugal buffer flow is supplied by the Mariotte bottle, MB, shown in Fig. 12.19. It enters the buffer chambers, BC, via the tubes, DT, from the distributor, D. After passing through the perforations in plates P, the buffer moves upwards past the electrodes, E, carrying the evolved gases t o the “balconies”, B, where the flow slows owing t o the large channel cross-section so as t o allow enough time for the gas bubbles t o escape before the fluid descends through the drain tubes, D, in order t o leave the cell via channels. d , and nipples, N. The flow scheme of the instrument is shown in Fig. 12.19. The Mariotte bottle, MB, is a source of buffer at a constant pressure head, which delivers a constant buffer influx into the cell via PVC tubes, DT. I n addition t o creating the centrifugal buffer flow, these tubes can also be used t o create a lateral buffer flow through the annulus. By transferring an appropriate number of tubes from the right to the left buffer chamber, one can change the direction o f the lateral buffer flow or alter its magnitude. The “see-saw”, SS, provides a fine control for the lateral buffer flow by changing the relative level of the end points of the drainage tubes, dt. The Mariotte bottle also feeds the collector-compensator, CC, via a variable-flow resistance, RV. The sample is injected from a syringe, SY, which is driven by a motor, MD, and drive, DR, via tubing, CT, into the injector capillary, Cap. As an example of the preparative throughput of the instrument, a normal run in cell separations will process 0.5 mI o f a cell suspension of 5 * 10’-10*cells/ml in one operation. Such a separation run may last from 30 t o 45 min. Serpentine fluid belt electrophoresis An alternative way of stabilizing a fluid curtain against thermal convection by inversion of nascent vortices is shown in Fig. 12.20. The fluid enters the serpentine flow channel in tube A at IN from a well centred injector and the vortices shown in tube A are inverted after transfer into tube B and inversions are repeated as the fluid moves from one vertical section into another. A similar sequence of inversions with consequent stabilization can also be obtained b y rotating the configuration of Fig. 12.20 through 90” [3,21]. Fig. 12.2 1A shows a serpentine cell and Fig. 12.2 1B and C show a dye separation illustrating the mode of action of such a serpentine separation cell [3]. This system has the advantage o f compacting a long separation path into a small space, but it has two serious disadvantages: (1) the serpentine channel is difficult t o make and (2) like the flat curtain

280

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

ENDLESS FLUID BELT ELECTROPHORESIS

28 1

4 Fig. 12.19. Endless fluid belt apparatus- flow system [S]. SCH: separation chambers; W: window; Cap: injection capillary; COL: collector; CT: collector tubing; CC: collector-compensator nipple; TT: test-tubes; BC: buffer compartments; EC: electrode compartments; E l , E,: electrodes; N: nipples of rigid drainage tubes (not shown); dt: narrow plastic drainage tubes; SS: “see-saw”; P: perforated plates; MB: Mariotte bottle; MT: tube of Mariotte bottle; CS: clean-out syringe; D: distributor; DT: distributor tubes; CT:coiled tubing linking capillary to sample syringe; SY: sample syringe; MD: motor drive; DR: screw and plate which drive the plunger of syringe SY; RV: regulator valve for control of collector compensator inflow.

Fig. 12.18. Endless fluid belt apparatus IS]. The magnets have been removed; they are shown in relation to the iron core C in Fig. 12.17. WO: Window opening; MA: mantle surrounding iron core; S: syringe needle for removal of bubbles from top of annulus; IN: injector mount; Cap: injector capillary; CC: cooling chambers in mantle; CO: gasketed opening for insertion of collector; COL: collector; CT: collector tubing; Wm: gasketed pug-in quartz window fitting into opening WO; CW: cooling water tubes; the arrows indicate the path of cooling water through these tubes and cooling chambers CC. The construction IS the same on thc back side of the mantle, which is not shown. BC: buffer chambers; OT,, OT,: openings for passage of magnet tunnel TU shown in Fig. 12.18B; the opening OT, is sealed by the plate TP of the tunnel of Fig. 12.18B; 0 , - 0 , : openings in the outer walls of the buffer chambers against which the four perforated gasketed plates P are pressed; CP: cooling pipe conveying coolant to core C; V: vent for removal of air bubbles from core C; E: electrodes; EC: electrode compartments; EP: electrode connection posts; B: four balconies; D: drainage tubes; d: continuation of drainage tube; N: terminal nipple of drainage tube for detachment of tubes, dt, shown in Fig. 12.19; the balconies are shown in place o n the left side of the cell and removed on the right; G : gasket; CB: anticonvection baffles; MG: magnet.

282

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fig. 12.20. Serpentine flow channel 1211. The buffer is transferred from left to right as it meanders through the channel. Engendered vortices are inverted relative to the gravitational field vector $as they are transferred from one vertical channel t o the adjacent one.

apparatus (“free-flow’’ electrophoresis), the serpentine flow channel must be closed off laterally by membranes with the consequent above-mentioned shifts in pH and conductivity in the separation space. These shifts in pH and conductivity may not be objectionable, however, if the system were used for continuous flow isoelectric focusing where a pH gradient must be maintained in the cell and the long migration path (in a small space) will provide the required long residence time of ampholytes to be separated in the electromigration column [25]. Examples of preparative separations by continuous flow deviation electrophoresis In order to illustrate the capabilities and effectiveness of the methods of deviation electrophoresis, we shall present some illustrative examples of separations. We shall not tabulate here the variety of buffers that have been used, as the buffer composition can be found in the literature cited. The buffers used in molecular separations differ from those employed in separations of cells notably in the use of solutes, such as sucrose, which must be used in the latter application at appropriate concentrations to protect the cells from osmotic damage. Other considerations in designing buffer solutions for cell separations are optimization of the viability of the cells and avoidance of cell aggregation. A detailed consideration of buffers suitable for cell separations by deviation electrophoresis has been published [26].

Molecular mixtures As the resolving power of deviation electrophoresis is inferior to that of methods in which chromatographic and filtration effects are combined with electrophoresis, relatively few uses have been made of it for molecular analysis. Nevertheless, the ease of continuous collection of separated fractions may sometimes outweigh the drawback of lower resolution. Fig. 12.14 presents an example of the separation and collection of lowmolecular-weight solutes (dyes) by circular endless belt electrophoresis [7]. Whereas separations in the endless belt are usually exhibited as photographs of separated streaks,

ENDLESS FLUID BELT ELECTROPHORESIS

283

the separations obtained in the flowing curtain apparatus are given in terms of the evaluation of the collected fractions. Thus, Fig. 12.22 shows the collection pattern resulting from curtain flow electrophoresis of a crude peptide fraction precipitated with acetone from a nucleosol of Yoshida ascites carcinoma cells [27]. Fig. 12.23 shows four consecutive photometric tracings obtained after 2 0 , 2 8 , 5 0 and 70 s, respectively, by curtain flow electrophoresis of serum proteins [20]. The diameter of the injected streak was about 0.05 mm and the injected sample volume was 0.1-0.3 pl, containing 3-1Opg of protein. Fig. 12.24 shows an interesting group of tracings obtained by the same instrument [20] using pathological sera.

Viruses Relatively little work has been done on virus separations by continuous flow deviation electrophoresis. The separation of strains of fd3 bacteriophages by flowing curtain electrophoresis has been reported [28]. In this study, it was possible to isolate from the main component two mutants present in minute amounts, which differed from it by one and two electric charges, respectively, in their protein coats. Fig. 12.25 shows the rapid separation of a mixture of two strains of tobacco mosaic virus, U1 and U2, into two widely spaced streaks achieved by endless belt electrophoresis [ 181. The mixture contained about 2% of the U2 and 0.5% of the U1 fraction. The picture shown was obtained by ultraviolet light photography. The scarce component, U1, is the faster and appears on the left because, as in all endless belt photographs shown here, the electromigration is from right to left. The streaks corresponding to the two virus components of the mixture are clearly separated (top arrows) after travelling 4 cm downwards from the injector (not shown), that is, after about 8 s residence in the electric field. The two streaks on the left (lower arrows) show how far the separation has increased in the upward ascending streaks after an additional half-revolution about the iron core (about 20 s). Micro-organisms We saw in Fig. 12.1 3 an example of a separation of two species of yeast, Rhodotorula and Saccharomyces cerevisiae,by circular endless belt electrophoresis [7]. A photograph of a similar separation in the non-circular endless belt apparatus of bacteria (Escherichia coli) and a species of yeast (Rhodotomla)is given in Fig. 12.26, accompanied by a histogram which shows that there is no overlap between the fractions in the collector [18]. An interesting observation that can be made with a number of pure strains of bacteria is shown in Fig. 12.27. A streak of a pure strain of Escherichiafreundii is shown t o be split into two electrophoretically distinct and widely separated components [29]. Similar observations have also been made on some strains of Escherichia coli [30]. Electron microscopy revealed that the two fractions into which the pure bacterial strains were split represented a piliated and non-piliated fraction. Cultures of each of these fractions gave rise to offspring that again consisted of a mixture of piliated and non-piliated bateria exhibiting the same electrophoretic separation pattern. Preparative separations of bacteria differing in their Gram-staining characteristics and resistance to antibiotics by means of flowing curtain electrophoresis have also been reported [31].

284

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

R

Fig. 12.21. (A) Perspective view of serpentine path continuous flow deviation electrophoresis cell [ 31. (B) Simultaneous separation of two streaks of a mixture of three dyes (injected a t the bottom) through two separate injectors in the serpentine flow cell. The buffer flow is dkected upward. Streaks in the two separation patterns (from left to right): Evans Blue, Rose Bengal and “Brush” green recording ink. (C) Collection pattern for the separated components of the two sample streaks [3]. The sequence of colours is the same as in B. Evans Blue appears in test-tubes 2,3,16 and 17; Rose Bengal in 6,7,18,20 and 21 and the green dye in tubes 9,10,11,23,24 and 25. Tubes 8 and 22 contain a dilute mixture of Rose Bengal with the green dye. The collected fractions are very dilute in tubes 6, 8, 11, 18 and 25. Test-tubes 4 and 19 are empty.

Subcellular particles Most of the work on the separation of subcellular particles has been performed with the flowing curtain apparatus developed by Hannig and his numerous co-workers at the Max-Planck Institute of Biochemistry, Martinsried, near Munich, G.F.R. We shall refer to Hannig’s review papers [31,32] for complete surveys of the activities in this field and limit ourselves here, within the scope of our methodological objectives, to a few representative illustrations.

ENDLESS FLUID BELT ELECTROPHORESIS

285

286

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

2.5

t

0

m

1.5

N

w

b 0 N

w

0.5

20

40 Froct No

-

60

80

Fig. 12.22. Flat curtain flow electrophoresis o f a crude peptide fraction precipitated with acetone from the nucleosol of Yoshida ascites carcinoma cells [27].

rnobility~105~cm volt“ 2~

sec-’

Fig. 12.23. Separation of serum proteins by the analytical fluid curtain apparatus in pH 8.8 Tris-borate buffer [20]. Separation times: (a) 20 s; (b) 28 s; (c) 50s; (d) 7 0 s. The overlapping of the components diminishes with separation time.

Of particular interest for fundamental biological investigations and for the study of biological cell membranes in particular are the successes in generating and separating b y deviation electrophoresis artificially created vesicles surrounded by “insideaut” and “outside-out’’ membranes [33]. It proved possible to detach from erythrocytes small protuberances as isolated vesicles. These particles have the same outer surface as the erythrocytes carrying the sialic acid. They migrate electrophoretically with the same mobility as the erythrocytes. On the other hand, it also proved possible to detach from erythrocytes as isolated vesicles tiny invaginations formed in their surface. These vesicles carry the inside membrane surface of the erythrocytes on their outside, i.e., the siahc acid layer is now on the inside [34,35].Electrophoretic analysis in the flowing curtain apparatus showed that these “inside-out’’ vesicles have a smaller electric surface charge density than the “outside-out’’ vesicles, as indicated by their slower electromigration rate. Fig. 12.28 shows the distribution of vesicles among the fractions obtained by the flowing curtain apparatus from a mixture of both types of vesicles [34]. It turned out that such membrane inversions, with consequent creation of insideaut

ENDLESS FLUID BELT ELECTROPHORESIS 4orbus- Waldenstrorn Vb (121)-37.6

I Nephrotic Alb(

287

syndrome

al) - 43.0

-

16.6

- 12.4 - 29.2

IgD - Plasmoc ytoma

Liver cirrhosis

41b (UI) - 41.8 a2 - 15.8 PI 11.4 /32 - 6.0 71 - 14.3 7, -107

Alb(a1)- 21.2

-

? 7

-

7.7

-22.4 -48.7

Fig. 12.24. Separations of pathological sera within 42 s with the analytical h i d curtain apparatus [ 201.

Fig. 12.25. Separationof two strains, U1 and U2 (slower component), of tobacco mosaic virus in the endless fluid belt apparatus: after about 8 s (arrows at top right) and after about 20 s (arrows at bottom left) [ 181.

vesicles, were also possible with other biological entities such as E. coli bacteria and mitochondria whose inner membrane surface thus became accessible to investigation [36,37]. Of no lesser interest is the demonstration of the possibility of electrophoretic isolation of undamaged subcellular cell components, such as renin granules from rabbit kidney cortex [38], rat liver lysosomes [39] and rat liver ribosomes [40].

Cells Electrophoresis of somatic cells presents considerable experimental difficulties that are

288

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fig. 12.26. Separation of bacteria and yeast cells: Escherichiu coli (faster component on the left) and Rhodotomla (slower component) [ 181. A: separation pattern; B: collection histogram.

Fig. 12.27. Splitting of Escherichiufreundii into fast (F) and slow ( S ) components observed with the endless fluid belt apparatus [ZS].F and S, F' and S' and F" and S" denote consecutive half-turns of the helix.

ENDLESS FLUID BELT ELECTROPHORESIS

289

I

Outside-out

6001

I

Inside-out

A

15

20

25

30

Fraction number

Fig. 12.28. Separation in the fluid curtain apparatus of “inside-out” and “outside-out” erythrocyte vesicles [ 34). Left peak: vesicles carrying sialic acid on the outside surface (as the intact erythrocytes). Right peak: a pure fraction of vesicles carrying the sialic acid on the inner surface.

not inherent in the electrophoretic process. They are based on the sensitivities of the cells to their chemical environment, the osmotic pressure of the ambient solution and the presence of factors that could affect the viability of the cell. Clumping of the cells, which accelerates sedimentation, should be carefully avoided. It is advisable to provide the instrumental parts t o which the cells could adhere (such as the collection tubing) with a special coating. Special buffers should be used to ensure the survival of the cells during the separation process. All of these factors have been reviewed in detail in connection with flowing curtain electrophoresis [26]. In the endless belt apparatus, cell losses due to cell adhesion in the collector system (including the collection tubes) vary greatly from one species of cell to another. Thus, with erythrocytes the cell recovery may be as high as 98%, while with lymphoid cells it may be below 50%. The following illustrations show cell separations in the non-circular endless belt apparatus. Fig. 12.29A and B shows two histograms for collection patterns of human blood cells from two hospital patients [18]. Fig. 12.30 shows an unusual separation pattern in which the erythrocyte streak splits into two distinct cell components. This blood sample

290

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

0 Erythrocytes 200 Granulocytes

Leukocytes

0 Erythrocytes

8

Fraction number

9

10

Fraction number

12

Fig. 12.29. Separation of human erythrocytes from leucocytes in the endless fluid belt apparatus [ 181. A : Histogram for collection pattern from a hospital patient. B: Histogram for collection pattern obtained from a different patient.

Fig. 12.30. Splitting of a donor’s erythrocytes into two electrophoretically distinct components as observed with the endless fluid belt apparatus [41].

was obtained from a donor who exhibited partial polyagglutinability [41]. Fig. 12.31 shows the distribution pattern in a collection of murine mesenteric node lymphocytes subjected to endless belt electrophoresis. This figure illustrates how the resolution is improved by reducing the injection rate, thus making the streak thinner, which confines the particles in it more to the central region of the annulus. Fig. 12.32A shows a photograph of a separation pattern of a contaminated suspension of human lymphocytes [5]. Fig. 12.32B shows the distribution of the lymphocytes and contaminants in the collection pattern, and Fig. 12.32C shows the distribution of E- and EA-rosette-forming cells in a lymphocyte collection pattern in which enrichment of B and T lymphocytes could be obtained in different fractions. The remaining illustrations were obtained with the fluid curtain apparatus. Fig. 12.33 shows a photograph of the separation path of a mixture of human and rabbit erythrocytes escaping at the top right from the meeting point of the separated streaks [42].

29 1

ENDLESS FLUID BELT ELECTROPHORESIS M u r t n e m e ~ e n t e r i cl y m p h node cells

0 P l o8/

30

M= High in) r a t e ( 1 0 0 ) e-e=tow inj r a t e ( 4 0 )

25 20 15 10 5 F r a c t i o n number - migration

0

Fig. 12.31. Separation of murine mesenteric lymph node cells with the endless fluid belt apparatus (51. The solid line represents a collection pattern obtained at high sample injection rate; the broken line, showing higher resolution, was obtained at a low injection rate.

This instrument was applied to investigations of significant biological problems and one of the important recent results is the isolation of a single-cell population by Heidrich and Dew [43] from a mixed-cell suspension derived from rabbit's kidney cortex. In this work it proved possible to isolate proximal and distal tubules as well as renin-active cells. Of particular importance is the work on the fractionation of lymphocytes into immunologically distinct sub-populations that has been carried out systematically with the fluid curtain apparatus. A complete bibliography can be found elsewhere [26] and we shall limit ourselves to the consideration of illustrative examples. Fig. 12.33 shows curves that represent the overall leucocyte distribution in a sample of human w h t e blood cells subjected to electrophoretic analysis and the distribution curves for granulocytes, lymphocytes, eosinophils and monocytes inferred from an analysis of the collected fractions [3 11 . Fig. 12.34 shows a trimodal electrophoretic distribution profile for murine spleen cells [44]. The electroinigration is towards the left. The peak of the fraction of highest mobility corresponds t o T cells, whereas the two peaks of lower mobility (to the right of it) are formed by surface immunoglobin-carrying lymphocytes (Ig') and other types of cells. The legend indicates the cell sub-populations as established by physical and imniunological tests on the collected fractions. These examples suffice to demonstrate that continuous flow deviation electrophoresis is a powerful technique for fundamental investigations in cell biology and has potential practical value in medical research and diagnosis. Measurement of electrophoretic mobility The principles according to which electrophoretic mobilities can be measured with a continuous flow deviation electrophoresis apparatus are essentially the same for fluid curtain electrophoresis as for endless fluid belt electrophoresis. In the former, the slope of a streak against the vertical is a measure of the electrophoretic mobility of the streak components, whereas in the latter the pitch of the helix (i.e., the distance between

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

292

Fraction number - migration

Fig. 12.32. Separations of suspensions of human lymphocytes with contaminants by means of the endless fluid belt apparatus. The lymphocytes were prepared by the Ficoll-Isopaque interface centrifugation method [ S 1. (A) Photograph of separation pattern. (B) Collection patterns of lymphocytes and contaminants. (C) Distributions of the E-rosette- and EA-rosette-forming cells in a lymphocyte collection pattern obtained with the endless fluid belt apparatus. The vertical scale on the right applies to the curve drawn through the circles (0).The percentage scale on the left applies to the curves drawn through the crosses (x) and squares (B). WBC = White blood cells. (Graphs (B) and (C) have been obtained by Dr. R.C. Seeger and MI. J. Rosenblatt in collaborative experiments).

adjacent turns) measures the electrophoretic mobility. This statement is correct, however, only in the absence of electroosmotic streaming, which in fact is normally present. Thus, owing to electroosmotic convection even electrically neutral particles will deviate from a zero-deviation path. We must therefore establish a baseline, i.e., the path of an electrically neutral particle against which the deviation of streaks of charged particles will be measured [5, 181. The neutral dye ApolIon (Microchemical Specialities, Berkeley, Calif,, U.S.A.) can serve this purpose.

293

ENDLESS FLUID BELT ELECTROPHORESIS

Granulocytes

0

0

Lymphocytes

30

40

35

45

50

Fraction number

Fig. 12.33. Separation of human leucocytes by the fluid curtain apparatus in phosphate buffer (pH 7.3) with a field intensity o f 80V/cm [ 311. LJ--o--o: Overall distribution curve; -: granulocytes; . lymphocytes; -:eosinophile granulocytes; - - -: rnonocytes.

-._.-.

--

We shall now describe how absolute mobilities can be measured with the endless fluid belt apparatus. An analogous procedure could be used with the fluid curtain apparatus. To measure absolute mobilities, we must know the intensity of the electric field and the distance traversed in the field direction by the particles under observation against the zero-mobility reference streak in a known time. The electric field can be found as the ratio of the voltage between two points, such as between the metal tube, S , and a straight metal wire inserted through the capillary, Cap, in Fig. 12.8. The migration time can be chosen to equal one buffer revolution time in the annulus. This can be timed by injecting a pulse of dye through the injector, IN. The separation between the particles under test (mixed with Appollon dye) and the Apollon zero-mobility reference streak is measured after one turn. From these measurements, we can calculate the horizontal distance covered in 1 s due to electrornigration by the observed particles in a field of 1 V/cm, i.e., their mobility, p . Usually, it is more convenient to make relative mobility measurements, for which we require a standard of known mobility. The dye brilliant blue (K & K Labs., Hollywood,

294

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Fraction ( mrn) 380 3 5 8 335 313 29.1 2 6 8 2 4 6 Electrophoretic migration path

224

201

Fig. 12.34. Electrophoretic distribution profie of spleen cells from conventionally raised C3D2F1 hybrid mice [44]. - - -: All nucleated cells; -A-A-: O* lymphocytes; - 0 - 0 -: all Ig+ lymphocytes; G-@-: ring-like fluorescing Ig+ lymphocytes; --o--o-: spot-like fluorescing Ig' lymphocytes.

- -

Calif., U.S.A.) is suitable for this purpose. It preserves a constant ionic charge in the pH range 3.3-9.3. If we inject brilliant blue of mobility p B and Apollon together with a particle, U, of unknown mobility, we obtain a triple streak. If we consider a horizontal line intersecting with these streaks at some distance from the injector at the points A, B and U, the horizontal distance between the Apollon and brilliant blue streaks will then be AB and the distance between Apollon and the streak of unknown mobility will be AU. The unknown mobility, p u , is then given by

! % = -AU PB

AB

(12.14)

The electrophoretic mobility of a species of ions or biological particles depends on the nature of the ambient buffer solution, the chemical composition, ionic strength and pH of which must be specificed in order t o make the given mobility value meaningful. There is, however, another important parameter that must also be specified, viz., the temperature at which the mobility has been measured. This is a less simple matter as there is a temperature gradient within the fluid belt or curtain with a temperature maximum in the mid-plane between the walls. This problem has been solved by using brilliant blue (whose streak is centred in the annulus) as a thermometer, as it were [ 591. If we tabulate the absolute mobility of brilliant blue as a function of temperature, we can determine the temperature at the centre of the annulus from the deviation of brilliant blue streak in a known electric field [5,18,19]. As the material of unknown mobility lies in the same fluid layer as the brilliant blue streak, its mobility has thus been determined for the temperature indicated by the absolute mobility of brilliant blue. In this connection, we recommend the use of the newly proposed Tiselius unit of electrophoretic mobihty (1 TU = 1O-' cmz s-' V-' = 10-9m2 s-' V-'), w h c h has been introduced as an analogue to the Swedberg unit used in sedimentation [I].

