Immobilized Enzymes and Cells - PDF Free Download

Contributors to V o l u m e 136 Article numbersare in parenthesesfollowingthe names of contributors. Affiliationslisted ...

Author: Nathan P. Colowick | Nathan P. Kaplan | Klaus Mosbach

23 downloads 2225 Views 35MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Contributors to V o l u m e 136 Article numbersare in parenthesesfollowingthe names of contributors. Affiliationslisted are current. OBSIDIANA ABRIL (25), Research and De-

PETER S. J. CHEETHAM (40), PPF Internavelopment, Angenics, Inc., Cambridge, tional, Ashford, Kent TN24 OLT, England Massachusetts 02139 ICHmO CHIBATA (41, 43), Research andDeL. ANDERSSON (I 1), Ferring AB, S-200 62 velopment Headquarters, Tanabe SeiyMalm6, Sweden aku Co., Ltd., Yodogawa-ku, Osaka 532, Japan RICHARD ARMENTA (9), Molecular Devices Corporation, Palo Alto, California 94304 DEBBIE C. CRANS (25), Department of Chemistry, Colorado State University, D. AURIOL (23), BioEurope, F-31400 TouFort Collins, Colorado 80523 louse, France PIERO CREMONESl (14), ltalfarmaco, MASAKI AZUMA (36), Technical Research S.p.A., Milan, Italy Laboratories, Kyowa Hakko Kogyo Co., Ltd., Hofu Plant, Hofu-shi, Yamagachi M. J. DANIELS (34), Genencor, Inc., South Prefecture 747, Japan San Francisco, California 94080 JOHANN BADER (28), Deutsches Patentamt, M. DELUCA (8), Department of Chemistry, D-8000 Miinchen 2, Federal Republic of University of California, San Diego, La Germany Jolla, California 92093 J. L. BARET (38), Novo lndustrie Enzymes ROBERT K. DINELLO (9), Chiron CorporaS.A., F-75017 Paris, France tion, Emeryville, California 94608 L. A. BEHIE (30), Department of Chemical G. FLEMINGER (17), Department of Biotechand Petroleum Engineering, The Univernology, The George S. Wise Faculty of sity of Calgary, Calgary, Alberta, Canada Life Sciences, TeI-Aviv University, RaT2N 1N4 mat-Aviv, 69 978 TeI-Aviv, Israel LARRY G. BUTLER (22), Department of BioSABURO FUKUI (27), Department of Induschemistry, Purdue University, West Latrial Chemistry, Faculty of Engineering, fayette, Indiana 47907 Kyoto University, Yoshida, Sakyo-ku, Kyoto 606, Japan G. J. CALTON (45), RhOne.Poulenc Research Center, Savage, Maryland 20763 MURRAY C. FUSEE (42), Research Division, W. R. Grace & Co., Washington ReBERNARD CAMBOU (12), Laboratory of Apsearch Center, Columbia, Maryland plied Biochemistry, Department of Ap21044 plied Biological Sciences, Massachusetts Institute of Technology, Cambridge, G. M. GAUCHER (30), Division of BiochemMassachusetts 02139 istry, Department of Biological Sciences, Faculty of Science, The University of CalM. L. CAMPBELL (45), RhOne.Poulenc Regary, Calgary, Alberta, Canada T2N 1N4 search Center, Savage, Maryland 20763 S. GESTRELIUS (32), Ferring AB, S-200 62 GIACOMO CARREA (14), Istituto di Chimica Malm6, Sweden degli Ormoni, Consiglio Nazionale delle Ricerche (C.N.R.), 20131 Milan, Italy IAN GIBBONS (9), Biotrack, Inc., Sunnyvale, California 94086 Z. M. S. CHANG (7), Artificial Cells and Organs Research Centre, Faculty of Medi- HELMUT G/3NTHER (28), Lehrstuhl fiir Orcine, McGill University, Montreal, Queganische Chemie und Biochemie im Orbec, Canada H3G 1 Y6 ganisch-Chemischen lnstitut der Technisix

X

CONTRIBUTORS TO VOLUME 136

chen Universitgit Miinchen, D-8046 REINHOLD KELLER (47), Hoechst AktiengeGarching, Federal Republic of Germany sellschaft, D-6230 Frankfurt am Main 80, Federal Republic of Germany NORIO HAGI (46), Biotechnology Research Center, Toyo Soda Manufacturing Co., ALEXANDER M. KLISANOV (12), LaboraLtd., Yamaguchi-ken 746, Japan tory of Applied Biochemistry, Department of Applied Biological Sciences, MassaS. HARBRON (13), Department of Biochemchusetts Institute of Technology, Camistry, University College London, London bridge, Massachusetts 02139 WCIE 7JE, England ANDREAS KONNECKE (18), Department of URSULA HAUFLER (26), Fachbereich BioloBiochemistry, Biosciences Division, Karlgic, Universitgit Bremen, D-2800 Bremen Marx-University, DDR-7010 Leipzig, 33, Federal Republic of Germany German Democratic Republic LARS HEDSYS (21), Pure and Applied BioY. KONNO (49), Dowa Engineering Co., chemistry, Chemical Center, University Ltd., Tokyo 100, Japan of Lund, S-221 O0 Lund, Sweden RIET HILHORST (20), Department of Bio- BETTINA Koep (29), Henkel KGaA, D-4000 Diisseldorf 1, Federal Republic of Gerchemistry, Agricultural University Wamany geningen, 6703 BC Wageningen, The FRIEDRICH KRAUSS (28), Lehrstuhlfar OrNetherlands ganische Chemie und Biochemie im OrBERNARD L. HIRSCHBEIN (25), Imaging Reganisch-Chemischen lnstitut der Technissearch, Polaroid Corporation, Camchen Universitdt Miinchen, D-8046 bridge, Massachusetts 02139 Garching, Federal Republic of Germany T. IMAIZUMI (49), The University of Tokyo, L. J. KRICKA (8), Department of Clinical Shibuya-ku, Tokyo 113, Japan Chemistry, Wolfson Research LaboratoKEIICHI INUZUKA (36), Kyowa Hakko ries, University of Birmingham, BirmingKogyo Co., Ltd., Ube Plant, Ube-shi, ham B15 2TH, England Yamaguchi Prefecture 755, Japan MARIA-REGINA KULA (2), lnstitut far EnSHIGEAKI IRINO (46), Chemical Research zymtechnologie der Universitgit DiisselCenter, Toyo Soda Manufacturing Co., dorf in der Kernforschungsanlage Jiilich Ltd., Yamaguchi-ken 746, Japan GmbH, D-5170Jiilich 1, Federal Republic HANs-DIETER JAKUBKE (18