CONCLUSION

295

CONCLUSION Continuous flow deviation electrophoresis can be conducted without recourse to porous media or similar means of stabilization against thermal convection in flow channels of different shapes. We have considered examples of a flat curtain, a serpentine path and circulation in an endless belt about a horizontal axis. In the last two cases, additional stabilization is achieved by periodic inversion of nascent thermal convection vortices. An additional advantage of the latter systems is the possibility of compressing a long migration path into a small space. The endless belt system possesses an additional, unique advantage, viz., the use of a magnetic field for generation of buffer flow. The smoothness and constancy of the flow produced by the interaction of the electrophoretic current with the constant field of a permanent magnet is not the rnain advantage of this system, which obviates the need for a complex pump for this purpose. The principal advantage of this magnetohydrodynamic drive is the possibility of onutting lateral confines in the flow channel, which thus becomes a sandwich with the fluid contained between two parallel dielectric walls. Thus, membranes, which are indispensable in the flat and serpentine fluid curtain apparatus, can be omitted. This omission not only leads to simpler operation and maintenance, but is also a very significant improvement because omission of membranes eliminates membrane-induced temporal and spatial shifts in pH and electrical conductivity in the separation space and permits convective streak colliniation 1451. The flat fluid curtain system has received the greatest amount of work in its development and has been put to many uses in research, not only in the laboratory of its origin in Munich but also, as a result of its commercial availability, in many other laboratories. The diverse uses and applications of this apparatus are apparent from the bibliography. The endless belt apparatus, however, is not yet available commercially and has been confiied, owing to the complexity of its construction, to instruments emanating from one laboratory. It has therefore been used by only a few workers in few investigations. The development of the fluid curtain and the endless belt apparatus has proceeded concurrently since about 1960, and not without mutual benefits, which can be seen from the chronology of new improvements although not always from the references. It is to be expected that this symbiotic relationship between these methods will continue in the future. The examples shown not only illustrate the great potential of these methods as powerful research tools, but also suggest the possibility of diagnostic uses in medicine and therapeutic uses through preparative separations on a large scale of components which could be administered to patients.

ACKNOWLEDGEMENTS Our work o n cell separations has been greatly aided by the advice and cooperation of Dr. Charles W.Boone, Head of the Cell Biology Laboratory of the National Cancer Institute. In the separations of human B and T lymphocytes we enjoyed the cooperation of Dr. Robert Seeger of the UCLA Department of Pediatrics and of his assistant Mr. Joseph

29 6

CONTINUOUS FLOW DEVIATION ELECTROPHORESIS

Rosenblatt (cf. Fig. 32C). We received the greatly appreciated similar cooperation of Dr. Benjamin Bonavida of the UCLA Department of Medical Microbiology and Immunology in work aiming at separation of murine B and T lymphocytes. Last, but not least, I am indebted to Professor K. Hannig and Dr. H. G . Heidrich of the Max Planck Institute of Biochemistry in Martinsried near Munich for having so kindly provided numerous illustrations on the fluid curtain apparatus and its uses.

SYMBOLS AND UNITS E B J

Electric field intensity in V/cm Magnetic field intensity in gauss Electric current density in A/cm* Electric conductivity in (ohm cm)-' U Electrophoretic mobility in Tiselius units: TU (1 TU = (cm/s)/(v/cm)) [ 11 P W Electroosmotic mobility in TU 7) Fluid viscosity in poise U Fluid curtain velocity in cm/s Central maximum velocity in cm/s uo Ve Electrophoretic velocity in cm/s Electroosmotic velocity at the wall in cm/s v, Velocity o f a particle due to electro-osmosis and electrophoresis in cm/s u* X, Y, Z Coordinate axes Q! Angle of deviation in degrees Z Distance of a point from center of fluid ribbon in cm h Thickness of fluid ribbon in cm d Internal diameter of injector tube, or initial thickness of injected streak in cm R Resolving power Vertical force density (forcelunit volume) in dynes/cm3 fx Horizontal force density (forcelunit volume) in dynes/cm3 f Y

REFERENCES 1 2

N. Catsimpoolas, S. Hjertkn, A. Kolin and J. Porath, Nature (London), 259 (1976) 264. K. Hannig, tl. Wirth, B. H. Meyer and K. Zeiller, Hoppe-Seyler's 2.Physiol. Chem., 356 (1975)

3 4 5

A. Kolin and P. Cox, Proc. Nat. Acad. Sci. US.,52 (1964) 19. A. Kolin,Proc. Nat. Acud. Sci. US.,56 (1966) 1051. A. Kolin, in N. Catsirnpoolas (Editor),Methodsof Cell Separation, Vol. I, Plenum, Ncw York,

6 7 8

A. Kolin, Proc. Not. Acad, Sci. U S . , 46 (1960) 509. A. Kolin,J. Chromatogr., 26 (1967) 164.

1209.

1979, pp. 93-180.

A. Strickler and T. Sacks, in N. Catsimpoolas (Editor), Isoelecmk Focusingand Isorachophoresis, New York Academy of Sciences, New York, 1973, pp. 497-514. 9 H. L. Dryden, I;. P. Murnaghan and H. Bateman, in Hydrodynamics, Dover, New York, 1956, p. 184. 10 H. Lamb, in Hydrodynamics, Dover, New York, 1945, p. 582. 11 M. Smoluchowski, in L. Graetz (Editor), Handbuch der Eiektrizitat und Magnetismus, Vol. 2 , Barth, Leipzig, 1921, p. 366.

REFERENCES

297

12 C. C. Brinton and M. A. Lauffer, in M. Bier (Editor), Electrophoresis, Vol. 1, Academic Press, New York, 1959, pp. 428487. 13 H. Svensson and I. Brattsten, Ark. Kemi, 1 (1949) 401. 14 W. Grassman and K. Hannig, Naturwissenschaften, 37 (1950) 496. 15 J . Barrollier, E. Watzke and H. Gibian, 2. Naturforsch. B , 13 (1958) 754. 16 K. Hannig, Z. Anal. Chem., 181 (1961) 244. 17 K. Hannig, Hoppe-Seyler’s Z. Physiol. Chem., 338 (1964) 211. 18 A. Kolin and S. J. Luner, Anal. Biochem. ,30 (1969) 111. 19 S. J. Luner, Thesis, University of California, Los Angeles, 1969. 20 K. Hannig, H. Wirth, R. K. Schindler and K. Spiegel, Hoppe-Seyler’s 2. Physiol. a e m . , 358 (1977) 753. 21 A.Kolin,Proc. Nut. Acad. Sci. U.S.,51 (1964) 1110. 22 H. Lamb, in Hydrodynamics, Dover, New York, 1945, p. 582. 23 H. L. Dryden, F. P. Murnaghan and H. Bateman, in Hydrodynamics, Dover, New York, 1956, p. 184. 24 B. Bergrahm, Sci. Tools, 14 (1967) 34. 25 T. S. Fawcett, in N. Catsimpoolas (Editor), Isoelectric Focusing and Isotachophoresis, New York Academy of Sciences, New York 1973, pp. 209 and 112-126. 26 K. Zeiller, R. Loser, H. Pascher and K. Hannig, Hoppe-Seyler’s Z. Physiol. Chem., 356 (1975) 1225. 27 P. Spitzauer, A. Schweiger and K. Hannig, Hoppe-Seyler’s Z. Physiol. Chem., 354 (1973) 1327. 28 G. Braunizer, G. Hobom and K. Hannig, Hoppe-Seylerk Z. Physiol. Chem., 338 (1964) 278. 29 A. Kolin, Proc. 1st European Biophysics Congress, Baden, Austria, 1971, p. 481. 30 R. M. Owen, Thesis, Calif. State College, Long Beach, Calif., 1972. 31 K. Hannig, in D. Click and R. M. Rosenbaum (Editors), Techniques of Biochemical and Biophysical Morphology, Wiley-Interscience, New York, 1972, p. 21 1. 32 K. Hannig, Mitt. Max-Planck Ges., p. 185. 33 H. G. Heidrich, Mitt. Max-Planck Ges., 6 (1974) 464. 34 H. G. Heidrich and G. Leutner, Eur. J. Biochem., 41 (1974) 37. 35 G. Leutner and H. G. Heidrich, The mechanism of “inside-out” vesicle formation (in preparation). 36 H. G. Heidrich, R. Stahn and K. Hannig, J. Cell Biol., 46 (1970) 137. 37 H. G. Heidrich and W. L. Olsen, J. Cell Biol., 67 (1975) 444. 38 M. E. Dew and H. G. Heidrich, Hoppe-Seyler’s Z. Physiol. Chem., 356 (1975) 621. 39 K. Hannig, R. Stahn and K. P. Maier, Hoppe-Seyler’s Z. Physiol. Chem., 350 (1969) 784. 40 A. Schweiger and K. Hannjg, Hoppe-Seyler’s 2. Physiol. Chem., 348 (1967) 1005. 41 P. Sturgeon, A. Kolin, K. S. Kwak and S. J . Luner, Haematologia, 6 (1972) 93. 42 M. Ganser, K. Hannig, W. F. Kriismann, G. Pascher and G. Ruhenstroth-Bauer, Klin. Wochenschr., 46 (1968) 809. 43 H. G. Heidrich and M. E. Dew,J. Cell Biol., 74 (1977) 780. 44 K. Zeiller, G. Pascher and K. Hannig, Immunology, 31 (1974) 863. 45 A. Kolin, in P. G. Righetti, C. J. van Oss and J. W. Vanderhoff (Editors), Electrokinetic Separation Methods, Elsevier/North-Holland Biomedical Press, Amsterdam, New York, Oxford, 1979, p. 169.

This Page Intentionally Left Blank

Chapter 13

Preparative electrophoresis in gel media Z. HRKAL

CONTENTS Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel composition and concentration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gel dimensions and sample load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Voltage gradient and the electrophoresis time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Duration of electrophoresis and flow-rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elution by diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrophoretic elution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Continuous separation and elution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prcparative gel electrophoresis with discontinuous elution. . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

299 299 300 300 30 1 30 I 30 1 302 304 305

INTRODUCTION The excellent resolution and simplicity of analytical gel electrophoresis led to attempts to scale up the analytical procedure and to attain a high resolution of the components of macromolecular species and their isolation by gel electrophoresis on a preparative scale. It was soon realized that a simple increase in analytical gel electrophoresis in rod shaped gels does not yield the desired results, giving low recoveries and inadequate purity of sample components. Numerous designs of special apparatus for large-scale gel electrophoresis have been published during the last 15 years [ 1-20]. However, polyacrylamide gel electrophoresis has found little acceptance in biochemistry as a preparative technique and few significant isolations have been described that would have not been possible by means of conventional fractionation procedures. It is to be hoped that an understanding of the technical problems and the variables that influence the separation of macro-ionic mixtures on a preparative scale will lead t o an increase in the popularity of this valuable technique. The following parameters have to be carefully selected in order to achieve optimal resolution and recovery in preparative applications of gel electrophoresis: gel composition and concentration, gel dimensions, sample load, voltage gradient, duration of electrophoresis, and flow-rate.

GEL COMPOSITION AND CONCENTRATION

Two kinds of media are employed for gel electrophoresis on a preparative scale, namely polyacrylamide and agarose gels, of which polyacryiamide is by far the most frequently

300

PREPARATIVE ELECTROPHORESIS IN GEL MEDIA

used. The gels and buffer systems are those described for analytical gel electrophoresis in Chapters 4 and 5, The gel solution is degassed in order to avoid air bubbles, then poured into a gel mould and quickly overlayered with water. The gel mixture is polymerized for 1 h at the temperature at which the electrophoretic separation is to be performed. It is important not to disturb the gel solution by touching it, etc., during polymerization, in order to obtain a flat gel surface. The optimal gel concentration should be found empirically with the analysed mixture using analytical gel electrophoresis at several gel concentrations. By plotting log R F against gel concentration (Ferguson plot) for the components of the system to be resolved, the gel concentration at the maximal separation can be determined as 0.368RF (ints), where RF ( i t s ) is the relative mobility at the point of intersection of the Ferguson plot 1211. For the separation of proteins within the molecular weight range 104-105 an acrylamide gel concentration of 7% is commonly used.

GEL DIMENSIONS AND SAMPLE LOAD Joule heat produced in electrophoretic experiments does not permit a thickness of the acrylamide gel layer of greater than 15 mm to be used, otherwise the heat cannot be evenly dissipated and curvature of migrating bands occurs. For gels of circular geometry this leads to a maximal cross-sectional area of 1.8 cm2. As the permittable sample load is proportional to the gel cross-sectional area, it is desirable to employ large gel surface areas. To achieve this, Jovin et al. [ 3 ]introduced an annular gel geometry; the gel is moulded in the form of a hollow cylinder of thickness 1.5 cm, the outer face being cooled with a cooling jacket and the inside with help of a cooling finger. With this arrangement, the cross-sectional area can be increased to 15 cm2. The sample load depends markedly o n the migration properties of the components to be separated. For two adjacent bands that are separated by one bandwidth [21], a maximum of 1-2mg/cm2 can be applied. The greater the distance between the components, the larger are the sample loads that can be employed, Hence the term preparative electrophoresis may relate to sample loads of 1, 10 or even 100mg/cmz for the same apparatus and gel, depending on the properties of the species to be separated. As the recovery is directly proportional to the sample load, it is advantageous to employ as high a sample load as can be resolved with a given system. With a sample size of less than 1 mg/cm2 a recovery as low as 30%is not unexpected.

VOLTAGE GRADIENT AND THE ELECTROPHORESIS TIME The electrophoresis time is inversely proportional to the voltage gradient and therefore the use of high voltage gradients would be preferable. However, the Joule heat limits the maximal voltage t o ca. 10 V/cm in the separation gel. The temperature of gel also influences the resolution, as with the increasing temperature the electrophoretic mobility of the macro-ions increases, with resulting shortening of the electrophoretic time and a decrease in zone spreading. If there is n o risk of sample denaturation or inactivation, the electrophoresis should be performed at 25°C.

DURATION O F ELECTROPHORESIS AND FLOW-RATE

30 1

DURATION OF ELECTROPHORESIS AND FLOW-RATE

A convenient flow-rate corresponding to one exchange of elution chamber volume per minute is usually employed, which leads to an elution time of about 5 h for 6-cm gels. As with continuing electrophoresis the slower migrating bands would be eluted in substantially higher buffer volumes than the rapidly migrating bands of the same thickness, it is of advantage t o employ a gradual deceleration of the elution buffer flow. This can be achieved by controlling the elution pump speed with an electronic system [22] or with a low-speed motor-driven cam [23]. In recent years, two distinct methods of preparative gel electrophoresis have been developed: (1) In the first method, the protein (or other macro-ion) mixture is first separated by electrophoretic migration on an acrylamide gel; the zones are then localized, cut out, minced and the material either eluted or released from the gel by electrophoretic migration. This procedure is especially useful for coloured macro-ions (e.g., haemoproteins). ( 2 ) In the second procedure, compounds that are separated on a gel column are allowed to migrate off the end of the gel into an elution chamber and from there eluted by a buffer flow into a W monitor and further to a fraction collector. The elution buffer may be delivered either continuously or discontinuously (intermittent elution). Each of these methods has advantages and drawbacks determined by the relative number and proportions of fractions, size of the components and their mobility and diffusion rate. Elution by diffusion The gel is removed from the gel tube by injection of water from a syringe and the portions containing the desired fractions are separated with the aid of a string and minced with scissors. Exhaustive elution is then performed with a suitable buffer, the eluate separated by centrifugation and the procedure repeated three times. The supernatants are pooled and concentrated on an Amicon ultrafiltration cell using Diaflo UM-I0 membranes. This procedure is time consuming and gives low yields. Electrophoretic elution The separated sample components are localized, cut out and minced. The open tube, 1.2 cm in diameter, is closed with a dialysis menibrance and 3 ml of 0.025 M Tris-HC1 buffer (pH 8.0) containing 6% of sucrose is applied to the bottom of the tube. The solution is overlayered with the gel solution (2.17 ml of 1 M HJ’03, 0.17 g of Tris, 0.035 ml of tetraethylmethylenediamine (TEMED), 2.5 g of acrylamide, 0.63 g of bisacrylamide, 0.5 mg of riboflavin and water t o 100 nil) to form a gel ca. lOmm high. The gel mixture is photopolymerized for 1 h. The minced gel layer containing the desired component is packed on the gel layer and covered with Tris-HC1 buffer (pH 8.0) containing 6% of sucrose. The tube is filled with the same buffer without sucrose and the electroelution performed with Tris-glycine buffer (pH 8.9) (Tris 6.4g, glycine 4.0g, water to 1 1) in the

302

PREPARATIVE ELECTROPHORESIS IN GEL MEDIA

cathodic reservoir and 0.1 M Tris-HC1 buffer (pH 8.15) in the anodic reservoir. Electrophoretic elution is performed in the preparative electrophoresis apparatus for 30-60 min. After this period, the protein fraction is quantitatively eluted in the buffer compartment above the d d y s i s membrane [24]. Continuous separation and elution A number of instruments for the continuous separation of the components of macromolecular mixtures by means of preparative electrophoresis on acrylamide gel have been described, some of them being commercially available (LKB, Canalco). There are two critical points that have t o be resolved in order for a satisfactorily functioning preparative gel electrophoresis apparatus to be obtained: (a) adequate cooling of the gel layer to prevent a temperature gradient across the gel, leading t o band curvature; (b) construction of the elution chamber so as to ensure rapid elution of the migrating components without excessive dilution. Dissipation of the Joule heat is mostly effected by adopting the Jovin et al. [3] system for cooling both the outer and inner surfaces of hollow cylindrical gels. The design of the elution system is of the utmost importance for successful separations in a continuous arrangement. Nees [25] designed and evaluated three constructions of the elution chamber, one of circular and two of rectangular shape. In the circular cross-section design, an elution chamber with the four inlets and the eluent outlet in the centre gave the best flow characteristics (sharp decrease in protein concentration with eluent volume). Even better results were obtained with an elution chamber of rectangular cross-section with air bubbles being introduced into the elution system at regular intervals. Nees [25] described the construction of apparatus with a rectangular elution chamber for preparative electrophoresis on a gel slab. An additional advantage of this construction is good heat dissipation from the slab-shaped gels. Smaller devices, with a cross-sectional area of 5-15 cm2, for the separation of milligram amounts of protein mixtures are mostly based on a design of Jovin et al. [3].Acrylamide gel in the shape of a hollow cylinder is prepared in a cylindrical glass column with a removable plastic floor. The outer jacket and inner cooling finger provide cooling during polymerization and electrophoresis. For the separation, the plastic bottom is replaced with an elution chamber consisting of a base formed by a porous glass filter and covered with a dialysis membrane. Elution is symmetrical with a radial flow of elution buffer from the circumference to a central capillary in the cold finger. The upper and lower electrode compartments are provided with circular platinum electrodes (Fig. 13.I). The basic gel and buffer system suggested by Jovin et al. [3] is as follows:

Separation gel (pH 8.9) Stock solutions: A. Acrylamide 30 g Bisacrylamide 0.8 g H20 to l0Oml

DURATION OF ELECTROPHORESIS AND FLOW-RATE

303

Fig. 13.1. (A) Preparative gel electrophoresis apparatus with continuous elution. 1, 2: Platinum electrodes; 3: upper buffer compartment; 4: lower buffer compartment; 5: separation gel; 6: cold finger with the central capillary; 7: cooling jacket; 8: glass membrane; 9: elution chamber. (B) Design of elution chamber.

24.0 ml 18.15g 0.23 ml to 100ml H20 C. Ammonium persulphate 0.14g HZO to l 0 0 m l Separation gel composition (7.5%) (volume parts): A 1 , B 1 , C 2

B.

HCl(lil4) Tris TEMED

Spacer gel (pH 6.9) D. Acrylamide Bisacrylamide

5.og 1.25 g H20 to loo& E. HJP03 (1M) 12.8 ml Tris 2.85 g HzO to 100ml F. hboflavin 0.002 g H*O to 100ml Spacer gel composition (volume parts): D 2, E 1, F 1

304

PREPARATIVE ELECTROPHORESIS IN GEL MEDIA

Upper electrode buffer (-) (pH 8.3) Glycine 2.88g Tris 0.6g H20 to I 1 Lower electrode buffer (+) (pH 8.1) 3 elution buffer HCl(1 M ) 60.0 ml Tris to pH 8.1 H2O to 1 1 Electrode buffer (pH 8.3)according to Canalco Tris 3.0 g Glycine 14.4 g H20 t o 11 Elution buffer electrode buffer diluted with water 1 :8 Preparative gel electrophoresis with discontinuous elution The continuous elution of the fractions emerging from the gel column requires complex equipment and may lead to considerable dilution of the isolated components. To avoid this, several designs of electrophoretic apparatus employing intermittent elution have been described. In the simple arrangement of the discontinuous elution system [ 161, the gel is formed in a cylindrical glass column 3 cm in diameter, over the lower end of which is fitted a dialysis bag containing 12-15 ml of electrode buffer. Into the dialysis bag two capillaries are inserted, which are joined to the pump and W monitor to form a closed circuit. The electrophoretic run is performed in the usual way. When a component of the mixture being analysed migrates into the dialysis bag, the W absorbance increases until it reaches a limiting value; the electrophoresis is interruped and the dialysis bag is removed and replaced with a new one. Isolation of further components is performed in a similar manner. The method is simple but suitable only for well separated bands. The equipment is shown in Fig. 13.2. A simple and ingenious principle of an electrophoretic apparatus with intermittent elution was described by Marceau et al. [ 181, as shown in Fig. 13.3. A solution of acrylamide and cross-linhg agent is allowed to polymerize in a Perspex cylinder. A dialysis membrane is stretched over a flat, fritted disc and fixed at the bottom of the gel tube. The elution chamber is created by pumping a small volume of elution buffer through an inlet between the membrane and the lower surface of the gel with the outlet closed by a valve. The Perspex material allows displacement of the gel column up and down and ensures close contact along the wall. At the time of elution, the valve is opened and the buffer is expelled from the elution chamber by the hydrostatic pressure exerted on the upper surface of the gel column. Brownstone [ l l , 201 described an apparatus that permits the separation of gram amounts of protein compounds in a gel cylinder 5 cm high and 14 cm in diameter. The gel block is cooled at the upper and lower gel surfaces and the cylindrical face is insulated

305

REFERENCES

outled closed

inlet closed

buffer in

I

3 gel clown

I

outlet opened

Fig. 13.2. Preparative gel electrophoresis apparatus with stepwise elution. 1 : Gel column; 2: dialysis bag; 3 . 4 : capillary tubes; 5, 6 : platinum electrodes: 7 : pump; 8: UV monitor.

Fig. 13.3. Principle of elution of electrophoretic apparatus with intermittent elution according t o Marceau et al. [ 181.

with help of an insulating jacket to prevent heat losses. The intermittent elution system and automatic recycling of the eluted material ensure high resolution and purity of the isolated fractions. It appears that, in spite of the efforts of many workers, simple and versatile electrophoretic apparatus with a continuous or intermittent elution system, yielding high resolution and recovery, still remains to be constructed.

REFERENCES 1 U. J. Lewis and M. 0. Clark, Anal. Biochcm., 6 (1963) 303. 2 D. Racusen and N. Calvanico,Anal. Biochem., 7 (1964) 62.

306 3 4

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

PREPARATIVE ELECTROPHORESIS IN GEL MEDIA T. Jovon, A. Chrambach and M. A. Naughton, Anal. Biochem, 9 (1964) 351. S. Raymond, Science, 146 (1964) 406. J. V. Mabel, Jr., Ann. N. Y. Acad Sci., 121 (1964) 382. S. Hjerth, S. Jerstedt and A. Tiselius, Anal. Biochem, 11 (1965) 211. P. H. Duesberg and R. R. Rueckert, Anal. Biochem , 11 (1965) 342. T. HollmBn and E. Kulonen, Anal. Biochem., 14 (1966) 455. A. H. Gordon and L. N. Louis, Anal. Biochem., 21 (1967) 190. I. Schenkein, M. Levy and P. Weis, Anal. Biochem., 25 (1968) 387. A. D. Brownstone, Anal. Biochem., 27 (1969) 25. W. S. Bont, J. Geels and G. Rezelman, Anal. Biochem, 27 (1969) 99. S. Hjertbn, S. Jerstedt and A. Tiselius, Anal. Biochem., 27 (1969) 108. N. P. B. Dudman and B. Zerner, Anal. Biochem, 57 (1973) 14. T. R. C. Boyde and M. A. Remtulla, Anal. Biochem, 55 (1973) 492. A. Z. Reznick, H. W. Allen and R. J. Winzler, Anal. Biochem., 52 (1973) 395. J . Pristach and M. Bach, Chem. Lisfy, 68 (1974) 733. N. Marceau, R. Blais and N. Balaux, Anal. Biochem, 68 (1975) 17. P. P. Jaarsveld, B. J. van der Walt and C. H. Le Roux, Anal. Biochem., 75 (1976) 363. A. D. Brownstone, Anal. Biochem., 70 (1976) 572. D. Rodbard, A. Chrambach and C. H. Weiss, in R. C. Allen and H. R. Maurer (Editors), Electrophoresis and Isoelectric Focusing in Polyacrylamide Gel, De Gruyter, Berlin, 1974, p, 63. L. Strauch, Protides Biol. Fluids, 15 (1967) 535. L. B. Ellwein, R. W. Huff and A. Chrambach, Anal. Biochem, 82 (1977) 46. T. Suzuki, R. E. Benesch, S. Yung and R. Benesch, Anal. Biochem., 55 (1973) 249. S. Nees, in R. C. Allen and H. R. Maurer (Editors), Electrophoresis and Isoelectric Focusing in Polyacrylamide Gel, De Gruyter, Berlin, 1974, p. 189.

Chapter 14

Preparative electrophoresis in columns P. JUST SVENDSEN

CONTENTS Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical application and procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conventional zone electrophoresis in porous media . . . . . . . . . . . . . . . . . . . . . . . . Disc electrophoresis in polyacrylarnide gel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Isoelectric focusing in a sucrose gradient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

307 308 31 5 315 3 19 3 20 324

INTRODUCTION Several electrophoretic methods can be applied to the separation of proteins on a preparative scale. The basic principles are zone electrophoresis, isoelectric focusing and isotachophoresis. Separations as obtained in analytical experiments are the goal of preparative runs, and this is achieved by increasing the sample load and the cross-sectional area and length of the separation path. In column electrophoresis, the separation path is the space inside a vertical tube connected t o an electrode at each end and filled with the electrolyte and an appropriate stabilizing medium t o counteract convection arising from differences in density due t o variation in protein concentration and heat generated by the current flow. The Joule heat is removed by surrounding the separation tube with a cooling jacket. If the temperature increase becomes critical at high voltage and current, cooling in the centre of the separation tube [ I ] is also necessary. Numerous stabilizing media are available, divided into three types: porous media, gel media and liquid media (e.g., density gradients). The design of a column, or more strictly the design of an elution system for each type of anticonvection medium, is quite different. A buffer flow may be established through a porous medium and zones may thus be washed out of the matrix when separation is completed. The migration of a specific protein zone may be controlled by establishing a continuous flow of buffer in a direction opposite t o that of the electrophoretic migration. Gel media d o not allow a buffer flow through the matrix, and electrophoresis must be continued until the protein zones migrate off the gel bed into a continuous buffer flow sweeping the lower end of the matrix. When the separation has been performed in liquid media, the entire contents of the column are eluted, i.e., the protein zones are collected together with the anticonvection medium. The sample load varies with the electrophoretic principle together with the type of anti-convection medium employed: 1-2 nig/cni2 per band can be applied in isoelectric

308

PREPARATIVE ELECTROPHORESIS IN COLUMNS

focusing in sucrose gradients and 50-75 mg/cm2 per band using isotachophoresis in polyacrylamide gel. Pre-fractionation by other means should be performed in order to remove the main bulk of impurities before applying the sample on the column. This increases the possible load for a specific protein as many impurities are often present in much larger amounts than the protein itself, and thus exceed the upper load limit of the system, Protein samples should always be dialysed againt the electrophoresis buffer. No combination of electrophoretic principle and anticonvection medium provides ideal conditions for the optimal separation of any protein sample. It is therefore evident that an apparatus with easily interchangeable elution systems, each designed for a specific type of supporting medium, is the most practical solution. In this chapter apparatus and elution systems for column electrophoresis are described and examples are given for each type of supporting medium.

APPARATUS Fig. 14.1 shows a general view of a laboratory-scale universal column electrophoresis apparatus. The design [see cross-sections in Fig. 14.2 (lower end) and Fig. 14.3 (upper end)] allows separations in all three types of supporting media and is an improvement on the LKB Uniphor column. Elution can be performed from all media with a maintained electric field, and for porous media a regulating buffer flow can be introduced at any time during the experiment. The numbering in Fig. 14.1 corresponds to that in Fig. 14.2 and 14.3. The apparatus is an all-plastic construction, except for a few parts to be mentioned 4 0 m m I.D.),carrying below, and consists of a cooling jacket (1) (Perspex, 50mm O.D., two threaded PVC rings (3) for assembling the apparatus. The separation tube (2) (Perspex, 30 mm O.D., 26 mm I.D., or glass, 30 mm O.D., 27 mm I.D.) can be varied in its length together with the cooling jacket in order to obtain any desired length of separation path. A porous membrane (Millipore polypropylene, 50pm) can be stretched and fixed over the end of the separation tube for supporting the matrix. The end-pieces (4) (Perspex) are identical, and that used as the upper end-piece is upside down in relation to the lower one. The end-pieces are fully cooled (5). Electrode houses ( 6 ) (Perspex) are attached t o the end-pieces by means of threaded rings (14) (PVC). Semipermeable membranes (M) are inserted between the end-pieces and the electrode houses [ 2 , 3 ] to separate the electrode buffers from the column but leaving a conducting connection between the parts. The electrode holders (7), carrying the platinum wire electrodes (10) and pins (9) for connection to the power supply, are held in place by threaded PVC rings (8). The electrode houses have inlets (1 la) and outlets (1 lb) for electrode buffers, which are circulated through the electrode houses by means of nonconducting pumps (Perspex). PVC couplings (12a) with safety locks (12b) facilitate connection and disconnection of the tubings to the electrode houses. By this arrangement a high buffer stability around the electrodes is achieved as any size of buffer reservoir can be used and the electrode buffer can be renewed or changed at any time during the run. Owing t o the membrane (M), no disturbances, convection or increased pressure are transferred into the column.

APPARATUS

309

Fig. 14.1. Universal column electrophoresis apparatus. 1: Cooling jacket. 2: Separation tube. 3: Threaded ring. 4: End-pieces. 6: Electrode house. 10: Electrode. 1 l a: Inlet for circulating electrode buffer. 12a: Coupling for 10 mm O.D., 7 mm I.D. silicone-rubber tubing. 12b: Safety lock. 14: Threaded ring. 16: Buffer inlet. 18: Buffer outlet/air escape. 19: Elution stopper. 20a: Threaded ring. 21a: Inlet for cooling water. 21b: Outlet for cooling water. 22a: Inlet for elution buffer, air escapc. 22b: Buffer inlet. 23: End stopper. 24: Threaded ring. 28: Capillary, 5 mm O.D., 1 mm I.D. 29: Elution outlet, polyethylene tubing. E: Elution chamber. (Numbering as in Figs. 14.2 and 14.3.) AU tubings and external connections have been omitted for clarity.

310

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.2. Cross-section of lower section of universal column electrophoresis apparatus. 1: Cooling jacket. 2: Separation tube. 3: Threaded ring. 4 : End-piece. 5: Cooling channel. 6: Electrode house. 7: Electrode holder. 8: Threaded ring. 9: Pin for connection to power supply. 10: Electrode. 1 la: Inlet for circulating electrode buffer. 1 lb: Outlet for circulating electrode buffer. 12a: Coupling for 10 mm O.D., 7 mm I.D. silicone-rubber tubing. 12b: Safety lock. 13: Coupling for 6 mm O.D., 3 mm 1.1). silicone-rubber tubing. 14: Threaded ring. 15: Conical clamp. 16: Buffer inlet. 17: Hollow screw. 18: Buffer outletlair escape. 19: Elution stopper. 20a: Threaded ring. 21a: Inlet for cooling water. 22a: Inlet for elution buffer, air escape. E: Elution chamber. M: Dialysis membrane. (Numbering as in Fig. 14.1.)

The elution stopper (19) has an inlet (1 6) and an outlet (18) for buffer. The outlet originates just below the elution chamber, to be described below, and serves as an air escape when the elution stopper is being filled with buffer. A conical clamp (1 5) and a hollow screw (17) are used to stretch and hold the polyethylene elution outlet tubing (29) in Fig. 14.1, as described below, When the threaded PVC ring (20a) is tightened, the elution stopper is forced on to the lower end of the separation tube (2) and the elution chamber (E) is formed. A cooling water inlet (21a) and outlet (2 lb) are provided. When the apparatus has been assembled, the cooling water runs directly from the lower end-piece

APPARATUS

31 1

Fig. 14.3. Cross-section of upper section of universal column electrophoresis apparatus. 1 : Cooling jacket. 2: Separation tube. 3: Threaded ring. 4: End-piece. 5: Cooling channel. 9: Pin for connection to power supply. 10: Electrode, 21b: Outlet for cooling water. 22b: Buffer inlet. 23: End stopper. 24: Threaded ring. 25: Rubber rings. 26: Insert. 27: Threaded insert for attaching accessories. M: Dialysis membrane. (Numbering as in Fig. 14. 1 .)

into the cooling jacket and vice versa at the upper end-piece. The elution buffer inlet (22a) is located on exactly the same level as the lower end of the separation tube and is used as an air escape when the lower end-piece is filled with buffer. The upper end of the column is closed with a stopper (23) containing two rubber rings (25) and inserts (26 and 27) and held in place by a threaded PVC ring (20b). A pressure can be exerted on the rubber rings by turning the threaded ring (24) (PVC), thereby obtaining a leak-proof sealing of the capillary tube (5 mm O.D., 1 mrn I.D., glass) (28) in Fig. 14.1, which is placed in the centre of the stopper ( 2 3 ) . The glass capillary tube can be pushed downwards close to the surface of the column bed for sample application, and upwards to level with the lower end of the stopper (23), thus acting as an air escape when the upper end-piece is filled with buffer through the inlet (22b) from an outside buffer reservoir. The upper end of the capillary tube is closed with clamped siIicone-rubber tubing. The insert (27)has threads for attaching further accessories, e.g., a movable elution chamber (not shown).

312

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.4. Simplified column electrophoresis apparatus. 1: Cooling jacket. 2: Separation tube. 3: Threaded ring. 4: Lower end-piece. 5 : Air escape, connection to manometer tube. 6: Threaded ring. 7: Elution stopper. 11: Inlet for circulating electrode buffer. 12: Outlet for circulating electrode buffer. 13: Inlet for elution buffer. 14: Inlet for cooling water. 15: Upper end-piece. 16: Buffer reservoir. 17: Electrode. 18: End stopper holding electrode and with pin for connection to power supply. E: Elution chamber. (Numbering as in Fig. 14.5.) All tubings and external connections have been omitted for clarity.

By abandoning some of the demands for general universal application, the column design can be simplified [4], as shown in Fig. 14.4 (general view) and Fig. 14.5 (crosssection of the lower end). The number of parts has been greatly reduced, but the geometry of the separation path, the free choice of either a Perspex or a glass separation tube

313

APPARATUS

A-A

Fig. 14.5. Cross-section of simplified column electrophoresis apparatus. 1 : Cooling jacket. 2: Separation tube. 3: Threaded ring. 4 : Lower end-piece. 5 : Air escape, connection t o manometer tube. 6: Threaded ring. 7: Elution stopper. 8: Electrode. 9: Pin for connection to power supply. 10: Provision for conical clamp and hollow screw (15 and I 7 in Fig. 14.2). 11: Inlet for circulating electrode buffer. 12: Outlet for circulating electrode buffer. 13: Inlet for elution buffer. 14: Inlet for cooling water. IS: Elution chamber. A-A: View of elution stopper turned through 90". (Numbering as in Fig. 14.4.)

and the possibility of adding central cooling are maintained (see Fig. 3 in ref. 5). The number of possible elution techniques is reduced as only elution chambers with a dialysis membrane can be used. The upper electrode (17 in Fig. 14.4) is placed directly in a limited volume of buffer contained in the built-in upper buffers reservoir (16 in Fig. 14.4). For isoelectric focusing and electrophoresis in discontinuous buffer systems (disc electrophoresis and isotachophoresis), these limitations can be neglected for separations of macromolecules. The electrode (8 in Fig. 14.5) is placed in the elution stopper (7) and, when the electrode buffer is circulated continuously through the end-piece via the inlet (1 1) and the outlet (1 2) from a large outside buffer reservoir, the buffer stability is maintained even during long runs. The buffer reservoir should be cooled. The end-pieces in this construction can be coupled to the cooling jacket and separation tube in the same way as shown in Fig. 14.1, and the parts of the two designs are thus fully interchangeable and can be combined. Further, when a type of end-piece as 4 in Fig. 14.5 is attached to both ends of the cooling jacket and separation tube and combined with stoppers of the type as 23 in Fig. 14.3, a cooled column for chromatography, e.g., gel filtration and ionexchange chromatography, is obtained. The cross-sectional area of the above instruments is 5.3 cm2 (4.2 cm2 with central cooling). If the cross-sectional area is increased, central cooling must be incorporated. In Fig. 14.6 a design for large-scale electrophoresis in discontinuous buffer systenis is

314

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.6. Large-scale column electrophoresis apparatus. 1: Cooling jacket. 2: Separation tube (Pyrex glass). 3: Central cooling (Pyrex glass). 4: Air escape, connection to manometer tube. 5: Threaded ring. 6: Elution stopper with built-in electrode. 7: Inlet for circulating electrode buffer. 8: Elution outlet with coupling to polyethylene tubing. 9: Outlet for circulating electrode buffer. 10: Inlet for elution buffer. 11: Inlet for cooling water. 12: Outlet for cooling water. 13: Buffer reservoir. 14: Electrode. 15: Threaded ring. 16: Stopper for holding and centering cold finger 17, and for holding electrode with pin for connection to power supply. 18: Outlet for cooling water. 19: Inlet for cooling water. 20: Threaded ring. E: Elution chamber. A11 tubings and external connections have been omitted for clarity.

PRACTICAL APPLICATION AND PROCEDURES

315

shown. The cross-sectional area of the separation path is 20.6 cmz with central cooling. The separation tube and cold finger are made of glass. Employing the isotachophoretic principle [S] in polyacrylamide a load of 1-1 .5 g per band can be applied. The design is of the same type as shown in Figs. 14.4 and 14.5 and is therefore very compact.

PRACTICAL APPLICATION AND PROCEDURES Conventional zone electrophoresis in porous media The column shown in Fig. 14.1 is assembled using a Perspex separation tube 60 cm long and fitted with a porous polypropylene membrane stretched over the lower end of the separation tube by means of a ring (8 in Fig. 14.7). The electrode houses are attached to the end-pieces and dialysis membranes are inserted between the end-pieces and electrode houses. The electrodes are brought into place and connected t o the power supply (LKB 2103). Cooling is started, the temperature of the cooling water being 3°C. To obtain an elution system as shown in Fig. 14.7, the elution stopper (19 in Fig. 14.2) is prepared in the following way. A dialysis membrane is stretched over the upper end of the elution stopper by means of a Perspex ring (R in Fig. 14.7). Polyethylene tubing, moulded by means of a special tool at one end as shown (4 in Fig. 14.7), is pushed through a hole cut in the centre of the dialysis membrane and all the way through the elution stopper. When the hollow screw (17 in Fig. 14.2) is tightened, the conical clamp (1 5 in Fig. 14.2) grips firmly around the tubing. The tubing is then pulled downwards and, as it stretches, it slides through the conical clamp and a perfect seal is obtained between the dialysis membrane and the silicone-rubber insert (3 in Fig. 14.7) and between the flanged end of the polyethylene tubing and the dialysis membrane. The elution stopper is pushed into the lower end-piece but is not tightened by means of the threaded ring (20 in Fig. 14.2) until after the end-piece and elution stopper have been filled with buffer. The elution outlet tubing (29 in Fig. 14.1) is connected to an LKB Perspex pump. An extension tube of the same diameter as that of the separation tube is attached to the upper end-piece. A 300-ml volume of Ultrogel AcA 44, suspended and decanted twice in 0.2 M Tris-0.05 M HCl (pH 8.6), is diluted to a total volume of 450 ml with water and poured into the column. During the packing, a flow-rate of 25 ml/h is maintained. After packing, the stopper (23 in Fig. 14.1) is placed in the upper end-piece. A capillary tube fitted with a piece of silicone-rubber tubing with a clamp is inserted into the stopper. The upper end-piece is filled with buffer through the inlet (22b in Fig. 14.3) from a Mariotte flask. When all air has escaped through the capillary, it is closed by means of the clamp on the silicone-rubber tubing. Buffer in a volume twice that of the column bed is then passed through the column at 25.4 ml/h for 24 h. The capillary is pushed downwards to a position 1 mm above the surface of the column bed. An LKB Perspex pump is connected to the capillary and 3 ml of a dialysed mixture of human serum orosomucoid (ca. 40 mg) and human serum transferrin (ca. 60 mg) are layered on top of the column bed. The displaced buffer volume escapes through 22b (Fig. 14.3) into the Mariotte flask. After application of the sample, the capillary is returned to the upper position and closed.

316

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.7. Cross-section of the elution chamber. 1 : Inlet for elution buffer, air escape. 2: Buffer outlet,

air escape. 3: Silicone support for 4 . 4 : Elution outlet (flanged polyethylene tubing). 5:Anticonvection medium. 6: Separation tube. 7 : Cooling channel. 8: Ring for stretching the supporting membrane for the column bed. 9: Elution stopper. 10: End-piece. 11: O-ring. R: Perspex ring for stretching the dialysis membrane, M . The buffer flow in the elution chamber is indicated by arrows (porous media).

The flow characteristics of a packed porous column bed should always be tested by permitting a sample to pass through the matrix. This is demonstrated by the elution profiles shown in Fig. 14.8A (W elution profile) and Fig. 14.8B [immunochemical elution profile (fused rocket immunoelectrophoresis, see ref. 6)]. Fractions of 5.5 ml were collected in an LKB Ultrorac at 4°C. It can be seen that smooth symmetrical bands are obtained. An identical sample was then applied on the column and buffer circulation through the electrode houses was started. A buffer flow of 5.3 ml/h in the direction of the elution chamber was established. After 1 h, 400 V/62 mA were applied over the column with the anode at the upper end-piece. The power supply (LKB 2103) was set to constant voltage at 400 V. The current was set to be maximum 65 mA and the power to a maximum of 25 W.This polarity will cause the sample proteins to migrate electrophoretically towards the upper end of the column, but the constant flow of buffer in the opposite direction will counterbalance this [3,7], i.e., the movement of the protein zones in relation to the matrix can be regulated b y varying the voltage or the buffer flow. The voltage and flow should be set so that the fastest migrating protein, in this instance the orosomucoid, moves downwards as slowly as possible. Owing to the design of the elution chamber, fractions are collected continuously during electrophoresis. The current was discontinued after 46 h and the buffer flow increased to 12.7 ml/h. Fractions of 5.5 ml were collected during the entire run.

PRACTICAL APPLICATION AND PROCEDURES

317

Fig. 14.8. (A) UV elution profile from experiment testing the flow characteristics of Ultrogel AcA 44. (B) Immunochemical elution profile of the same experiment. 0: Orosomucoid. T: Transferrin. The fraction numbers are indicated. Sample volume in the application wells, 3 bl; 2 nil of anti-total serum were used (Dakopatts, Copenhagen, code 100SF).

The U V elution profile is shown in Fig. 14.9A and the imniunochemical elution profile in Fig. 14.9B. The vertical dotted line indicates the point at which the current was broken (46 h). The proteins have been completely separated, but as the experiment could have been continued for 8 4 h more for the top fraction of orosomucoid to emerge, increased separation can be achieved [ 3 ] .In fact, the technique corresponds to running in a much

318

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.9. (A) UV elution profile from electrophoresis experiment with a regulating buffer counter flow. (B) Immunochemical elution profde of the same experiment. The dotted line indicates the time at which the current was disconnected. 0: Orosomucoid. T: Transferrin. The fraction numbers are indicated. Sample volumes in the sample wells, 4 pl; 2 ml of anti-total serum were used (Dakopatts, Copenhagen, code 100SF).

longer column: in this instance, running for 130 h corresponds to a column length of 150 cm. If, in a porous medium, electrophoresis is performed with migration towards the lower end of the column, continuous elution is also possible. The solid Perspex ring (R in Fig. 14.7) is replaced with a porous polyethylene ring [8] of the same design. The inlet (1 in Fig. 14.7) is connected t o a pump delivering the same volume per unit time as the pump connected to the outlet (4 in Fig. 14.7). A continuous flow of buffer will then sweep the elution chamber, and protein zones are collected during electrophoresis as they migrate off the column bed and enter the elution chamber.

PRACTICAL APPLICATION AND PROCEDURES

319

Disc electrophoresisin polyacrylamide gel [ 9- I 1 1 A Perspex separation tube 20 cm long is fitted with a porous polypropylene membrane at the lower end and placed in a cooling jacket for polymerization (not shown). The temperature of the cooling water is 3°C. A 50-ml volume of a gel mixture is poured into the separation tube, overlayered with distilled water and polymerized for 60 min under fluorescent light. The gel concentration is 8.5%, with 3% cross-linking wit% Bis (N,N’methylenebisacrylamide). The final buffer concentration is 0.03 M H$04-0. 18 M Tris [2-amino-2-(hydroxymethyl)propane-l,3-diol], pH 8 $5.A catalyst as described elsewhere in this book [ S ] is used. A stacking gel is then polymerized on top of the first one with a gel concentration of S%, with 15% cross-linking with DATD (N,N’-diallyltartardiamide). The final buffer concentration is 0.03M H3P04-0.033MTris, pH 6.25. The apparatus shown in Fig. 14.4 is assembled with the separation tube containing the gels and cooling is started (temperature 3 O C ) . The upper part is filled with 0.2 M glycine0.02SM Tris, pH 8.4. The stopper with the electrode (18 in Fig. 14.4) is attached t o the top of the reservoir (16 in Fig. 14.4) and 3 nil of a saniple (dialysed against the Trisglycine buffer) containing ca. 60 mg of human serum transferrin and 40 nig of human serum orosomucoid are applied on top of the stacking gel. The elution stopper (7 in Fig. 14.4) is prepared as described above, but a porous polyethylene ring (R in Fig. 14.10) is used instead of a solid Perspex ring. The elution stopper is pushed into the lower endpiece (4 in Fig. 14.4), which is filled with elution buffer (0.03M H2S04-0.066MTris, p1-I 7.1) from a Mariotte flask connected to the inlet (13 in Fig. 14.4). The buffer surface must be level with the surface of the buffer in the reservoir (16 in Fig. 14.4). The tubing connected to the outlet ( 5 in Fig. 14.4) can be used as a manometer to facilitate the adjustment. The air escape (5 in Fig. 14.4) is closed when all air has left the end-piece. The threaded ring (6 in Fig. 14.4) is tightened and the elution chamber is formed as shown in Fig. 14.10. Circulation of buffer through the elution stopper is started through the inlet ( I I in Fig. 14.4) and the outlet (12 in Fig. 14.4). The anode buffer is the same as the elution buffer. The elution outlet tubing is connected to an LKB Perspex pump, flow-rate 25.4 ml/h. The apparatus is connected to the power supply and electrophoresis is performed at a constant current of 10 m A . Fractions are taken every 20min and the experiment is run for 30 h. The elution profiles are shown in Fig. 14.1 1 A and B. The discontinuity of the orosomucoid peak is due to the gel concentration, which corresponds t o the “stacking limit” for orosomucoid found in a Ferguson plot with the same buffer system and crosslinking [5]. The front of the peak still remains “stacked” and the rear of the peak is “unstacked”, i.e., the orosomucoid has not been completely overtaken by the glycine. At a gel concentration of 10% the orosomucoid would have migrated behind the phosphate-glycine boundary, and at a gel concetration of 6% the orosomucoid would have remained fully “stacked” between the phosphate and the glycine, but leaving the transferrin behind in the glycine zone. By adjusting the gel concentration very precisely, such “selective stacking” [I 21 can be utilized for protein separation. When €-aminocaproic acid is used instead of glycine in an experiment with the same gel buffer as used here and with a gel concentration of 6%, both proteins will remain stacked as the net mobility of e-aminocaproic acid is much lower than that of glycine and transferrin.

320

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.10. Cross-section of the elution chamber. R: Porous ring for stretching the dialysis membrane, M. The buffer flow in the elution chamber is indicated by arrows (gel media). For details, see text.

By adding a limited amount of glycine t o the sample in such an experiment, orosomucoid and transferrin will be “spaced” by the glycine, which then migrates between the two proteins, i.e., orosomucoid is stacked between phosphate and glycine, and transferrin is stacked between glycine and eaminocaproic acid. By adding a number of “spacer” ions with different net mobilities intermediate to those of the leading and terminating ions, the possibility of stacking different sample molecules between different boundaries arises (“cascade stacking” [13]). Ampholine displays a wide range of such intermediate net mobilities and, when added to the sample, separation of proteins is obtained (“isotachophoresis” [5]) in a migrating pH gradient. Further elution techniques for use with gel supporting media are described elsewhere in this book [5]. Isoelectric focusing in a sucrose gradient [ 14,151 The apparatus shown in Fig. 14.1 is assembled using a glass separation tube 20 cm long and fitted with a cold finger (see Fig. 16.1). The elution stopper (19 in Fig. 14.1) is prepared as described above, but a ring (R in Fig. 14.12) is used. Note that this ring does not have a flange and the elution outlet tubing extends into the separation tube. The elution stopper is pushed into the lower end-piece and the threaded ring (20a in Fig. 14.1) is tightened. An elution system as shown in Fig. 14.12 is formed. The cold finger is adjusted so that a space of 0.5 mm is left between the lower end of the cold finger and the upper edge of the extended elution outlet tubing. Cooling is started, the temperature of the cooling water being 3°C.The elution stopper is filled with 0.01 M HQ04 in 60%(w/v)

PRACTICAL APPLICATION AND PROCEDURES

32 1

Fig. 14.1 1. (A) UV elution profile from disc electrophoresis experiment in polyacrylamide. (B) lmmunochemical elution profile from the same experiment. 0: Orosomucoid. T: Transferrin. The fraction numbers are indicated. Sample volume in the sample wells, 5 pl; 2 ml of anti-total serum were used (Dakopatts, Copenhagen, code 1OOSI’).

sucrose. The lower end of the separation tube is filled through the inlet (1 in Fig. 14.12), also with 0.01 M H3P04 in 60% (w/v) sucrose, t o approximately 1 cm above the lower end of the cold finger. An LKB Perspex pump connected t o the elution outlet is run briefly in order t o remove the air from inside the tubing. A second pump is connected t o the inlet (1 in Fig. 14.12), care being taken that n o air is trapped inside the tubing. From the top

322

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.12. Cross-section of elution chamber. 1 : Buffer inlet. 4: Elution outlet tubing (extended). R : Perspex ring for stretching the dialysis membrane, M. The buffer flow during elution is indicated by arrows (liquid media). The contour of the cold finger is indicated by a dotted line. For details, see text.

along the wall of the separation tube a linear sucrose gradient is introduced, from 55% (w/v) sucrose, 2% Ampholine 5-7 to 0% sucrose, 2% Ampholine 5-7. The flow-rate is 50 ml/h and the total volume 70ml. When 10 ml have entered, 10 mg of human serum transferrin in 700 pl of 0.1 M KCl are added to the mixing chamber, then 5 ml of 2% Ampholine are added on top of the gradient and finally 0.01 M NaOH solution is pumped into the upper end-piece. The inlets (1 l a in Fig. 14.1) in the electrode houses (6 in Fig. 14.1) are closed with clamped silicone-rubber tubing, The lower electrode house is filled with 0.01 M H3P04 in 60%(w/v) sucrose and the upper electrode house with 0.01 M NaOH solution. The apparatus is connected to the power supply (anode at the lower endpiece). The starting voltage is 18OV at 2 mA, increasing and stabilizing during the experiment to 500 V at 1.5 mA. After 48 h, elution is performed in the following way to maintain the electrical field during elution [4]. The pump connected to the outlet tubing (4 in Fig. 14.12) is set to a flow-rate of 25.4ml/h. The pump connected to the inlet (1 in Fig. 14.12) is set to a flowrate of 5.3 m l b , conveying 0.01 M H$04 in 60%(w/v) sucrose into the column. In this way, a very stable boundary is formed at the level of the upper edge of the outlet tubing. The sample zones in the density gradient will in this boundary be directed to the outlet and thus be eluted smoothly, without disturbing the above gradient. As the current passes freely through the column, equilibrium is maintained during the elution. Contact with the upper electrode is maintained by connecting the inlet (22b in Fig. 14.1) to a Mariotte flask containing 0.01 M NaOH solution. The elution profiles are shown in Fig. 14.13A and B. As the density of the eluate

PRACTICAL APPLICATION AND PROCEDURES

323

Fig. 14.13. (A) UVelution profileofelectrofocusingexperiment ina sucrosegradient. The pH measured in the fractions is superimposed on the diagram. (B) lmmunochemical elution profile of the same experiment. The fraction numbers are indicated. Sample volume in the sample wells, 2 ~ 1 2; rnl of antitotal serum were used (Dakopatts, Copenhagen, code 1OOSF).

decreases during the elution, the flow in the quartz cell in the Uvicord is downwards for maximum stability. Better resolution in the immunochemical elution profile could have been obtained by taking fractions at shorter intervals. The pHs measured d o not correspond to the p l values owing to the admixing of H3P04,but for preparative purposes it is good resolution that is desired, rather than p/ values. The current has to be discontinued

324

PREPARATIVE ELECTROPHORESIS IN COLUMNS

Fig. 14.14. Cross-section of the elution chamber. 4: Elution outlet tubing; R : Perspex ring for stretching the dialysis membrane, M. The buffer flow during elution is indicated by arrows (liquid media). The contour of the cold finger is indicated by a dotted line. For details see text.

during elution in runs for measuring the p l values, but diffusion will then decrease the resolution actually obtained as the time for smooth elution is long enough to spread the zones by diffusion. If runs for measuring p l values are performed, another arrangement of the elution system is employed (see Fig. 14.14). The ring (R in Fig. 14.14) is a solid Perspex ring. Before elution is started, the elution stopper is emptied and the current broken. The cold finger is placed 1 mm above the elution outlet and only one pump is used and connected to the outlet tubing (4). Otherwise, the procedure is the same as described above. Preparative column electrophoresis as described here is easy to perform as the same instrument (Fig. 14.1) can be modified in a simple way for separations in all anticonvection media without altering the general operation of the column. The possibilities seem very wide and it should be possible to scale up almost any analytical separation t o a preparative scale.

REFERENCES 1 J. Porath,Ark. Kerni, 11, No. 18 (1957) 161. 2 B. Bergrahm, Sci. Tools,14 (1967) 34. 3 P. J. Svendsen, Anal. Biochern., 25 (1968) 236. 4 P. J. Svendsen,Protides Biol. Fluids,17 (1970) 413. 5 P. J. Svendsen, in Z. Deyl (Editor), Electrophoresis. A Survey of Techniques and Applications. Part A: Techniques, Elsevier, Amsterdam, Oxford, New York, 1979, Chapter 16.

REFERENCES 6

I 8 9 10 11 12 13 14 15

325

P. 1. Svendsen, in Z.Deyl (Editor), Electrophoresis. A Survey of Techniques and Applications. Part A : Techniques, Elsevier, Amsterdam, Oxford, New York, 1979, Chapter 7. H. Hochstrasser, L. T. Skeggs, Jr., K . E. Lentz and J. R. Kahn, Anal. Biochem., 6 (1963) 13. B. Bergrahm and R. Harlestam,Sci. Tools, 15 (1968) 26. L. Ornstein, Ann. N. Y. Acad. Sci., 121 (1964) 321. B. J. Davies, Ann. N . Y. Acad. Sci., 121 (1964) 404. ‘T.Jovin, A. Chrambach and M. A. Naughton, A n d . Biochern., 9 (1964) 351. A . Chrambach, T. M. Jovin, 1’. J . Svendsen and D. Rodbard, in N. Catsirnpoolas(Editor),Methods ofprotein Separation. Vol. 2, Plenum, New York, 1976, p. 27. N. Y. Nguyen, D. Rodbard, P. J . Svendsen and A. Chrambach, Anal. Biochem., 77 (1977) 39. 1-1. Svensson, Arch. Riochern. Biophjx, Suppl. 1 (1962) 132. 11. Haglund,Methods Biochern. Anal., 19 (1971) 1 .

This Page Intentionally Left Blank

Chapter 15

Preparative isoelectric focusing P . BLANICKP

CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carrier ampholytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution and loading capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Separation of proteins from carrier ampholytes . . . . . . . . . . . . . . . . . . . . . . . . . Stabilizing medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pre-treatment of sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrical conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Detection methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pre-experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Difficulties i n preparative isoelectric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carrier ampholyte-protein interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . pHdrift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protein precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preparative isoelectric focusing technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Isoelectric focusing in a density gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Technical equipmcnt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Density gradient material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution and loading capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Difficulties due to protein precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Polarity of the isoelectric focusing system . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emptying of isoelectric focusing columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flat-bed isoelectric focusing in granulated gel . . . . . . . . . . . . . . . . . . . . . . . . . . . . Technical equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Granulated gel material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution and loading capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protein detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protein elution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zone convection isoelectric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Technical equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution and loading capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Continuous flow isoelectric focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Technical equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution and loading capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

328 328 328 328 328 329 329 330 330 331 331 331 331 331 332 332 332 332 333 334 334 334 335 335 336 336 336 337 337 338 338 338 338 338 339 339 339 339 340 340 341 34 1 342

328

PREPARATIVE ISOELECTRIC FOCUSING

INTRODUCTION Apart from its analytical applications, isoelectric focusing is being used with considerable success in preparative work with proteins on account of its high concentrating effect, resolving power and loading capacity. As theoretical aspects of isoelectric focusing in general, as formulated by Svensson [ 1, 21, have been presented in Chapter 9 , only special comments important for preparative work with isoelectric focusing will be briefly mentioned here. This chapter deals with preparative isoelectric focusing techniques, e.g. those in which it is possible to use more than l00mg of the sample: - isoelectric focusing in a density gradient - flat-bed isoelectric focusing in a granulated gel - zone convection isoelectric focusing continuous-flow isoelectric focusing. ~

GENERAL ASPECTS Carrier ampholytes

Properties The most important properties of the carrier ampholytes [3] in preparative isoelectric focusing are as follows: - good buffering capacity at the p l - one of the factors limiting the load capacity - good conductivity at the p l - molecular weight lower than that of proteins to be separated - low light absorption at 280 nm. As the buffering capacity of the carrier ampholytes usually has to be stronger than that of the proteins being separated, a concentration of carrier ampholyte below 10% (wlv) is used, 1-2% (w/v) being most frequently used. Ampholine (LKB-Produkter, Bromma, Sweden) has so far been the most popular carrier ampholyte. In addition, Servalytes (Serva, Heidelberg, G.F.R.) and Biolytes (BioRad Labs., Richmond, Calif., USA.), for example, are also commercially available. Some modified protein hydrolysates can also be used as carrier ampholytes in preparative isoelectric focusing [4].

Resolution and loading capacity A good resolution [5,6] can be obtained with proteins with a low diffusion coefficient and a high mobility slope at the pl. A good resolution is further favoured by a high electric field strength and a pH gradient slope. In preparative isoelectric focusing, in contrast to the analytical procedure, proteins have to be separated so that they can be recovered in a pure state, even if they are applied in large amounts. In comparison with other electrophoretic methods, a good resolving

3 29

GENERAL ASPECTS

power and a high loading capacity are not contradictory in isoelectric focusing [7]. On the contrary, when a narrower pH range of the carrier ampholyte is used, the distances between the separated proteins are larger and, in addition, the protein concentrations are better distributed for focusing into zones (the focused zones are broader). Extremely sharp zones mostly do not represent a good resolution; in such a case, the pH gradient is steep and differences between zones with narrow isoelectric points are smqll. At present, in preparative isoelectric focusing one tends to use narrower pH ranges of the carrier ampholytes (1 -2 pH units), which are favourable in all of the techniques described below. Preparative isoelectric focusing with a narrow pH range can be used as the second step at a very early stage in protein purification, which follows the preliminary separation with a wide pH range (3.5-10). Carrier ampholytes with a pH range of less than 1.5 are not commercially available (Fig. 15.1). It is not difficult, however, to prepare very narrow pH ranges by isoelectric focusing of more concentrated carrier ampholytes from which the desirable pH range is selected and used thereafter. By this means the resolution and loading capacity can be increased [6].

2

4

6

8

10

pH

Fig. 15.1. Standard Ampholine pH ranges (LKR).

Separation of proteins from carrier ampholytes After preparative isoelectric focusing, eluted fractions contain a certain amount of carrier ampholytes (and/or density gradient material) in addition to the separated proteins. Carrier ampholytes can be separated from proteins by gel filtration on Sephadex G-50 [8], ion-exchange chromatography [9] and salting-out [ 101. Dialysis is a time-consuming procedure and is rarely used [ I 1 J. Stabilizing medium In contrast to most other electrophoretic methods, isoelectric focusing leads to an equilibrium in a pH gradient in which ampholytes (proteins) are concentrated at positions in the gradient where the pH is equal to the isoelectric point of the ampholyte (protein) [l]. In the steady state and with a large potential gradient applied, the conductivity of the system is low. Therefore, the bands of focused proteins are sharper and more concentrated in comparison with the protein bands in other electrophoretic methods. In the pH gradient range used, the stabilizing medium must be electrically neutral and chemically neutral and chemically stable. All kinds of convective flows must be excluded,

330

PREPARATIVE ISOELECTRIC FOCUSING

and an osmotic flow which spoils the protein zones and changes their location in the system is equally undesirable. The stabilizing medium must also be able to support zones of very high density. The medium for preparative isoelectric focusing can be stabilized by suitable density gradient material [ 121, Sephadex [ 13,141, polyacrylamide [ 151, a free-flow procedure [ 161 and positive vertical density gradient ampholytes [ 171. Polyacrylamide, however, which is a good medium for analytical assays, is rarely used in preparative isoelectric focusing owing to difficulties in separating the proteins from the gel. Selective elution of protein zones from the gel by using a suitable ampholyte or buffer, in which the protein to be isolated has a reasonable electrophoretic mobility, seems promising [ 181. Pre-treatment of sample

As the salt content of the sample must be kept at a minimum, desalting of the sample by molecular sieving, by dialysis or by ultracentrifugation is a standard prerequisite in preparative isoelectric focusing, Most proteins require a certain ionic strength or dipole moment of the solvent in order to form stable solutions. In such instances, 1% glycine is used as the dialysis medium, which, owing to its zwitterionic character, increases the dipole moment of the solvent without affecting the isoelectric focusing [19,20]. Desalting of the sample against diluted carrier ampholyte solution is also used. Electrical conditions As the technical equipment for preparative isoelectric focusing is larger than that used for analytical purposes, the experimental time is particularly important. As in any separation technique, isoelectric focusing exposes a protein to a more or less artificial environment which might cause deleterious structural changes [21]. Another factor specific to isoelectric focusing is the time course of isoelectric precipitation. It is therefore obvious that the experimental time should be kept as short as possible. A short experimental time (and also the actual resolution) depends on the electric field strength, which can be increased to a certain limiting value. This limit is set by the conductivity distribution along the pH gradient and the cooling capacity of the system [6]. One of the best means of reducing the experimental time is to maintain optimal electrical conditions during the whole experiment. A power supply with a constant power is suitable for this purpose. Before choosing the electrical conditions, it is necessary to remember the high conductance of the system at the beginning of the run (carrier ampholytes are in a mixed state); further, in the fully established pH gradient excessive heat production within the lowconductivity zones may occur [22]. The capacity of the cooling system also depends on the construction of the cooling system used, apart from the temperature of the cooling system (usually 2-10°C) (thin layers are more intensively cooled than columns). In addition, the experimental time depends on the pH range chosen, increasing with the use of narrower pH ranges [22]. If the separated protein is labile by nature, introduction of the sample after the pH gradient has been established helps to shorten the running time.

GENERAL ASPECTS

33 I

Detection methods In order to exploit the high resolving power of isoelectric focusing, there is a need for a sensitive and specific detection of focused proteins. The most usual means of obtaining a protein elution profile is monitoring by UV absorptiometry. This method however, is less specific when a mixture of proteins with narrow isoelectric points is analysed, and a far more detailed picture can be obtained by using immunoelectrophoretic techniques ~31. Measurements of biological activity should be carried out after adjusting the pH to a physiological value. In most instances the carrier ampholytes or the density gradient material, which interfere in such measurements, should be removed (see Separation of proteins from carrier ampholytes, p. 329). A decrease in biological activity due to oxidation of -SH groups may occur during isoelectric focusing. This can be prevented by the addition of a small amount of a reducing agent, e.g. a thiol compound [24]. A protective layer of ascorbic acid placed at the anodic side of the pH gradient has also been found to the effective [25]. As the pH depends on temperature, it is essential to measure the pH at the same temperature as is used in isoelectric focusing. Pre-experimental Before starting preparative work with proteins using isoelectric focusing, it is useful to obtain preliminary information about the sample to be separated as far as its protein isoelectric profile, p l solubility and pH stability of individual components and influence of the carrier ampholyte on the determination of biological activity are concerned. A considerable proportion of difficulties arising in the course of isoelectric focusing can thus be eliminated. Difficulties in preparative isoelectric focusing Carrier arnpholyte -pro tein interaction

A reversible weak carrier ampholyte-protein complex analogous to protein complexes with other buffering substances or those formed under conditions of increasing ionic strength is assumed to occur in isoelectric focusing. The existence of an irreversible strong carrier ampholyte-protein complex has not been proved so far [6,8]. On the contrary, some workers have reported a pronounced stabilizing effect of carrier ampholytes (Ampholine) on certain proteins [26,27]. pH drift

In contrast to the theoretically assumed stationary state of the established pH gradient, a slow, mainly cathodic, drift of the pH gradient occurs within the isoelectric focusing process. This causes a flattening in the middle of the pH gradient with a concomitant steep rise in the region around pH 8. Thus the resolving power is decreased at this pH.

332

PREPARATIVE ISOELECIRIC FOCUSING

The shift in the pH gradient is clearly visible and pronounced during isoelectric focusing in polyacrylamide gel, but it has also been reported to occur in other isoelectric focusing techniques [28,29]. The cause of this effect is not clear and diverse explanations have been put forward; the effect may be due to electroosmosis, ampholyte diffusion into electrode solutions, water acting as an ampholyte and its focusing to its isoelectric point, anodic migration of carbonate formed by adsorption of carbon dioxide in alkaline electrode solution, electrochemical modification of the carrier ampholyte (both its anodic oxidation and cathodic reduction) or nonequilibrium between electrophoretic migration and diffusion. The shift in pH hardly occurs in isoelectric focusing of proteins at acidic pH [29].

Protein precipitation Precipitation of proteins is another factor that may have an unfavourable influence on isoelectric focusing. It may occur for various reasons, such as low stability (solubility) at the isoelectric point, exposure to a denaturing pH within the pH gradient or in the electrode solutions, and loading more material than the isoelectric solubility permits. Protein precipitation during isoelectric focusing can be prevented by using non-ionized solubilizing agents. The most commonly used is urea, the solubilizing effect of which is thought to be due to the weakening of hydrophobic bonds with a certain degree of unfolding of the peptide chain. In addition, it prevents protein aggregation. Its disadvantage is that it has a slight influence on the isoelectric points of proteins [30,31]. Some uncharged and zwitterionic detergents (e.g., Triton X-100, Brij-35, Emulgophene, alkylbetaine, sulphobetaine) have been used successfully as solubilizing agents. Their concentration in the whole pH gradient varies between 0.1 and 1% [32]. A 1% glycine solution can be used to prevent precipitation of proteins in solutions desalted before isoelectric focusing (see Re-treatment of sample, p. 330). Another means of preventing protein precipitation during isoelectric focusing and simultaneously of limiting the influence of precipitated contaminants is to use narrower pH ranges. The loading capacity of the system is thereby increased and the precipitation of protein contaminants in crude samples containing only a minute amount of the component to be purified will occur outside the resolution range of the pH gradient.

PREPARATIVE ISOELECTRIC FOCUSING TECHNIQUE Isoelectric focusing in a density gradient

Technical equipment

A density gradient as a medium suitable for stabilization of the pH gradient was first used by Svensson [33]. Isoelectric focusing in a density gradient is the most frequently used preparative isoelectric focusing technique, mainly owing to the commercial availability of the equipment. Two types of column for preparative isoelectric focusing (1 10 and 440 ml) are

PREPARATIVE ISOELECTRIC FOCUSING TECHNIQUE

333

t

--a --b -c

I Fig. 15.2. Scheme of preparative column isoclectric focusing in a density gradient (LKB). a: Inner water jacket; b: outer water jacket; c: annulus between the water jackets.

manufactured by LKB, each consisting of a glass column with an inner and outer water jacket for efficient cooling. The pH gradient is established in the annulus between the water jackets. At each end of the gradient is an electrode solution, either acidic or basic. One electrode is a loop of a wire in the upper compartment of the column and the other is wrapped around a PTFE rod in the innermost tube of the column. The electrodes are placed so that no bubbles formed on them can interfere with the separation (Fig. 15.2).

Density gradient material The density gradient material must satisfy the conditions required for the carrier medium. Also, it must be readily soluble in water and its solution must not be lightabsorbing in the UV region of the spectrum or interfere with the determination of biological activity. An adequate purity is also important; eg., impure sucrose (see below) may contain protein contaminants that focus at characteristic bands [34]. Sucrose is the most fequently used material for preparing the density gradient. In addition, glycerol [ 3 5 ] , sorbitol [ 3 6 ] . ethylene glycol [37], ficoll [38] and dextran [39] have been used. The choice of the density gradient material is of some importance in isoelectric focusing. For instance, instead of sucrose, which tends to be dissociated at alkaline pH, it is necessary to use another material (glycerol) in the separation of proteins that are isoelectric in the pH range 8- 1 1 .

334

PREPARATIVE ISOELECTRIC FOCUSING

Resolution and loading capacity The carrying capacity of the density gradient, which is directly associated with the range of the pH gradient of choice, is very important in preparative isoelectric focusing, A broad pH’range (a steep pH gradient) results in very sharp and concentrated focused protein zones; the loading capacity of the system may be exceeded and droplet sedimentation of the zones may occur. The use of narrower pH ranges, which leads to broadening of the zones, and thus to their better resolution, is therefore fundamental. With such a higher resolution capacity, the carrying capacity of the density gradient increases proportionally with the square of the length of the focused zone [7].Thus, when the maximal theoretical amount of protein in a zone 0.1 cm long in a 110-ml column (LKB) is estimated to be 1 mg, a 5 c m zone would carry 2500 mg. Separation of whale-sperm myoglobin in a pH gradient of 7-9, containing 74% of the theoretical maximum (in this instance 1074mg), has been reported [7]. The loading capacity of the system makes it possible to use larger amounts of the sample to be separated (more than 100 mg) [32]; for instance, 7.3 g of the sample in a 440-ml column (LKB) was used for the purification of cuglucosidase from Qtophuga johnsonii in the pH range 6-8 [40].

Difficulties due to protein precipitation Protein precipitation in preparative isoelectric focusing in a density gradient is undesirable. The precipitate may slowly sediment, thereby contaminating other neighbouring zones or causing their coprecipitation. Very often precipitates have a tendency to stick to the column wall and cause “tailing” upon elution. In addition, heavy precipitates can block the outlet of the column. Causes of protein precipitation are mentioned above. As pointed out below, the undesirable effect of a protein precipitate can be limited by malung a suitable choice of the polarity of the isoelectric focusing system, by the method of sample introduction and by suitable elution. Protein precipitation can be prevented by adding of solubilizing agents or using a narrower pH range in which a larger amount of protein can be focused without exceeding the solubility limit.

Polarity o f the isoelectric focusing system When choosing the polarity of the isoelectric focusing system, the conductivity distribution in a fully established pH gradient (which exhibits a minimum between pH 6 and 7) and the conductivity decrease at the bottom of the column (as a result of increased viscosity) must be considered. In preparative isoelectric focusing, a very crude material with a low protein content to be separated is often applied. In such a case, it is useful to establish the pH value at which the contaminants are precipitated, prior t o isoelectric focusing. The polarity of the system is then chosen so that contaminating precipitates develop at the bottom of the column, below the focused zones analysed or isolated. Hence the precipitate sediments without causing coprecipitation with other proteins. When a heavy precipitate that blocks the outlet of the column is formed, stepwise elution from the top of the column is used.

PREPARATIVE ISOELECTRIC FOCUSING TECHNIQUE

335

Sample introducrion Recommended maximal amounts of buffer salt that can be present in the sample without affecting isoelectric focusing are 0.5 mmole for the 1 10-ml column and 1.5 mmole for the 440-ml column (LKB) [32]. A larger amount of buffer salt increases the column conductivity and leads to an undesirably long focusing time and deformation of the zones. There are three major ways of introducing the sample. The first is inherent to isoelectric focusing, e.g., its concentrating ability. For the preparation of the density gradient, an unconcentrated desalted sample solution is used instead of water; thus, the sample is present throughout the whole content of the column. The disadvantages of this method of sample application are (1) that part of the sample comes into contact with the electrode solutions (an acidic and a basic solution) and ( 2 ) that proteins pass through unfavorable pH values which may cause their denaturation. In the second method of sample application the concentrated protein sample is introduced at a position in the density gradient that is expected to be close to the isoelectric point of the protein to be separated. The density of the protein sample has to be in agreement with the density gradient. The third method is used when prolonged exposure of the sample may influence the stability of the protein. In such a case, the sample is applied after the pH gradient has been formed. As in the previous methods of application, the sample density has to be adjusted as required. This method of sample introduction is suitable for minimizing the exposure time and for bringing about a rapid separation, and is used in two-step preparative isoelectric focusing: in the first step the contaminants move away rapidly in a short time and the zone containing the separated protein is used again in the second step of isoelectric focusing, where a narrow pH range is used. The running time and electrical conditions are chosen as mentioned above. If the water temperature is 4"C, the maximal voltage and electric power are 1600 V and 15 W for the 110-ml column and 2000 V and 30 W for the 440-nil column (LKB) [22]. The construction of very short columns with a vertical cooling system, in which the solution is stabilized by means of a strong density gradient, makes it possible to shorten considerably the running time [7,41]. Emptying of isoelectric focusing columns At the end of the run, after the voltage has been switched off, the electric field strength will reduce to zero; the balance between electric migration and diffusion is thus broken and broadening of the focused zones due to extensive diffusion occurs. Therefore, it is recommended that the elution time is reduced as much as possible. A hydrostatically forced elution does not offer a constant flow-rate owing to viscosity differences along the column. A more controlled method involves stabilizing the column level by pumping a certain volume of distilled water on to the top of the column. Elution flow-rates of 60 ml/h for the 1 10-ml column and 240 ml/h for the 440-ml column (LKB) are recommended. When dealing with crude samples that can block the outlet of the column owing to the formation of heavy precipitates, stepwise sectioning of the gradient from the top by means of narrow polyethylene tubes is recommended. Precipitates, which often have a

336

PREPARATIVE ISOELECTRIC FOCUSING

tendency to stick to the column walls, are liberated from the walls without control or they are dissolved as a result of the pH change. Thus, an undesirable fusion of already separated protein zones occurs. Siliconization of the column walls may prevent this effect [421. Flat-bed isoelectric focusing in granulated gel

Technical equipment Flat-bed isoelectric focusing in granulated gel is the second most frequent technique of preparative isoelectric focusing, owing especially to its simple and rapid performance [28,43,44]. This method of preparative isoelectric focusing does not affect the precipitation of proteins during separation; this is an advantage, especially when working with crude material. The high loading capacity of the gel bed permits up to l o g of protein mixture to be introduced [45]. The equipment for flat-bed isoelectric focusing in granulated gel (manufactured by Desaga, Heidelberg, G.F.R., and LKB, for example) consists of a horizontal trough with a glass plate and a cooling block. For large-scale work, equipment with dimensions of 20 x 20 x 1 cm and 40 x 20 x 1 cm (Desaga), which is also useful for analytical purposes, is available [28] (Fig. 15.3).

1

1

1

1

I

1

!

Y

a b s d e Fig. 15.3. “Double chamber” for flat-bed isoelectric focusing in a granulated gel (Desaga). a: Filterpaper pad soaked with electrode solution; b: cooling block; c: bottom glass plate; d: gel layer; e: trough.

Granulatedgel material Sephadex G-75 “Superfine” and G-200 “Superfine” gels (Pharmacia, Uppsala, Sweden) and Bio-Gels (Bio-Rad Labs.) are the commonly used gel-bed materials. The most convenient is Sephadex G-75 “Superfine”, owing to its weak electroendosmotic effect and good mechanical stability. In gel-bed preparation it is necessary to expect that in

PREPARATIVE ISOELECTRIC FOCUSING TECHNIQUE

337

commercial preparations various amounts of charged water-soluble contaminants could be present which can interfere with the pH gradient formation. Grandated gel is prepared by briefly swelling an appropriate Sephadex gel in distilled water (1Omin); after efficient washing with water and absolute ethanol, the gel is dried in vacuo [46]. This technique is successful when the experimental conditions are standardized. As the swelling properties of the Sephadex gel may vary from batch t o batch, the final degree of evaporation for each of them must be known. An initial gel slurry (5%, w/v) is prepared by adding the gel to the carrier ampholytes. A homogeneous suspension is introduced into a trough and the water content is decreased to the calculated volume by evaporating excess of water in a light stream of air. The trough is placed on a cooling block and both ends of the gel bed are connected with electrodes by means of strips of filter-paper. These strips are soaked with appropriate electrode solutions (e.g., 1 M H2S04,2 M ethylenediamine). The pH range of the carrier ampholytes should be chosen in such a way that the isoelectric point of the protein to be separated is in the middle of the pH gradient.

Resolution and loading capacity In this technique, loading capacity is defined as the amount of protein (mg) per millilitre of the gel bed; a loading capacity of 5-10 mg/nil has been selected for protein mixtures, irrespective of the pH range. For a single protein the capacity represents 0.25- 1 mg/ml when using the pH range 3-10 and 2-4 mg/ml when using narrower pH ranges. The separation of 10 g of a highly heterogeneous enzyme, Pronase E, provides a good example of the high resolution and loading capacity (1 2 mg/ml) of the technique [45]. The protein recovery in flat-bed gel isoelectric focusing (indicated by absorbance measurements) varies between 80 and 90% for a single protein and protein mixtures. In cases when the pls of separated proteins are close and when it is necessary to use a shallow pH gradient, granulated gel beds are useful for obtaining narrower pH ranges (less than 1.5 pH unit). The electrical conditions depend on the gel suspension volume, sample composition and pH range chosen. According to the efficiency of the flat-bed water cooling (2-10°C), the focusing time is usually between 14 and 16h. Sample introduction

Before being applied, the protein sample is desalted as mentioned above, Large sample volumes can be simply included in the original gel suspension. However, the use of certain pH values can decrease the isoelectric focusing recovery. In an open system the protein sample can be introduced as a zone in any position in the gel bed, either before or after establishing the pH gradient. When the separated protein is by nature labile, the running time should be as short as possible; it is suitable to introduce the sample as a zone after the pH gradient has been established. Precipitates that might occur do not interfere with the run even if a large amount of crude protein mixture is applied. These precipitates are trapped within the gel and do not affect the separation.

338

PREPARATIVE ISOELECTRIC FOCUSING

Protein detection Apart from the detection possibilities mentioned above, a special printing technique has been established in order to reveal the protein spectrum focused in the gel bed [45]. Small amounts of proteins are adsorbed by the filter-paper placed on the top of the gel layer, and the filter-paper is then dried and stained by one of the usual protein-staining techniques. The protein spectrum can be evaluated in situ by densitometry using transmission at 280 nm, a silica plate being used instead of the bottom glass plate [45]. Pvotein elution After the printing method has revealed the positions of the zones, the gel bed is sectioned by a fractionating grid. Major protein zones, which are visible in the gel bed as transparent bands, are additional points of orientation in sectioning. In single gel-bed sections, the pH can be directly measured immediately after the isoelectric focusing has been terminated. Gel sections are transferred into small columns and proteins are eluted with small amounts of buffer. By changing the ionic strength and pH, the eventual precipitate can often be dissolved without a decrease in biological activity. Zone convection isoelectric focusing Principle Zone convection isoelectric focusing is a technique in whch ampholytes are fractionated in a free solution [ 17,47491. During stationary electrolysis in non-stabilized or incompletely stabilized liquid, inhomogeneities in the density of the solution occur and vertical density gradients of dissolved ampholytes are formed. An apparatus has been constructed on the basis of these observations that uses a density gradient of ampholytes for liquid stabilization against thermal convection. Technical equipment The apparatus consists of two separated cooling parts, the trough and the lid (Fig. 15.4) [17]; the space between them forms a series of “U-tube units” (Tiselius cells). The trough is filled with carrier ampholytes of the required concentration and pH range. The level of the liquid in the compartments rises after fitting the lid and thus communication between them is ensured. The level of the surface of the liquid will be lowered after lifting the lid at the end of the separation; thus communication between the compartments will be cut off and the risk of broadening of focused zones is eliminated. The pH is measured directly in the apparatus before the contents of single units are soaked out. The large cooling surface permits the use of a high voltage (800-2400 V); in spite of this, the running time is usually longer (about 48 h). The best stability of the system has been found to be at the temperature of the maximal density of water (about 4°C) [50]. Zone convection isoelectric focusing in free solution can be achieved in an apparatus

PREPARATIVE ISOELECTRIC FOCUSING TECHNIQUE

a

IW

b

d

1 C

339

Fig. 15.4. Zone convection isoelectric focusing apparatus [ 171. a: Upper part of the apparatus (lid); b: bottom of the apparatus (trough); c: running space; d: cooling water.

that consists of a coil of polyethylene or glass tubing where each turn corresponds to one compartment [51,52]. However, fraction collection in this system is difficult.

Resolution and loading capacity The resolution and the loading capacity depend on the dimensions of the apparatus, the number of compartments and the pH range of the carrier ampholytes. If the pH gradient is linear and none of the zones occurs in more than two compartments, then the resolving power of the system can be established by dividing the number of pH units applied by half the number of compartments in the apparatus. It is understood that the resolving power is defined as the minimal difference between the p l s of two proteins that can be separated [ 171. Zone convection isoelectric focusing allows one to obtain narrow pH ranges of carrier ampholytes, which are necessary for increasing the resolution and loading capacity in other preparative techniques.

Sample introduction In contrast to other isoelectric focusing techniques, the protein sample can often be used directly without previous purification or dialysis and it can be introduced into any compartment either before or after establishment of the pH gradient (after lifting the lid and soaking out an appropriate volume of carrier ampholytes). Zone convection isoelectric focusing enables a larger amount of crude material to be introduced; proteins are isolated with nearly 100%recovery [17]. Protein precipitation, if it happens to occur, does not affect the isoelectric focusing run; the precipitate rests at the bottom of the compartments, where it can be dissolved by changing the pH and ionic strength. Continuous flow isoelectric focusing

Principle This technique originates from continuous flow electrophoresis [53,54] where the electric field is orientated at right angles to the direction of electrolyte flow: proteins drift

340

PREPARATIVE ISOELECTRIC FOCUSING

away under the influence of the liquid flow of the electrolyte and migrate towards the corresponding electrodes. The ratio of the liquid flow t o the electrophoretic migration must be constant. In continuous flow isoelectric focusing, ampholytes migrate in the electric field until they reach their isoelectric points; here they have zero electrophoretic mobility and move only in the direction of liquid flow. Stability of the system is ensured either by a laminar flow of the electrolyte in the narrow space (0.3-0.5 mm) between two flat glass plates (free flow) [16,55] or by the capillary or density gradient systems [56]. In free-flow apparatus it is not possible to decrease the liquid flow-rate below a certain limit in order to maintain the stability of the system (the effects of convection and gravitation must be minimal). When proteins and ampholytes migrate to p1 positions slowly (narrow pH ranges), capillary systems (Sephadex G-100 beads, graded particles of polyacrylamide gel) or a density gradient (sucrose) are preferred, because in these systems there are no limitations with regard to minimal flow-rates [56].

Technical equipment Modified apparatus for continuous flow electrophoresis is used [53,57], consisting either of two parallel glass cooling plates (e.g., 23 x 30 x 0.3 cm) filled with an appropriate anticonvection medium (Sephadex, polyacrylamide, sucrose) or of two parallel glass cooling plates closer together (0.3-0.5 mm) and without an anticonvection medium (freeflow apparatus). The electrode compartments are separated by semipermeable membranes (Fig. 15.5). The construction of some apparatus copes with great differences in conductance between ampholytes in their mixed state and in focused conditions [56]. A constant flow-rate of the carrier ampholytes and the protein sample through the inlet tubing is usually maintained by a multi-channel peristaltic pump. The flow-rate of the appropriate electrode solutions (dilute acid and base) is also regulated. In apparatus for continuous flow isoelectric focusing in a density gradient, the horizontal flow of carrier ampholytes and a single sucrose density gradient are regulated by a peristaltic pump in both the inlet and outlet tubings [56]. A large potential gradient (50-100V/cm) can be used because of the large cooling surface [16,56].

Resolution and loading capacity The loading capacity depends particularly on the flow-through time. Several grams of sample (up to 500 mg in a day) can be separated when a sufficiently long flow time is used [56]. The limit of the loading capacity has not been clearly determined. As in previous isoelectric focusing techniques, the resolving power depends on the pH range applied; also, it depends on the number of outlet tubings of the cell. Protein zones can be collected without a decrease in resolution because the pH gradient is influenced by the electric field up to the outlet tubings. Precipitates formed in capillary stabilized systems prevent the smooth flow of the electrolyte through the cell. In this event, solubilizing agents or a density gradient can be used. The main advantage of capillary stabilized systems is that it is not necessary to eliminate proteins from density gradient material.

34 1

CONCLUSION

-

1

/

4

a h -~

I I

1

t

1

+I

Fig. 15.5. Scheme of an apparatus for continuous flow isoelectric focusing [ 5 6 ] .a: Electrode compartment; b: semipermeable membrane; c: running space.

Sample introduction In continuous flow isoelectric focusing, the protein sample is introduced continuously a t a chosen location in the focusing cell. I n contrast t o continuous flow electrophoresis, the protein sample can be supplied over the whole width of the cell. The outlet position is not changed for several days (as has been found in separating coloured proteins) [56].

CONCLUSION Isoelectric focusing is one of the most sensitive methods for studying proteins on both analytical and preparative scales, owing t o its ability t o distinguish proteins according t o their characteristic isoelectric properties. A high resolution and loading capacity together with a high recovery are the major advantages of preparative isoelectric focusing techniques in general. The most popular preparative isoelectric focusing technique is that with an increasing density gradient. This technique allows one t o introduce a large amount of material t o be separated (up t o 7 . 3 g) and successfully eliminates protein precipitation. The commercial availability of suitable equipment is also an advantage. Flat-bed isoelectric focusing in granulated gel has the advantage of simplicity, a relatively short running time, the possibility of introducing larger amounts of sample (up t o 10 g) and rapid evaluation of the protein pattern (print detection technique and the possibility of using densitometry in situ). Appropriate technical equipment is commercially available. The other techniques, which have certain disadvantages with respect t o protein precipitation (continuous flow isoelectric focusing) and longer running times (zone convection isoelectric focusing) are at present being further developed.

PREPARATIVE ISOELECTRIC FOCUSING

342

REFERENCES

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

H. Svensson,Acta Chem. Scand., 15 (1961) 325. H. Svensson, Acta Chem. Scand., 16 (1962) 465. 0. Vesterberg, Ann. N. Y.Acad. Sci., 209 (1973) 23. P. Blanickf and 0. Pihar, Collect. Czech. Chem. Commun., 37 (1972) 319. H. Svensson,J. Chromatogr., 25 (1966) 266. H. Rilbe, Ann. N. Y. Acad. Sci., 209 (1973) 11. H. Rilbe and S . Pettersson, in J. P. Arbuthnott and H. A. Beeley (Editors), Isoelectric Focusing, Butterworths, London, 1975, p. 44. 0. Vesterberg, Sci. Tools, 16 (1969) 24. W. D. Brown and S . Green, Anal. Biochem., 34 (1970) 5 14. P. Nilsson, T. Wadstrom and 0. Vesterberg, Biochim. Biophys. Acta, 221 (1970) 146. 0. Vesterberg, Protides Biol. Fluids, 17 (1970) 383. H. Haglund, Sci. Tools, 14 (1967) 17. B. 1. Radola, Biochim. Biophys. Acta, 194 (1969) 335. H. Delincde and B. J. Radola, Biochim. Biophys. Acta, 200 (1970) 404. D. Wellner and M. B. Hayes, Ann. N. Y. Acad. Sci., 209 (1973) 34. N. Seiler, J. Thobe and G. Werner, Hoppe-Seiler’sZ . Physiol. Chem., 351 (1970) 865. E. Valmet, Sci. Tools, 15 (1969) 8. A. McCormick, L. E. M. Miles and A. Chrambach, Anal. Biochem., 75 (1976) 314. 0. Vesterberg, T. Wadstrom, K. Vesterberg, H. Svensson and B. Malmgren, Biochim. Biophys. Acta, 133 (1967) 435. T. Wadstrom, Biochim. Biophys. Acta, 147 (1967) 441. S . Jacobs, Analyst (London), 98 (1973) 25. H. Lundin, S . G. Hjalmarsson, H. Davies and H. Perlmutter, LKB Application Note, No. 194 (1975). M. Jirka and P. Blanickf, Biochim. Biophys. Acta, 295 (1973) 1. N. Catsimpoolas, FEBS Lett., 4 (1969) 259. S. Jacobs,Protides Biol. Fluids, 18 (1971) 499. N. Ui, Biochim. Biophys. Acta, 257 (1972) 350. T. Wadstrom, Biochim. Biophys. Acta, 168 (1968) 228. B. J. Radola, Biochim. Biophys. Acta, 295 (1973) 412. J. C. Fawcett, in P. G. Righetti (Editor), Progress in Isoelectric Focusing and Isotachophoresis, North-Holland, Amsterdam, 1975, p. 99. N. Ui, Biochim. Biophys. Acta, 229 (1971) 567. N. Ui, Biochim. Biophys. Acta, 257 (1972) 350. A. Winter and C . Karlsson, LKB Application Note, No. 219 (1976). H. Svensson, Arch. Biochem. Biophys., Suppl., 1 (1962) 132. C. Earland and D. B. Ramsden,J O?romatogr.,35 (1968) 575. 0. Vesterberg and B. Berggren,Ark. Kemi, 27 (1966) 119. 0. Vesterberg, Methods Enzymol., 22 (1971) 387. E. Ahlgren, K. E. Eriksson and 0. Vesterberg, Acta Chem. Scand., 21 (1967) 937. G. V. Sherbet, M. S . Lakshimi and K . V. Rao, Exp. Cell Res., 70 (1972) 113. E. M. Leise and T. LeSane, Prep. Biochem., 4 (1974) 395. J. C. Janson, Acta Univ. Ups., 1972. H. Rilbe, Ann. N. Y . Acad. Sci., 209 (1973) 80. C. J. Smith and J. P. Arbuthnott, J. Med. Microbiol., 7 (1974) 41. B. J. Radola, Biochim. Biophys. Acta, 194 (1969) 335. B. J. Radola, Ann. N.Y. Acad. Sci., 209 (1973) 127. B. J. Radola, Biochim. Biophys. Acta, 386 (1975) 181. A. Winter, H . Perlmutter and H. Davise, LKBApplicution Note, No. 198 (1975). H. Rilbe,Protides Biol. Fluids, 17 (1970) 369.

REFERENCES 48 49 50 51 52 53 54 55 56 57

E. Valmet, Protides Biol. Huids, 17 (1970) 401. C. Rose and N. M. G. Harboe, Protides Biol. Fluids, 17 (1970) 397. A. Tiselius, Trans. Faraday Soc., 33 (1937) 524. V. Macko and H. Stegeman, Anal. Biochern., 37 (1970) 186. J . Bours,Exp. Eye R E S . 16 , (1973) 501. H. Svensson and I. Brattsten, Ark. Kemi, 1 (1949) 401. W. Grassmann, Z . Angew. Chern., 62 (1950) 170. 2. Pruslk, J. Chrornatogr.,91 (1974) 867. J. C . Fawcett, Ann. N. Y. Acad. Sci., 209 (1973) 112. 11. C. Mel, J. 7’heor.Biol., 6 (1964) 307.

343

This Page Intentionally Left Blank

Chapter 16

Preparative isotachophoresis P. JUST SVENDSEN

CONTENTS Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals and solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solutions.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practicalapplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Refercnces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

345 346 346 346 348

355 362

INTRODUCTION The separation of proteins on a preparative scale by employing the isotachophoretic principle [ 1-31 has many advantages over conventional zone electrophoresis. In zone electrophoresis the sample must be applied highly concentrated in a thin uniform application zone prior to electrophoresis. During electrophoresis, free diffusion will cause the bandwidth of the separated zones t o increase, and the resolution may decrease drastically. Moreover, the sample molecules will pass through a gel matrix over a period of time that is inverseiy proportional to their mobility, i.e., slowly migrating molecules are eluted in considerably larger volumes than rapidly migrating molecules. Ornstein [4 J and Davies [ 5 ] introduced “steady-state stacking” as a method for concentrating the sample into a thin starting zone as the initial step of the electrophoresis experiment, and in a second step converted the system into conventional zone electrophoresis. When employing polyacrylamide gel as an anticonvection medium, the method is known as disc electrophoresis. Steady-state stacking and isotachophoresis are identical: the sample molecules are concentrated between a rapidly migrating ion (the leading ion), with a mobility higher than that of any of the sample molecules, and a slowly migrating ion (the terminating ion), with a mobility lower than that of any of the sample molecules. A buffering counter ion with a charge of sign opposite to that of the leading ion, the sample molecules and the terminating ion, is present throughout the system. At equilibrium, the sample molecules are arranged into a series of consecutive zones according to decreasing mobility, in immediate contact with each other and all migrating at the same velocity. The concentration of a constituent in a zone is constant and a function of the concentration of the leading ion, and the mobilities of the leading ion, the constituent and the counter ion. Svendsen and Rose [6] introduced the use of Ampholine carrier anipholytes as “spacers” in isotachophoresis. Ampholine is a mixture of polyamino and polycarboxylic acids displaying a wide range of p l values and mobilities that coincide with and fall

346

PREPARATIVE ISOTACHOPHORESIS

between those of, e.g., human serum proteins. When Ampholine is stacked, a smooth migrating pH gradient is formed, within which sample molecules migrate until they reach a region with a pH where they attain a net mobility that matches the net mobility of that region. In principle this is very similar to isoelectric focusing, in which sample molecules are separated according to their PI, or in fact to zero net mobility. By employing the isotachophoretic principle for preparative purposes, the drawbacks and difficulties mentioned above are completely eliminated. In this chapter, the technique of preparative isotachophoresis using Ampholine carrier ampholytes is described, including the design of the apparatus, the chemicals and the solutions, and also the influence that a systematic variation of the parameters has on the performance of the method.

CHEMICALS AND SOLUTIONS Chemicals When the following chemicals are used, good results are obtained in preparative isotachophoresis: Acrylamide, N,N’-methylenebisacrylamide (Bis), N,N’-diallyltartardiamide(DATD), N,N,N‘,N’-tetramethylethylenediamine (TEMED), riboflavin-5‘-phosphoric acid monosodium salt (FMN) dhydrate, E-aminocaproic acid (EACA), 2-(N-morpholino)ethanesulphonic acid (MES), 2-amino-2-(hydroxymethyl)propane-1,3-diol (Tris), reagent grade, glacial acetic acid, p.a., orthophosphoric acid, p.a. (85%), sulphuric acid, p.a., ammonium peroxydisulphate, p.a., Ampholine carrier ampholytes, KAWI (mixed-bed ionexchange resin). The reagents should be of the highest purity available.

Solutions Gel buffers (leading electrolyte stock solutions): Glacial acetic acid 3.0 ml pH 4.0 Tris 1.o g Distilled water to 100 ml TEMED 300 pl pH 4.35 Glacial acetic acid 3.0 ml 1.7 g Tris Distilled water to 100ml TEMED 300 pl Glacial acetic acid 3.0 ml pH 4.8 Tris 3.2 g to 100ml Distilled water TEMED 300 p1 pH 6.0 MES 7.3 g Tris 1.5 g Distilled water to 100ml TEMED 300 pl

CHEMICALS AND SOLUTIONS

MES Tris Distilled water TEMED pH 6.25 1 M H3P04 Tris Distilled water TEMED pH 7.05 1M H$04 Tris Distilled water TEMED pH8.1 lMHQ04 Tris Distilled water TEMED pH8.5 lMH3P04 Tris Distilled water TEMED The pH must be checked water (at 25°C). pH6.2

347

7.3 g 2.0g to l0Oml 300 pl 30 ml 4.0 g to l0Oml 300 p1 30 ml 6.0 g to l 0 0 m l 300 pl 30 ml 12.0g to l 0 0 m l 300~1 30 ml 22.0 g to l0Oml 300 1 after the addition of TEMED and diluting 1 : 9 with distilled

Elution electrolyte and anode buffer as used: pH 7.1 1 M H2S04 121 I d Tris 32.0g Distilled water to 4000 nil Terminating electrolyte and cathode buffer as used : pH8.9 EACA 60.0 g Tris 3.0 g Distilled water to 2000 rnl Gel solutions (monomer): T40 C5 Acrylamide 38.0 g DATD 2.0 g Distilled water to 100 ml T40 C3 Acrylamide 38.8 g Bis 1.2g Distilled water to 100rnl The gel solutions are treated batchwise with 20g of KAWI resin for I5 min, filtered through filter-paper and then stored in dark bottles in a refrigerator at 4°C for several weeks.

348

PREPARATIVE ISOTACHOPHORESIS

Catalyst solutions: No. 1 Riboflavin-5’-phosphate 6.0 mg Distilled water to l 0 0 m l No. 2 Ammonium dioxypersulphate 200 mg Distilled water t o l0Oml The catalyst solutions are freshly prepared weekly and stored in dark bottles at 4°C. Gel mixture: 6.25 ml T40 C5 (DATD) or T40 C3 (Bis) Gel buffer stock solution 5.oml Catalyst No. 1 5.0 ml Catalyst No. 2 5.0 ml Distilled water to 50ml Before polymerization, the gel is overlayered with distilled water (5 mm). The gels are polymerized by illumination between 2 LKB-Polylite units for 60min at the temperature at which the experiment is to be run. When using the acidic gel buffers (pH 4.0-4.8), illumination should be continued for 3 h . Polymerized gels can be stored for several weeks at 4°C if the overlayered distilled water is replaced with gel buffer diluted 1 : 9 and the top of the separation tube is sealed with Parafilm.

APPARATUS The design of an apparatus developed in this laboratory [7] and made of Perspex and PVC takes advantage of the electrode arrangement by Bergrahm [8] and Svendsen [9], and is an improvement over the LKB Uniphor preparative electrophoresis apparatus with respect to handling, versatility and cooling. It also offers the possibility of inspecting all critical areas during an experiment. The elution system has been optimized according to the theory by Svendsen [ 101 using another design of the elution stopper by Bergrahm and Harlestam [I I]. Fig. 16.1 shows a cross-section of the apparatus. The apparatus is compatible with the base plate of the LKB Uniphor and is assembled as follows. The cooling jacket (1) is fastened to the clamp on the base plate. The lower end-piece (3) is attached to the c o o h g jacket by means of a threaded ring (28). The separation tube (2), containing the polymerized gel, is inserted and rests on the lower end-piece. The upper end-piece (4) is attached and when the threaded ring (22) is tightened a complete seal by means of O-rings is obtained. Circulation of the cooling water is started, through the inlet (1 2) and the outlet (14). The electrode houses (21) are attached to the endpieces (3 and 4) by means of threaded rings (31). A dialysis membrane (M) is placed between the end-pieces and the electrode houses. The electrodes (20) are inserted into the electrode houses and fastened with threaded rings (18). Circulation of electrode electrolytes is started through the inlets (IOa) and the outlets (lob) from cooled exterior electrolyte reservoirs. Terminating electrolyte is circulated through the upper electrode house and elution electrolyte is circulated through the lower electrode house. Connection to the power supply is obtained via pins (19). The separation tube is fdled with

349

APPARATUS

24

25

I

Fig. 16.1. Cross-section of universal column electrophoresis apparatus. 1 : Cooling jacket. 2: Perspex separation tube. 3: Lower end-piece. 4: Upper end-piece. 5 : Elution stopper. 6: (a) Conical clamp; (b) hollow screw. 7: Elution buffer inlet. 8: Outlet for elution buffer/air escape. 9: (a) Coupling for 10-mm silicone-rubber tubing; (b) coupling for 6-mm silicone-rubber tubing. 10: (a) Inlet for circulating electrode buffer; (b) outlet for circulating electrode buffer. 11: Inlet elution buffer/air escape. 12: Inlet for cooling water. 13: Cooling channel in end-piece. 14: Outlet for cooling water. 15: Inlet for terminating electrolyte. 16: Insert for capillary tube. 17: End stopper. 18: 'Threaded ring. 19: Pin for connection to power supply. 20: Platinum clcctrode. 21: Electrode house. 22: Threaded ring. 23: Stopper for holding and centering cold linger. 24: Cold finger (Pyrex glass). 25: Separation tube (Pyrex glass). 26: Inlet for cooling water. 27: Outlet for cooling water. 28, 29, 30, 31: Threaded rings. E: Elution chamber. M: Dialysis membrane.

350

PREPARATIVE ISOTACHOPHORESIS

APPARATUS

35 1

Fig. 16.2. Cross-section of the elution chamber. 1: Inlet for elution buffer/ai escape. 2: Ring for stretching the membranes, M. 3: Silicone support for 4 . 4 : Elution chamber outlet (polyethylene tubing). 5 : Gel. 6 : Gel support (porous polypropylene filter). 7 : Cooling channels. 8: Outlet for elution buffer/air escape. 9: Perspex separation tube. 10: Ring for stretching the gel support. (A) Membrane M is a dialysis membrane and ring 2 is made of porous PVC. (B) Membrane M is a porous polypropylene filter and ring 2 is made of Perspex. (C) Membrane M is a porous polypropylene filter and ring 2 is made of porous PVC. The buffer flow is indicated by arrows.

terminating electrolyte from a Mariotte flask through the inlet (15) and the sample, having a density higher than that of the terminating electrolyte, is applied on top of the gel by means of a syringe fitted with thin polyethylene tubing. The stopper (17) is inserted and fitted with a capillary tube (not shown), which is held in place by an insert (1 6) and threaded ring (29). The upper end-piece is filled through the inlet (1 5 ) with terminating electrolyte until all air has escaped through the capillary tube, which is then closed by means of clamped silicone-rubber tubing. The elution stopper (5) is prepared according to the type of elution technique to be employed. In the technique shown in Fig. 16.2B the following procedure is used. A porous polypropylene membrane is stretched over the end of the elution stopper by means of a solid Perspex ring ( 2 in Fig. 16.2B). Polyethylene tubing with one end moulded as shown (4) is pushed through a hole in the centre of the porous polypropylene membrane, the silicone-rubber membrane support (3), the conical clamp (6a) and the hollow screw (6b), which is then tightened. This causes the conical clamp to exert a firm grip on the polyethylene tubing. On being pulled, the polyethylene tubing stretches slightly and slides in the conical clamp, thereby sealing the porous polyethylene membrane on to the silicone-rubber support. The elution stopper is inserted into the lower end-piece and pushed upwards b y means of a threaded ring (30) until this engages the thread of the lower end-piece by one tum. The Mariotte flask containing the elution buffer is connected to the inlet (7) and silicone-rubber tubings fitted with clamps are

352

PREPARATIVE ISOTACHOPHORESIS

connected to the outlets (8 and 11). The polyethylene tubing (4) is connected to an LKB Perspex pump. The apparatus is tilted slightly to the right, the outlet (8) is opened and elution electrolyte will enter the lower end-piece through the inlet (7) and air trapped under the polypropylene membrane escapes through the outlet (8), which is then closed. The Perspex pump is started and, when the outlet (1 1) is opened, air trapped under the polyacrylamide gel matrix will escape through the outlet (1 l), which is left open until the elution stopper has been pushed all the way up and tightened by turning the threaded ring (30). The apparatus is brought to a vertical position and connected to the power supply (LKB 2103), an LKB Uvicord and a fraction collector (LKB Ultrorac). If the separation tube (2) is made of Perspex, a slight hydrostatic pressure should be applied over the gel bed. This is achieved by arranging the level of the Mariotte flask connected to the inlet (1 5 ) in the upper end-piece 15 cm above the level of the Mariotte flask connected to the inlet (7) in the elution stopper. When a glass separation tube (25) is used, the Mariotte flasks are arranged on the same level. Finally, the current is switched on. The cross-sectional area of the gel bed is 5.3 cm2 and, with a constant current of lOmA and a temperature of 5°C in the cooling water, the experiment runs for 22-24 h. With a separation tube ( 2 in Fig. 16.1) of length 20 cm the distance between the electrodes is 45 cm. Depending on the gel buffer used, the starting voltage is 200-400 V, increasing to 800-15OOV at the end of the experiment. Optimal cooling of the gel bed is achieved when a glass separation tube (25) is used together with a cold finger (2 1) also made of glass and centred by means of a stopper (23), which then replaces stopper 17. The cooling water inlet (26) of the cold finger is connected to the cooling water outlet (14) in the upper end-piece. The elution chamber (E in Fig. 16.1) is shown enlarged in Fig. 16.2, and the flow of the elution electrolyte is demonstrated for three different elution techniques. In Fig. 16.2A, the membrane (M) is a dialysis membrane and the ring (2) is made of porous PVC. The inlet (1) is connected to the elution electrolyte reservoir (Mariotte flask) and the polyethylene tubing (4) is connected to an LKB Perspex pump. Anode buffer is circulated below the membrane through the inlet (7 in Fig. 16.1) and the outlet (8). When the Perspex pump is started, elution electrolyte will enter through the inlet (1) and the porous ring ensures smooth sweeping of the elution chamber. At a flow-rate of 30 ml/h the elution chamber is swept once per minute. The bottom of the elution chamber must be funnel-shaped with an angle seen from the central outlet of 165". If the bottom is kept horizontal, severe remixing will occur. This technique can be used for macromolecules that do not absorb on or migrate through the dialysis membrane. Such problems are completely eliminated by employing the technique shown in Fig. 16.2B. The dialysis membrane is replaced with a porous polypropylene filter and the porous ring (2) with a solid Perspex ring. The inlet (1) is clamped after the air has escaped. The elution electrolyte reservoir is connected to the inlet (7 in Fig. 16.1). The outlet (8) is clamped when the air below the membrane has escaped. When the Perspex pump is started, elution electrolyte will enter the elution chamber from below through the porous membrane and sweep the elution chamber more efficiently than when using the dialysis membrane technique. At a flow-rate of 30 ml/h using Tris-sulphate as recommended, the upward flow in the polypropylene membrane exceeds the velocity at which sample molecules would migrate in the porous membrane

353

APPARATUS

B 1

I

Fig. 16.3. Cross-section of simplified column electrophoresis apparatus. 1 : Perspex separation tube. 2: Cooling jacket. 3 : Lower end-piece. 4: Upper end-piece. 5 : (a) Elution stopper; (b) view of elution stopper turned through 90". 6: Provision for accoinodating conical clamp and hollow screw (see 6a and 6 b in Fig. 16.1). 7: Inlet for circulating clcctrode buffer. 8: Outlet for circulating electrode buffer. 9: Pin for connection t o power supply. 10: Electrode in elution stopper. 11: Inlet for elution buffer. 12: Air escape, connection to manometer tube. 13: Inlet Tor cooling water. 14: Outlet for cooling water. 15: Pin for connection t o power supply. 16: Stopper. 17: Access to column. 18: Electrode. 19: Reservoir for terminating electrolyte. 20: Stopper for holding and centering cold finger. 21: Pin for connection to power supply. 22: Electrode. 23: Cold finger (Pyrex glass). 24: Separation tube (Pyrex glass). 25: Inlet for cooling water. 26: Outlet for cooling water. 27: Threaded ring. E: Elution chamber.

354

PREPARATIVEISOTACHOPHORESIS

Fig. 16.4. (A) Ferguson plots of serum proteins and narrow-range Ampholine [IS]. (B)Isotachophoresis in small tubes. Sample: proteins used for the Ferguson plot. Spacers: narrow-range Ampholine used for the Ferguson plot. (C) pH versus position in gel bed containing a migrating Arnpholine carrier arnpholyte gradient [ 61.

in that buffer. Therefore, n o bleeding occurs through the membrane, and even small molecules that would certainly diffuse through a dialysis membrane are collected. As the flow-rate of the elution electrolyte approaches zero at the periphery of the elution chamber, optimal elution is obtained by using the technique shown in Fig. 16.2C. The membrane is a porous polypropylene filter and the ring (2) is made of porous PVC. A second Perspex pump is connected to the inlet (1) after the air has escaped. If the Perspex pump connected to the polyethylene tubing (4)maintains a flow-rate of 30ml/h and the Perspex pump connected to the inlet (1) maintains a flow-rate of 6ml/h, a full sweep is completed every 2 min [lo]. When the performance of the elution technique in Fig. 16.2A is acceptable, which is the case for many proteins, the design of the electrophoresis apparatus can be greatly simplified. Such a design is shown in Fig. 16.3, and is based on the design of Svendsen [12]. The upper electrode (18) is placed directly into the terminating electrolyte reservoir (19) on top of the separation tube (1). The lower electrode (10) is placed directly in the elution stopper (5a and 5b). Elution electrolyte is circulated from a cooled external reservoir through the stopper via the inlet (7) and the outlet (8). Elution electrolyte from a Mariotte flask is conveyed to the lower end-piece (3) through the inlet (1 1) before the

PRACTICAL APPLICATION

355

elution stopper is tightened. Air escapes through the outlet (12) w h c h is connected to a silicone-rubber tubing serving as manometer for adjusting the level in the Mariotte flask containing the elution electrolyte t o match the level of the terminating electrolyte in the reservoir (1 9) in the upper end-piece (4). More details concerning these designs are shown in ref. 20.

PRACTICAL APPLICATION In Fig. 16.4A the mobilities of human serum transferrin and human serum orosomucoid are compared with that of narrow-range Ampholine preparations (a generous gift from LKB) by means of a Ferguson plot [ 131. Disc electrophoresis was performed in small tubes (5 mm I.D.) in increasing concentrations of polyacrylamide gel. The gel buffer with pH 8.5 was used in the separation gel and the gel buffer with pH 6.25 was used in the stacking gel. After electrophoresis, the proteins and Ampholine were precipitated by immersing the gels in picric acid-acetic acid-water (ref. 14, p. 27) containing 1% of lanthanum acetate [ 151. After 5 min, 6% (v/v) of a solution containing 0.25% (w/v) of Coomassie Brilliant Blue R250 in water was added, and the gels were left overnight at room temperature. The R E is calculated by dividing the migration distance of the protein bands and Ampholine bands by the migration distance of the boundary between the leading ion (phosphate) and the terminating ion (glycine). The detection of the boundary is easily affected, as the lanthanum ion will precipitate the phosphate, but not the glycine and Tris. Picric acid has been found to be the best fixative for Ampholine. The RE values are plotted against the total concentration of polyacrylanlide gel (%T) on logarithmic paper. The RE values for Ampholine 6.65 are estimated by recalculating data from an experiment with asparagine as terminator at the same operative pH. In this instance the curves were computed by means of a weighted linear regression analysis [ 161, and it is evident that the mobility range of Ampholine is compatible with the mobility range of human serum proteins up to a certain polyacrylamide gel concentration. The pH in the glycine zone behind the moving boundary is 9.25 (25”C), and the mobility of Ampholine in the p l range slightly below this pH must therefore be extremely low. Assuming that a Ferguson plot would exhibit the same slope for such Ampholine molecules (same molecular weight), the R E values will coincide with those of the proteins at considerably higher gel concentrations. This will be demonstrated below. Above RE = 1 .O the sample molecules migrate fully stacked, and below R E = 1 .O the sample molecules migrate unstacked or as bands in normal zone electrophoresis. If the glycine is replaced with EACA, transferrin will remain stacked at gel concentrations below ca. 15% and orosomucoid at gel concentrations below 20%, and the same applies to Ampholine within the plrange 3.5-9.5 (total range) as the pH in the EACA zone behind the boundary is 9.77 (25°C). The narrow-range Ampholines were then tested as “spacers” for isotachophoresis. Fig. 16.4B shows three experiments with a mixture of human serum transferrin and human seruni orosomucoid run in small tubes with Ampholine 6.65, Ampholine 8.15 and a 1 : 1 mixture of both. After electrophoresis, the gels were immersed in 3.5% perchloric acid and, after 5 min, 3% (v/v) of a solution containing 0.25% (w/v) of Coomassie Srilliant Blue

356

PREPARATIVE ISOTACHOPHORESIS

Fig. 16.5. (A) Computer input/output data for some phosphate systems with three different terminating ions (constituent data from ref, 21). (B) UV elution profiles of three preparative isotachophoretic experiments demonstrating the performance of different terminating ions. The apparatus shown in Fig. 16.1 was used.

(3150 was added [ 171, and the gels were left overnight. The Ampholine 6.65 spaces the orosomucoid but not the transferrin, the Ampholine 8.15 spaces the transferrin but not the orosornucoid, and the mixture spaces both proteins. The gel concentration was 4.8%, the leading electrolyte was Tris-phosphate (pH 8.5) and the terminating ion was EACA. In Fig. 16.4C the shape of the pH gradient obtained by preparative isotachophoresis in polyacrylamide gel is demonstrated. The leading electrolyte was Tris-acetate (pH 4.5) and the terminating ion was glycine. A I-ml volume of Ampholine 3-8 was added. The cross-sectional area of the apparatus was 3 cm’. The experiment was discontinued before the leading ion boundary had migrated out from the gel and after the terminating ion had entered the gel. The gel was then cut into ten equally large sections, which were eluted separately with distilled water. The curve shows the pH as a function of the position of the corresponding section of the gel bed. The discontinuity in the first section is caused by Tris-glycine buffer (used as elution electrolyte in this early experiment [ 6 ] ) that had diffused into the lower part of the gel bed. The choice of the terminating ion is very important. In Fig. 16.5A an example of computer input/output data is shown, using phosphate as the leading ion, Tris as counter ion and three different terminating ions: glycine, 8-alanine and EACA. The program used for the computation is a modified version of the Isogen program [2]. If we select the

PRACTICAL APPLICATION

357

system with a pH of 8.0 in the phosphate (leading ion) buffer, then the pH in the terminating ion zone will be 9.09 when glycine is chosen as the terminating ion. The net mobility of glycine at this pH is 6.34 (column: N-MOB). If fi-alanine is chosen as the terminating ion, the pH in the terminating ion zone behind the phosphate boundary will be 9.39. The net mobility of 0-alanine at this pH is 3.1 3, and 0-alanine is thus better than glycine for stacking slowly migrating molecules. The best choice, however, is E-aminocaproic acid (EACA). With EACA the pH in the terminating ion zone is 9.61 and the net tnobility is 1.67. In Fig. 16.5B the performance of the three terminators is demonstrated. A 1.6-ml volunie of Ampholine 7-9 was used in all three experiments. The three gels were cast at the same time in three Perspex separation tubes ( 2 in Fig. 16.1) at 4°C. Gel buffer of pH 8.1 was used. The sample was IgG (slow), prepared by salting-out and ionexchange chromatography. The sample was dialysed against the terminator before the Anipholine was added. The U V elution profiles were obtained with an LKB Uvicord 111 at 280 and 250 nm. It is evident that EACA is the best choice, as it is capable of eluting IgG (slow). This terminating ion is therefore used in all runs performed in this laboratory. Theoretically, the leading electrolyte should be preferred as the anode buffer and elution buffer, but when the pH of the leading electrolyte or the conductivity is low it becomes necessary to use a substitute. A problem also arises if expensive leading ions are employed (large volumes). Tris-sulphate at pH 7.1 can be used in all systems in which Tris is the counter ion and does not influence the performance of the systems described. This elution buffer is therefore used in all runs performed in this laboratory. As isotachophoresis is an equilibrium system, the reproducibility is very good. Two identical experiments were run on two different days. The sample was 30 mg of a haemoglobin prepared from human red blood cells. The gel buffer used was Tris-MES (pH 6.2), and 1.6 nil of Ampholine 6-8 was added to the sample. The result of the two runs is shown in Fig. 16.6. It can clearly be seen that even small details are satisfactorily reproduced. The U V elution profiles were obtained at 250 nm using an LKB Uvicord I. The concentration of a constituent within a stacked zone is constant, and consequently the length of the pH gradient in isotachophoresis using Ampholine carrier ampholytes as spacers is proportional to the amount of spacer material added. In four experiments the amount of sample was kept constant (30 mg of haemoglobin) and increasing amounts of Ampholine 6-8 were added. Tris-MES (pH 6.0) was used as the leading electrolyte. The UV elution profiles (LKB Uvicord I, 250 nm) of the four experiments are shown in Fig. 16.7A-D. Owing to the large amount of Ampholine in the run shown in Fig. 16.7D, the volume of the gel bed was doubled in order for the gel bed to be able to contain and equilibrate the very long gradient. The spacer action of the Ampholine is very convincing. The protein zones are affected by the increased amounts of spacers, and the distance between two protein peaks is also a function of the increased amounts of spacers. This indicates that the ranges of mobilities and p1values of Ampholine d o coincide with and fall between those of the proteins. The choice of Ampholine p l range must always be considered. Using Tris-acetate gel buffers (from pH 4.0 to 4.9) and Tris-MES gel buffers (from pH 5.9 to 6.9), the Anipholine is chosen so that the pZ values of the proteins to be separated are within the p l range of the Ampholine. This can be achieved by mixirig two or more Arnpholine p l ranges. Using Tris-phosphatc gel buffers (from pH 6.25 to 8.5), the Ampholine can be selected as

358

............................. .

A :. . ; . :. .. . :A w . ,

.

9.

PREPARATIVE ISOTACHOPHORESIS

.? % 7 ! & * % 1 '

1

,

,

- .

.

.

,

~.

I

+

tt -

-

-

-..

. .

Fig. 16.6. UV elution profile of two identical experiments performed to demonstrate the reproducibility of preparative isotachophoresis using the apparatus shown in Fig. 16.1 [ 181.

_ a

,:A

.......

,

I

.^a.

.,._, ........... "

1.3-"+*

u a a w

B

I

Ix)

/ I

'I

:

I

,

.

a

........................................ ..-_........ ,-*. *.,- ..... ~

L_

:D

Lt 3 I-"

c.

?e

* 34

a Ix)

Fig. 16.7. W elution profiles of four preparative isotachophoretic experiments performed to demonstrate the spacer function of Ampholine carrier ampholytes [ 181. Curves A, B and C are from identical experiments except for the amounts of Ampholine added. Curve D is from an experiment with twice the amount of polyacryhmide gel. The apparatus shown in Fig. 16.1 was used.

..

PRACTICAL APPLICATION

359

Fig. 16.8. W elution profiles of two preparative isotachophoretic experiments performed to demonstrate the spacer function of two different Ampholine ranges on the same sample. The apparatus shown in Fig. 16.1 was used.

above and then the p l range found is adjusted by adding 0.5 unit for gel buffer of pH 6.25, increasing to 2 units for gel buffer of pH 8.5. As the titration curves for Ampholine and proteins are different, adjustments may have to be made: if a protein is eluted in the front of the Ampholine gradient the plrange is lowered, and vice versa, if a protein is eluted in the rear of the gradient the pZ range is increased. The total amount of Ampholine is recommended t o be 0.3 ml(40%) per square centimetre of gel bed cross-sectional area with a length of the gel bed of 1Ocm. A current density of 2 mA/cm2 (constant current) is recommended at 5°C. The experiment then runs overnight. The spacer function of two different Ampholine ranges is shown in Fig. 16.8A and B. The sample consisted of gamma-globulins prepared by salting-out and ion-exclmge chromatography. The preparation also contained some ceruloplasmin. The gel buffer was Tris-phosphate (pH 8.1). The UV elution profiles were obtained with an LKB Uvicord 111 at 280 and 250 nm. In the experiment shown in Fig. 16.8A, I .6 ml of Ampholine 6-8 was added. The ceruloplasmin peak is marked with an arrow and is spaced only slightly apart

360

PREPARATIVE ISOTACHOPHORESIS

Fig. 16.9. W elution profdes of two preparative isotachophoretic experiments performed to demonstrate the resolution of the Same sample using (A) a gel buffer with low pH and (B) a gel buffer with high pH. The apparatus shown in Fig. 16.3 was used, fitted with a Pyrex glass separation tube and Pyrex glass cold finger.

from the gamma-globulins. In the experiment shown in Fig. 16.8B, 1.6 ml of Ampholine 7-9 was added and spaced the ceruloplasmin much further apart from the gammaglobulins. The separation of the gammaglobulins was also improved, as can be seen from the figure, but IgC and IgA were not separated in this experiment. At low pH, i.e., using gel buffers with low ionization of the leading ion, the proteins start to subfractionate into the genetic variants, as demonstrated in Fig. 169A. A gel buffer of pH 4.8 was used, the sample was human serum transferrin and 1.2 ml of Ampholine 5-7 was added to the sample. The elution profile was obtained with an LKB Uvicord I1 at 280nm. In Fig. 16.98, an experiment with an identical sample is shown. A gel buffer of pH 8.1 was used and 1.2 ml of Ampholine 6-8 was added to the sample. The elution profde was obtained with an LKB Uvicord I1 at 280 nm. Note that the transferrin is eluted as one peak, in contrast to the experiment shown in Fig. 16.9A, and occupies much less space in the pH gradient. This should be taken into consideration, as the risk of contamination with other proteins under such circumstances is much lower, and a

PRACTICAL APPLICATION

36 1

Fig. 16.10. UV elution profiles and irnmunochernical elution profiles of two preparative isotachophoretic experinients using the same sample and Ampholine range in different gel concentrations. (A) UV elution profile and (B) fused rocket immunoelectrophoresis of an experiment in 5% gel; (C) UV elution profile and (D) fused rocket immunoelectrophoresis of an experiment in 10%gel. The apparatus shown in Fig. 16.3 was used.

considerably higher yield is often obtained. Subfractionation should therefore always be performed after the protein has been purified. As mentioned above, it should be possible to perform preparative isotachophoresis with Ampholine carrier ampholytes at high gel concentrations by adding Ampholine with a high plrange. This is demonstrated in Fig. 16.10. Two experiments were run with the same sample and Ampholine, but at different gel concentrations. The sample was a mixture of human serum transferrin and human serum orosomucoid, and 0.8 ml of Ampholine 6-8 and 0.8 ml of Ampholine 8-9.5 were added to the samples. The gel concentration in the experiment shown in Fig. 16.10A was 5%) and in that shown in Fig. 16.1OC 10%. A gel buffer of pH 8.1 was used. The UV elution profiles were obtained with an LKB Uvicord I1 at 280 nm. Immunochenucal elution profiles were also obtained by means of fused rocket immunoelectrophoresis [ 191 using anti-human serum from rabbits (DAKO-immunoglobulins, Copenhagen, Denmark) (Fig. 16.1OB and D), and it can

362

PREPARATIVE lSOTACHOPHORESIS

be seen that the proteins were completely separated in both instances. Increasing the gel concentration, owing to the sieving effect, will cause the protein bands to shift to a position in the migrating Ampholine gradient with a higher pH, but the resolution is maintained. This can be utilized to facilitate the separation of proteins with different slopes in a Ferguson plot (different molecular weights). Preparative isotachophoresis as described above is most versatile for protein separation. It is easy to handle as sample application is not critical and it is further simplified as the elution and terminating electrolytes recommended are used with all of the gel buffers specified. It offers a high resolving power at high sample loads (75 mg/cm2 per band) and the resolution is easily optimized by systematic variation of the parameters. The practical application of preparative isotachophoresis with Ampholine carrier ampholytes as spacers is even simpler than conventional zone electrophoresis. A combination of gel buffer, Ampholine range and gel concentration that has once been demonstrated to solve a specific separation problem will work permanently owing to the high reproducibility of the method.

REFERENCES 1 F. Kohlrausch,Ann. Phys., 62 (1897) 209. 2 R . Routs, Electrolyte Systems in Isotachophoresis and Their Application to some Serum Separations, Thesis, Technical University of Eindhoven, Eindhoven, 1971. 3 T. M. Jovin,Biochemisrry, 12 (1973) 871,879 and 890. 4 L. Omstein, Ann. h! Y. Acad. Sci., 121 (1964) 321. 5 B. J. Davies,Ann. h! Y. Acad. Sci., 121 (1964) 404. 6 P. 3. Svendsen and C. Rose, Sci. Tools, 17 (1970) 13. 7 P. J. Svendsen, unpublished work 1970. 8 B. Bergrahm, Sci. Tools, 14 (1967) 34. 9 P. J. Svendsen, Anal. Biochem., 25 (1968) 236. 10 P. J. Svendsen, Sci. Tools, 19 (1972) 21. 11 B. Bergrahm and.R. Harlestam, Sci. Tools, 15 (1968) 26. 12 P. J. Svendsen,Protides Biol. Fluids, 17 (1970) 413. 13 K. A. Ferguson,Metabolism, 13 (1964) 985. 14 N. H. Axelsen, J. KrQll and B. Weeke, A Manual of Quantitative Immunoelectrophoresis: Methods and Applications, Universitetsforleget, Oslo, 1973. 15 P. J. Svendsen and A. Chrambach, in preparation. 16 D. Rodbard and L. Graber, Quantitative Polyacrylamide Gel Electrophoresis: Fortran IV Programs for Data Analysis, personal communications. 17 A. H. Reisner, P. Nemes and C. Bucholtz, Anal. Biochem., 64 (1975) 509. 18 P. J. Svendsen, Sci. Tools, 20 (1973) 1 . 19 P. J. Svendsen, in 2. Deyl (Editor), Electrophoresis. A Survey of Techniques and Applications. Part A : Techniques, Elsevier, Amsterdam, Oxford, New York, 1979, Chapter 7. 20 P. J. Svendsen, in 2. Deyl (Editor), Electrophoresis. A Survey o f Techniques and Applications. Part A : Techniques, Elsevier, Amsterdam, Oxford, New York, 1979, Chapter 14. 21 T. M. Jovin, M. L. Dante and A. Chrambach, Multiphasic Buffer Systems, Catalogue (PB-196090), NTIS,Springfield, Va.

Chapter 17

Preparative isotachophoresis on the micro scale L. ARLINGER

CONTENTS Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Small-scale preparative isotachophoretic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle of preparative capillary method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Capillary elution system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fraction collection device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resolution of fraction collection procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Desorption of sample from the strip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of preparative capillary method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biological activity assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Radioactivity measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lmmunoelectrophoretic identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enzyme identification by zymogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

363 364 365 365 367 368 370 370 371 373 374 376 376

INTRODUCTION The aim of this chapter is to survey the isotachophoretic methods used for the preparation of components on the micro scale. This immediately raises the question of what the terms micro and preparative mean. Traditionally, they have to some extent been regarded as contradictory, which is both confusing and erroneous. The purpose of a preparative procedure is to separate and collect sample components within whatever size range is chosen and with maximal yield and purity of the component of interest. By nucro scale we understand a total sample load in the microgram to milligram range and below. The development of isotachophoretic equipment during the last decade has mainly followed two lines, the analytical capillary column [ 1,2] and the preparative wide-column [ 1, 31 techniques. The sample size in the analytical technique has been a few nanomoles (or nanoequivalents) of each component, while the widecolumn technique has been able to separate up to several hundred milligrams of protein [4]. The wide-column preparative technique is considered to be a macro technique and has been thoroughly discussed in Chapter 16. The capillary technique used in the preparative mode will be discussed in detail in this chapter. A short survey will also be made of some different types of isotachophoretic separations that have been performed for (fully or partly) preparative purposes. There have been several publications in which sample sizes between those of the above two techniques are considered.

364

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

In this chapter, a basic knowledge of the isotachophoretic principles will be presumed. They can be found in Chapters 10 and 16.

SMALbSCALE PREPARATIVE ISOTACHOPHORETICMETHODS Vestemark’s early work with thin-layer techniques [5] included separations of, e .g., serum proteins and amino acids on cellulose powder and paper, and he was able to collect sections of interest after the bands had been revealed. Vestermark and Sjodin also later used cellulose acetate strips t o isolate radioactively labelled substances synthesized by enzymatic reactions with different sugars [6-81, using 1 0 - 2 0 ~ 1of the incubation mixtures as samples. In those studies the isotachophoretic technique was also used in a twodimensional combination with zone electrophoresis. Cellulose powder was also used by Eriksson [9] as a support for thin-layer isotachophoresis. He isolated and identified various compounds in haemolymphs of different insects, The small scale used enabled him to study metabolites in individual insects after elution of scraped-off sections with water. Agarose gel has been used as the supporting slab medium in isotachophoretic studies of proteins by Uyttendaele and co-workers [ 10,111. This medium is suitable for the immunoelectrophoretic technique that was used for the characterization of the separated bands. The agarose gel technique, however, was not used in a directly preparative manner, but it is likely to work as well as the above-mentioned techniques. The electroendosmosis generally caused by the agarose gel, however, has frequently been a limitation to its use outside the standard zone electrophoretic applications. This was one reason why Brogren [ 121 chose polyacrylamide as the gel slab material for protein separations by isotachophoresis. These were also combined with crossed immunoelectrophoresis and electrofocusing, and detection was effected by the usual staining procedures. Brogren and Peltre [13] also used Sephadex G-75 Superfine for separations in a granulated bed. The same fractionation procedure and equipment were used as for electrofocusing according to the procedure of Radola [ 141, and detection can be effected by a print technique. Granulated gels, however, have not so far been used for isotachophoresis on the micro scale. Previously, some aspects of isotachophoretic separations of proteins had also been studied using disc electrophoretic equipment, i.e., in polyacrylamide gel rods of diameter 5 mm [ 151. The traditional method of pushing out the gel from the glass tube, slicing and eluting was utilized for the sample collection. Finally, HjertCn’s rotating tube equipment [ 16, 171 can be used for preparative separations. The liquid fractions were carefully withdrawn from the tube after separation when the current had been switched off. The above methods for preparative isotachophoresis on a small scale are basically analytical techniques used for preparative purposes. The usefulness of the thin-layer methods depends on the detection methods that can be applied, eg., fluorescence, radioactivity or visible colours of sample ions or enclosing marker substances [18, 191. The sample fractions are collected by cutting, scraping and subsequent elution

PRINCIPLE OF PREPARATIVE CAPILLARY METHOD

365

procedures, frequently giving low reproducibility and poor yields. In many instances there will be n o detection methods available before elution, e.g., for proteins that would irreversibly denature by t h e staining procedures generally used. Further, the sample zones can be skewed because of inhomogeneities in the support media, imperfect cooling capacity or disturbances caused b y evaporation or electroendosmosis. The skewness will make fractionation considerably more difficult (sometimes necessitating staining of several slices in order to establish the position of a band) and adversely affect the resolution obtained during the separation phase. In the small cylindrical column methods, both the gel tubes and the rotating quartz tube should be scanned prior t o fractionation in order t o find the positions of the sample components of interest. The final resolution o f the separation with the gel rods depends on, among other factors, the thickness of the slices, skewness of the zones and diffusion times. When fractionating with the rotating tube equipment, switching off the electric current through the free liquid phase allows diffusion t o decrease tlie maximal resolution obtained during the separation phase. Also, both the simple gel rod and the sophisticated rotating tube method are small-scale rather than micro-scale methods. Although also larger than the micro scale, there have been some developments to continuous flow instruments f o r truly preparative purposes [30,211. They might be optimized for smaller amounts of sample, but the instrunlent concept is complex and generally sensitive to disturbances.

PRINCIPLE OF PREPARATIVE CAPILLARY METHOD There is increasing interest in biochemistry and medicine in separating and obtaining pure fractions from micro amounts of samples. The isotachophoretic separations obtained in the free liquid phase in a capillary have demonstrated the capability of producing pure component zones with very high resolution [ 721. Small amounts of sample can generally be separated in a short time. Therefore, it was o f interest t o convert the analytical capillary method into a preparative method, provided that the high resolution could be maintained and little increase in the time and labour involved. The remaining part of this chapter describes the preparative capillary isotachophoretic procedure and equipment. It was found that the same high resolution was obtained in the preparative mode as in the analytical niode, the separation time was not increased and the UV detector permitted complete supervision of the fraction collection.

Capillary elution system It is well known that the mixing caused by a flow in a capillary tube is severe. The diffusion in the liquid across a sharp boundary between adjacent isotachophoretic zones quickly destroys the resolution as soon as the electric field is switched off. Therefore, fractionation must be performed while the voltage gradient in the separation capillary is still on. The stabilization capability of the strong voltage gradient, however, allows a more crude elution procedure to be used without distortions, as compared with the high demands that have t o be put on cylindrical symmetry and even flow distributions in widecolumn elution equipment [4].

366

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

a

b Ll

Y

L'

Fig. 17.1. Principle of preparative capillary isotachophoresis. (a) Schematic diagram of the basic analytical equipment; (b) the preparative version with the additional flow of leading electrolyte. L = leading electrolyte, T = terminating electrolyte, D = UV detector site, S = sample ions, v = migration velocity, m , , m 2= semipermeable membranes,fL = flow of L to elute the sample zones.

capillary column wall

L

1 mm

J

Fig. 17.2. T-connection unit used for sample elution in the LKB Tachophor capillary column. The flow of leading electrolyte from a precision syringe pump sweeps out the migrating sample ions through the T-connection, and is collected on the moving cellulose acetate strip.

367

PRINCIPLE OF PREPARATIVE CAPILLARY METHOD buffer and sample

t

,T-connection

unit

-

leading electrolyte flow

ionic migration

thermostatted liquid UV-detector slit’

Fig. 1 7 . 3 . Part of the capillary column at the UV detector site of the Tachophor. The figure shows the mounting of the T-connection unit in the thermostatted capillary column plate.

The principle of preparative capillary isotachophoresis is shown schematically in Fig. 17.1, As in the analytical mode (Fig. 17.la), the leading electrolyte reservoir is separated from the capillary compartment by a semipermeable membrane (m,) in order to prevent hydrodynamic flow. In the preparative modification, a second similar membrane (mz) is inserted between the terminating electrolyte reservoir and the capillary. This is to ensure that the eluting counter flow (fL) of leading electrolyte will leave the capillary through the small T-connection at the detector site without causing any zonedisturbing flow in the separation capillary column. After passing the UV detector (D), the migrating zones will be swept out through the T-connection (see Fig. 17.1b) if the liquid flow-rate (fL) is greater than the corresponding migration rate (v)of the ions. The importance of balancing the liquid flow and migration velocity for quantitative elution is described in detail elsewhere [23]. As liquid flow will dilute the sample zones, the flow-rate is set only a few per cent higher than the migration rate. The design of the elution system used in the LKB 2 127 Tachophor capillary isotachophoretic equipment is depicted in Fig. 17.2. It consists of a T-connection unit clamping the PTFE capillary separation tube. The inner and outer diameters of the tubing are 0.45 and 0.75 mm, respectively. A capillary channel 0.5 mm long, with a diameter of 60pm, is drilled in the T-connection unit and through the capillary wall. The sample zones are swept out through this orifice by the counter flow of the leading electrolyte. The capillary is mounted in a loop (see Fig. 17.3). The T-connection is positioned approximately 2 cm behind the UV detector, so that virtually no changes in the separation pattern take place during the short passage between detection and fractionation. Fraction collection device Isotachophoretic zones in the separation capillary can have volumes as small as 10-20 nl [22], which means that traditional methods of sample collection are impossible. The flow of sample ions eluted by the leading electrolyte is collected as a continuous stream on a cellulose acetate strip, passing over the T-connection outlet at a distance of less than 0.1 mm (see Fig. 17.2). The voltage difference between the capillary outlet and the strip contributes to a smooth and even collection of liquid, giving a narrow continuous track on the strip. The cellulose acetate strip used in the Tachofrac (the fraction collector of the

368

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

Fig. 17.4. Tachofrac fraction collector unit, designed for the Tachophor, mounted on the capillary plate. 1 = leading electrolyte reservoir, 2 = Tachofrac unit, 3 = cellulose acetate strip, 4 = UV detector case, 5 = guide wheel, 6 = lifting arm with rubber stopper, 7 = T-connection unit, 8 = slit for UV light, 9 = mechanics simultaneously punching the strip edge and giving electronic impulse to recorder, 10 = strip puller, 1 1 = separation capillary column, 12 = screw for setting the distance between strip surface and T-connection outlet.

Tachophor) is 5 mm wide and 0.1 1 mm thick. Fig. 17.4 shows the fraction collector mounted on a capillary plate in the Tachophor. The edge of the strip is automatically marked at 5 0 m m intervals by a marker positioned very close to the T-connection outlet. At the same time a signal is fed to the U V recorder, resulting in a spike on the chart and permitting positive identification between positions in the W pattern and on the strip. The time delay between the detection at the UV slit and the outlet at the T-connection is constant for a given electrolyte system and running current. The displacement between the strip and chart needs be established only once for such a system, by making a calibration run with a coloured sample. When the fraction collector is not activated, the T-connection outlet is automatically closed by a silicone-rubber cover to prevent clogging of the narrow hole. Owing to the small dimensions of the T-connection, filtration of the leading electrolyte solution through a non-fibrous filter is recommended. Resolution of fraction collection procedure One of the major advantages of capillary isotachophoresis is the high resolution obtained in the liquid phase. Therefore, if the preparative procedure is to add to the value

369

PRINCIPLE OF PREPARATIVE CAPILLARY METHOD

a

b time

uv ebs

ow

Fig. 17.5. UV recordings of dye materials at 278 and 254 nm compared with soft laser scanning at 635 nm of strips with the collected samples, showing the maintenance of high resolution during the preparative process. Samples: different mixtures of tetrasulphonated indigo and xylenol orange. The first blue zones behind the leading ion are shown. Upper traces: UV recordings; lower traces: laser scan results. The time bases differ by approximately 15%, and intensities are not matched. The spikes m denote the event marker pulses, simultaneously causing a print on the edge of the sample collection strip. In (a), the electrolytes contained spacing impurities other than those in (b). Leading electrolyte: (a) 5 mM HCI, 13 mM p-alanine with 0.25% HPMC (Methocel 90 HG hydroxypropylmethylceHulose, a non-ionic additive increasing resolution by decreasing electroendosmosis, 15,000 cP, from Dow Chemical Co.); (b) 10 mM HC1, 24 mM p-alanine and 0.1% HPMC, all at pH 3.6. Terminating electrolyte: 10 mM caproic acid in both. Current: (a) 80pA during the separation,decreased to 30pA during detection and sample collection; (b) 120 and 70 PA, respectively. Temperature, 20°C; chart speed, 3 cm niin-' ; capillary length, 4 3 cm. The slit for the laser beam of the scanner was set at about 0.5 X 0.5 mm.

of capillary isotachophoresis, it must ensure that the resolution is maintained. By strictly controlling the liquid flows in the whole system, the only remaining uncertainties are the loop mounting, the dead volumes and the flushaut process in the separation capillary at the T-connection. The T-connection unit is mounted on the capillary in the middle of a loop of the capillary behind the UV detector (Fig. 17.3). The possible influence of the loop on the resolution has been studied by comparative runs with and without loops of various diameters placed immediately in front of the W detector. With the loop radius used in the Tachophor, n o significant effect has been found. The dead volume of the T-connection is extremely small (2 nl) and does not influence the resolution. The counter flow is generally 2-3pl/min, maintained by an accurate microsyringe pump. Further, the smallest detectable zone length is of the order of 0.1 mrn, occupying 1 6 n l in the separation capillary [22]. It is very difficult to study the elution How at the T-connection but, by comparing the resolution on the strip with that of the UV detector, an indirect answer can be obtained. A series of separations of dyes have been made [23] in w h c h the strips produced were scanned with a laser densitorneter. The wavelength of this unit did not coincide with that of the W detector, but this would not influence the resolution of the resulting patterns.

370

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

Fig. 17.5 indicates strongly that the resolution is maintained, and no other experiments so far reported have given rise to any doubts. Desorption of sample from the strip When collecting samples on a solid support there is a possibility of irreversible adsorption of the sample, especially proteins. Cellulose acetate is a well known and accepted material within the protein field. Two series of experiments for studying the adsorption behaviour have been reported [23]. In the first, electrophoretic elution of transferrin from the strip was applied, followed by immunological quantitation. The second series utilized simply repeated soaking with leading electrolyte buffer, applied to a sample of cholinesterase on the strip, followed by zymogram detection. In both sets of experiments complete elution, within experimental error, was found. An efficient means of eluting the sample from the strip is to centrifuge the strip in a small tube at 6000-8000g, placing a drop of elution liquid on top of the strip. Alternatively, one could chop the strip, centrifuge the eluate out through a small hole in an insert or space the eluate from the strip pieces with an inert, porous material.

APPLICATIONS OF PREPARATIVE CAPILLARY METHOD The purpose of applying preparative capillary isotachophoresis to a sample could be as a late or final step in a purification process, t o obtain the final substance as pure as possible. An advantage of isotachophoresis is that each zone contains only the ion in question and the corresponding amount of the chosen counter ion. The most probable step after sample collection is elution of the cellulose acetate strip. The first application described below is an example in which the eluate was subject t o biological activity. Another possible purpose of the preparative procedure is to use further detection or identification methods without eluting the sample from the strip. There are, of course, a number of possible methods. In the'following part of this chapter, some experiments will be summarized in which detection and identification were effected by radioactivity measurement, immunoelectrophoresis and zymogram techniques. it should be pointed out, however, that the micro scale involved in capillary isotachophoresis places some restrictions on the choice of methods. The volume of liquid collected on the strip is very small, 0.5 pl/cm being an average value. This corresponds to amounts mostly in the range of 20-50nmole/cm collected in a narrow track on the strip, approximately 1 mm wide. The zone length in the separation column is extended approximately 3-fold when transferred to the strip. This elongation simplifies further treatment of a sample zone from the strip. Moreover, the spacer ion technique [19] can be utilized to remove neighbouring impurity zones and will further increase the possibilities of recovering highly purified substances quantitatively. There are many detection methods available for micro-scale application. There are also ways of increasing the amounts of sample, e.g., by making several identical runs or by increasing the concentration of the leading ion, which causes a proportional increase in the concentrations of the sample ions in their zones. The maximal amounts of sample

APPLICATIONS OF PREPARATIVE CAPILLARY METHOD

37 1

that can be separated depend on the mobility differences between the sample components. During studies of impure enzyme preparations a few hundred micrograms were injected in each experiment and separated successfully [24]. Biological activity assay

A series of purification steps of an extract of peptide hormones from porcine upper intestines results in a methanol-soluble fraction of thermostable peptides [ 2 5 ] .The content of secretin (a gastrointestinal hormone) in this material is about 1-3% by bioassay. Secretin is a heptacosapeptide amide with 3500 units per milligram. The usual purification procedure, which starts with several hundred kilograms of intestines, is lengthy and substantial losses are incurred both before and after obtaining the methanol-soluble fraction. The purpose of the study of the Tachophor-Tachofrac method was to find a micro-scale hgh-yield preparation step that would result, if successful, in a material both biologically active and sufficiently pure for amino acid sequence determination. This might be of interest for the isolation not only of secretin from different species, but also of other compounds with the use of smaller amounts of starting material. Samples of the extract fraction containing approximately 5Opg of peptide material in 7.5-lop1 aqueous solution were injected into the Tachophor and the fractions were

a

A

r

b

uv

abr

B

-

.

. .. . ...

tima

Fig. 17.6. Separations of porcine thermostable intestinal peptides run as positively charged ions (courtesy of M. Carlquist, Karolinska Institutet, Stockholm, unpublished results). (A) 7.5-pl sample containing approximately 50 pg of the methanol-soluble fraction (50-75 units/mg); (B)separation of less sample doped with pure secretin. After sample collection the sections a, b and s were tested for biological activity, which was found only in the s sections. For further details, see text. Leading electrolyte: 5 mM potassium acetate (pH 5.1),0.25% HPMC; terminating electrolyte: 5 M e a l a n i n e . Current, 50pA; capillary length, 43cm; detection at 254 nm; temperature, 13°C; chart speed, 5 cm/min, analysis time, approximately 15 min.

372

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

nnn n

EEE opa B

Q

\I

.

3t RCC

1 mln

Fig. 17.7. Capillary isotachophoretic separation of Y-labelled 5’-AMP (88 pmole) and 5‘-UMP (5 nmole) together with unlabelled 3’-AMP (0.77 nrnole), S’GMP and 5’CMP (both 2.5 nrnole). The upper line (RCC) shows the activity as detected by the RadioChromatogram Camera. The lower trace (LSC) shows liquid scintillation counting on 2.5-mm pieces of the strip. Leading electrolyte: 5 mM HCI, 9 mMp-alanine, 0.25% HPMC a t pH 3.4. Terminating electrolyte: l O m M caproic acid. Capillary length, 43 cm; current, 80 HA during separation and 30 pA during detection and sample collection; total time of experiment, 24 min; wavelength of recording, 278 nrn; temperature, 20°C.

collected on the cellulose acetate strip [26]. The W trace at a typical analysis is shown in Fig. 17.6A. After addition of pure secretin (2-3 times the amount believed to be present in the original sample), the trace shown in Fig. 17.6B was obtained (the total amount of sample injected was smaller than in A). The strip sections corresponding to the fractions in B were eluted and injected intervenously into an anaesthetized cat fitted with a pancreatic fistula for activity assay. The elution was made by shaking the strip, cut into pieces, with 0.5 ml of physiological saline and withdrawal of the solution with a syringe. The recovery of activity was approximately 70% of the estimated value. This rough elution procedure was obviously not quantitative, but nevertheless indicates the preservation of activity during the isotachophoresis steps. The fractions denoted a, s and b in Fig. 17.6A were likewise eluted and assayed, fraction s giving 2 units of activity, whereas the expected value was 2-4 units. In order to obtain a quantitatively accurate assay, not

APPLICATIONS OF PREPARATIVE CAPILLARY METHOD

373

Fig. 17.8. Capillary isotachophoretic and immunological analysis of a transferrin sample, purified by column isotachophoresis. The sample was 3 91 of a transferrin solution of unknown concentration with 4 pl of 0.8%Ampholine, pH 5-7. The part of the cellulose acetate strip corresponding t o the line above the IJV trace was cut out and placed on top of the antibody-free part of the agarose gel. The final precipitin pattern (total anti-human serum) is shown in the lower part of the figure. The middle trace is identical with the precipitin line but with its time axis adjusted t o that of the UV trace. Leading electrolyte: 5 mM tIC1, 6 mM Tris, 0.5% HPMC, pH 7.35. Terminating electrolyte: 10 ti# glycine, Ra(OtI), added t o pH 9.2. Capillary length, 4 3 cm; current, 90pA during separation and 45 pA during detection and sample collection; total time of experiment, 22 min; wavelength of recording, 254 nm; temperature, 20°C; chart speed, 3 cm/min.

less than 0.1-0.3 unit should be injected into the test animal. Fractions a and b showed no activity. Radioactivity measuEment Fig. 17.7 shows the separation of the 5’-monophosphates of uridine, cytidine, guanosine and adenosine together with adenosine-3‘-monophosphate, the 5’-AMP and 5’-UMP being labelled with I4C. The complete cellulose acetate strips were studied after sample collection from two identical separations 1231. One of the strips was screened for activity with an LKB 2 105 Radiochromatogram Camera, which gives mainly qualitative information. The other strip was cut by hand into 2.5-mm pieces, which were placed in vials and their activities measured with an LKB 81000 liquid scintillation counter. Thc results given by the camera were in good agreement with both the UV results and the counter results. The amount of 5’4JMP activity was 120 nCi in 5 nmole, whereas the 88 pmole

374

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

Fig. 17.9. Isotachophoretic and immunological analysis of human serum proteins. Sample: 1 . 0 ~ 1 serum and 0.5 HI of 1%Arnpholine, pH 6-8. The part of the cellulose acetate strip corresponding to the UV record (A) was cut out and applied to the agarose gel containing antibodies against whole human serum (B). The different proteins marked were identified by the doping technique or by running against monospecific antisera. Leading electrolyte: 5 mM MES [ 2-N-(morpholino)ethane 0.5% HPMC, pH 9.0. Termisulphonic acid]; 10 mM ammediol(2amino-2-methyl-l,3-propanediol); nating electrolyte: 5 mM EACA (caminocaproic acid) adjusted with Ba(OH), to pH 10.6. Analysis time, 40 min; current during detection, 50pA; chart speed, 3 cm/min; speed of the sample collection strip, 2.6 cm/min; capillary length, 43 cm; monitoring wavelength, 280 nm.

of 5‘-AMP contained 50 nCi of activity. The unevenness of the liquid scintillation counting zone of 5’-UMP was due to errors in the hand cutting of the strip. The results show clearly that the resolution of the activity measurements is comparable t o that of W detection. Immunoelectrophoretic identification The sample collector strip can be used directly for immunological characterization of the sample by means of crossed immunoelectrophoresis [27]. The strip can be placed directly on the antibody-containing agarose gel. Experiments have been reported [23] on transferrin and ceruloplasmin samples. The transferring sample shown in Fig. 17.8 had been purified by gel column isotachophoresis, and showed a number of fractions when Ampholine was used as a spacer gradient [28,29]. The collector strip corresponding to the line in the UV trace was tested immunologically against rabbit total anti-human serum and the precipitin line in the bottom of Fig. 17.8 was obtained. The middle trace is identical, but with the time scale adjusted to that of the W trace. It is obvious that all of the fractions gave a transferrin response, and n o other precipitin line could be detected on the immuno plate. These fractions could be due to minor modifications of the molecules without immunological importance. Previous studies [28,29] have shown that artefacts due t o the presence of Ampholine are unlikely. The resolutions of the immuno and UV traces are similar; the small decrease in the

APPLICATIONS OF PREPARATIVE CAPILLARY METHOD

375

1 r’ Y

d

Fig. 17.10. Separation of a commercial chotinesterase preparation and zymogram detection of activity on pre-treated sample collection strips. The axes are defied in (b), where REA is the relative enzyme activity, estimated from the zymogram colour intensities on a scale of 0-5. (a) No sample, peak pattern is from electrolyte impurities; (b) 20pg of chotinesterase; (c) I p1 of 0.8% Ampholine, pH 4-6; (d) 2Opg of cholinesterase and Ampholine as in (c); (e) 35 p g of cholinesterase and Ampholine as in (c). The nonenzymatic protein part of the 35-pg sample was too large to attain isotachophoretic equilibrium completely, which is the reason for the difference in the shapes of the zones directly after the leading ion. Leading electrolyte: 5 mM HCI, 6 mM Tris with 0.5% HPMC, pH 7.35. Terminating electrolyte: 10 mMglycine, Ba(OH), added to pH 9.2. Capillary length, 43 cm; current, 45pA; thermostate, 20°C; detection at 254 nm; separation time until the first peaks were detected, about 20 min.

immuno pattern can be ascribed to diffusion effects resulting from the 20 h of immunoelectrophoresis. When a commercial ceruloplasmin preparation was added to the transferrin sample, at least six immunologically different components were obtained, which migrated in the region between the impurities and the first transferrin fractions.

376

PREPARATIVE ISOTACHOPHORESIS ON THE MICRO SCALE

In another study [30], microgram amounts of human serum proteins were separated isotachophoretically by using different Ampholine spacers. The strips were analysed in the same way by immunoelectrophoresis. Agarose gel plates as large as 24 x 9 cm were used for the characterization of complete samples. The results from one of the separations are shown in Fig. 17.9, where 1 pl serum was injected together with 0.5 of 1% Ampholine, pH 6-8, The proteins listed were identified by doping with pure proteins, but unequivocal identification was achieved b y running against monospecific antisera in the intermediate gels. During the study it was also found that the collected sample strips could be kept frozen at - 18°C before immunoelectrophoresis for at least 2 weeks (longer times were not studied) with no deterioration of the sample. Other successful separations have been made in the presence of antioxidants such as mercaptoethanol and dithiothreitol in the electrolytes, and also reagents such as urea and Triton X-100 1241. Enzyme identification by zymogram One possibility of post-detector derivatization directly on the strip has been studied by zymogram identification of a commercial cholinesterase [ 2 3 ] .Before the sample was collected, the strip was soaked in a reagent solution producing a reddish brown diazo compound when in contact with the enzyme. The collection was made on the dried strip, which afterwards was soaked twice again in the reagent solution with drying in between, to reinforce the intensity of the zymogram colour. The colour intensities on the strips were estimated visually on a 5-grade scale and the results are shown in Fig. 17.10. Ampholine spacer gradients were used to clarify wluch of the many protein fractions had the proper enzyme activity. The possibility of purifying the cholinesterase sample and obtaining only highly active enzyme is apparent.

REFERENCES 1 €I. Haglund, Sci. Tools, 1 7 (1970) 2 . 2 F. M. Everaerts, J. L. Beckers and Th. P. E. M. Verheggen, Zsotachophoresis - Theory, Instrumentation and Applications, Journal of Chromatography Library, Vol. 6, Elsevier, Amsterdam, Oxford, New York, 1976. 3 P. J. Svendsen and C. Rose, Sci. Tools, 17 (1970) 13. 4 P. J, Svendsen in 2. Deyl (Editor), Electrophoresis. A Survey of Techniques and Applications. Part A: Techniques, Elsevier, Amsterdam, Oxford, New York, 1979, Ch. 4. 5 A. Vestermark, Cons Electrophoresis. an Experimental Study, University of Stockholm, Stockholm, 1966. 6 A. Vesterrnark and B. Sjodin, J. Chromatog,, 71 (1972) 588. 7 A. Vesterrnark and B. Sj0din.J. Chromatog., 73 (1972) 211. 8 B. Sjodin and A. Vesterrnark, Biochim. Biophys. Acta, 297 (1973) 165. 9 G . Eriksson,Acta Chem. &and., 21 (1967) 2290. 10 K. Uyttendaele, M. De Groote, V. Blaton, F. Alexander and €1. Peeters, Protides Biol. Fluids, 22 (1975) 743. 11 K. Uyttendaele, V. Blaton, F. Alexander, H. Peeters, M. De Groote, N. Vinaimont-Vandecasteele and J. Chevalier, in P. G. Righetti (Editor), Progress in Isoelectric Focusing and Isotachophoresis, North-Holland, Amsterdam, 1975, p. 341.

REFERENCES

377

12 C. H. Brogren, in B. J. Radola and D. Graesslin (Editors), EIectrofocusing and Isotachophoresis, Walter de Gruyter, Berlin, 1977, p. 549. 13 C. H. Brogren and G. Peltre, in B. J. Radola and D. Graesslin (Editors), Electrofocusing and Isotachophoresis, Walter de Gruyter, Berlin, 1977, p. 588. 14 B. J. Radola, Biochim. Biophys. Acta, 386 (1974) 181. 15 A. Chrambach, G. Kapadia and M. Cantz, Separ. Sci., 7 (1972) 785. 16 S . HjertCn, Chromatog. Rev., 9 (1967) 122, p. 199. 17 S . Hjertbn, Protides Biol. Fluids, 22 (1975) 669. 18 A. Kopwillem,J. Chromatogr., 82 (1973) 407. 19 A. Kopwillem, W. G. Merriman, R. M. Cuddeback, A. J. K. Smoka and M. Bier,J. Chromafog,, 118 (1976) 35. 20 W. Preetz, U. Wannemacher and S . Datta, 2 Anal. Chem., 257 (1971) 97. 21 J . S . Fawcett, Ann. N.Y. Acad. Sci., 209 (1973) 112. 22 L. Arlinger,J. Chromatogr., 91 (1974) 785. 23 L. Ar1inger.J. Chromatog., 119 (1976) 9. 24 A. Baldesten and S . C . Hjalmarsson, personal communication. 25 J . E. Jorpes and V. Mutt (Editors), Secretin, Cholecystokinin, Pancreozymin and Gastrin, Handbook Experimental Pharmacology X X X I V , Springer, Berlin, 1973. 26 M. Carlquist, personal communication. 27 N. H. Axelsen, J. Kr$ll and R. Weeke, A Manual of Quantitative Immunoelectrophoresis, Universitetsforlaget, Oslo, 1973. 28 L. Arlinger, Biochim. Eiophys. Acta, 393 (1975) 396. 29 L. Arlinger, in P. G. Righetti (Editor), Progress in Isoelectric Focusingand Isotachophoresis, North-Holland, Amsterdam, 1975, p. 331. 30 U. Moberg, S.C. Hjalmarsson, L. Arlinger and H. Lundin, in B. J . Radola and D. Graesslin (Editors), Electrofocusingand Isotachophoresis, Walter de Gruyter, Berlin, 1977, p. 515.

This Page Intentionally Left Blank

List of frequently occurring symbols

A A0

electrochemical equivalent (kg C-' ) internal diameter of injector tube, or initial thickness of injected streak (m) ionic species coefficient of interaction zone half-width (m) constants magnetic field intensity (T) common counter ion constrictive factor concentration of all solutes (mole concentration of ionic species A, in zone u (mole m-3) concentration of substance i (mole m-3) concentration of component i, z (mole m-3) molar concentration (mole d n ~ - ~ ) specific heat capacity o f the solution (J kg-' K-') diffusion coefficient (inz s-') relative correction factor thermal diffusion coefficient of component i, z (m K-' s-') wall thickness of the separation column, internal diameter of injector tube or initial thickness o f injected streak (m) electric field strength (V ni-') electric field strength at the location where the component concentration is c (V m-') diffusion gradient of the potential (V * m-') electric field strength in the leading electrolyte (V * ni-') ohmic gradient of the potential (V * m-') thermal gradient of the potential (V m-') electric field strength in zone u (V * m-') elemental charge (C) Faraday's constant (C mole-') signal intensity (height) of the photometric recorder (m) relative signal intensity (height) of the photonietric recorder vertical force density (force per unit volume) (N * m-3) horizontal force density (force per unit volume) (N constant (including Faraday's constant) thickness of fluid ribbon (m)

-

-

-

.

-

- -

-

-

-

-

-

LIST OF FREQUENTLY OCCURRING SYMBOLS

380

“zone height” of the ith substance on the isotachopherogram (m) “relative zone height” of the ith substance “zone height” of the leading electrolyte (m) height of the standard ion species zone (m) electric current (A) ionic strength of the solution (mole m-3) light intensity (W) unity vector substance flow of component i, z (mole m-’ * s-I) diffusion contribution of the substance flow of component i, z (mole m-’ s-I) migrational contribution of the substance flow of component i , z (mole m-’ s-l) convective contribution of the substance flow of component i, z (mole rn” * s-’) thermal contribution of the substance flow of component i, z (mole m-’ * s-l) unity vector current density (A m-2) apparent equilibrium constant convective equilibrium constant (mole drn-’) convective equilibrium constant (mole dm-’) dissociation constant of an acid (mole dm-3) dissociation constant of an acid characterizing first-step dissociation (mole dm-j) dissociation constant characterizing dissociation from the first to the second step (mole drn-’) dissociation constant of a base (mole * dm-3) retardation coefficient ionic product of water (mole’ dm-3) dissociation constants of an ampholyte (mole dm-3) unity vector Boltzmann’s constant thermal conductivity of the solution (W m-’ * K-’) thermal conductivity of the solid phase (W * m-’ K-’ ) const ants length of the separation column, length of the porous bed (m) length of the ith zone in an isotachopherogram (m) length of the mixed zone on the start (m) distance between the point of application and centre of a gel plate (in rheophoresis) (m) actual pore length, actual distance migrated (m) molar mass (kg * mole-’) relative molar mass mass (kg)

-

-

-

i

f

K K( A-’) K(X’-‘) KA KA 1 KAZ

-

-

-

1

M M m

-

38 1

LIST OF FREQUENTLY OCCURRING SYMBOLS

m* NA

nf tli,r

P P* P

mass of the liquid evaporated from unit surface of a porous medium during unit time (kg m-* s-l) Avogadro’s number (mole-’) fraction number amount of substance of component i, z (mole) surface area (m’) constant electric output (W)

-

-

isoelectric point (sometimes designated as pH&) PI pK = -1ogK pK, = - log K , where K , is the dissociation constant of the more acidic of the two ionizable groups pK- = -log K - , where K- is the dissociation constant of the less acidic of the two ionizable groups Q total charge (C) qj Joule’s heat (J) qk heat removed (J) change of thermal flow in unit volume Y* thermal flow generated by unit volume (W * m-3) 4; thermal flow leaving unit volume (W m-3) q: R molar gas constant (J * K-’ mole-’)

-

see (ui,s )re1 resolving power, resolution radius of a spherical particle or Stokes radius (m) outer radius of the cylindrical annulus (m) inner radius of the cylindrical annulus (m) internal radius of the separation column (m) exclusion radius (m) cross-section of a column or cross-section of a porous medium (m’) total cross-section of a porous medium (m’) migration distance of an unretained solute in a gel (in rheophoresis)

(m) migration distance of a retained solute in a gel (in rheophoresis) (m) rheophoretic contribution of migration distance (m) endoosmotic contribution of migration distance (m) electrophoretic contribution of migration distance (m) free cross-section of all capillaries (in the perpendicular direction t o their longitudinal axis) (m’) path travelled by an idealized particle in the direction of electrophoretic migration (m)

382

LIST O F FREQUENTLY OCCURRING SYMBOLS

Sm

,si

cross-sectional area of electrolytic medium, measured perpendicular to the direction of the current (m2) resulting migration distance of migrant 1,2, . . . ,i (m) absolute temperature (K) temperature of the external surface of the separation column (K) thermal flow (We m-*) time (s) time needed for the detection of the leading electrolyte (s) time needed for the detection of zone u (s) time needed for the detection of the first sample zone (s) total imposed voltage (V) mobility of ionic species A in the leading electrolyte (m' s-' V-') mobility of ionic species A, (m' s-* V-') mobility of ionic species A in zone u (m2 s - ' * V-') mobility of ionic species B (m' * s - ' * V-') actual mobility of a colloid particle (m' s- l - V-') effective mobility of the ith substance (m' s-' V-') macroscopic mobility of the ith substance (m2 s-' ' V-') mobility at zero gel concentration (m2 s - l - V-'1 relative mobility of the ith substance with respect to the sth standard (sometimes designed as R F ) actual mobility of the i , zth component (m2 s-' V-') absolute (limit) mobility of the i , zth component (m' s-'- V-') actual mobility of the leading electrolyte (m2 s-' V-') electroosmotic mobility (m2 s-' V-') actual mobility of zone u (in isotachophoresis) (m' s-'. V-') velocity of a particle due to electroosmosis and electrophoresis (m * s-') volume (m') injected volume (m3) volume of the separation compartment (m3) velocity vector (m s-*) horizontal velocity of a charged particle (m s-') velocity of the ith substance at concentration ci fluid curtain velocity (m s-') velocity of the leading zone in isotachophoresis (m s - l ) electrophoretic velocity (m s-') central maximum velocity (m s-') linear velocity of the osmotic flow (m 0 s - l ) maximum linear velocity of the osmotic flow (at the wall) (m s-') chart speed (m s-l) velocity of a retained molecule (m s-') velocity of the suction flow (m * s-') velocity of zone u (m s-l) relative velocity of zone u

-

-

-

-

-

-

-

-

0

VOS

Vr

%uck + VU

+r

V

-

-

383

LIST OF FREQUENTLY OCCURRING SYMBOLS

velocity of an unretained molecule (m s-') volume rate of flow of electrolytes through the chamber (m3 s-') (volume) velocity of the osmotic flow (m3 s-') macroscopically determined positional coordinate (m) ion of a weak electrolyte X, carrying a charge z molar concentration of the Xz component (mole dm-') peak position of molecular species A, B, C (m) position coordinate, distance between the point of injection and the point of detection (m) macroscopically determined position coordinate (m) molar fraction of component i, z in the ith substance acrylamide bisacrylamide concentration (g per 100 ml) macroscopically determined position coordinate (m) number of elementary charges of the component; the sign of z determines the polarity of the ion polarity determination: sgn z = 1 for z > 0 sgn z = 0 for z = 0 sgnz=-1 forz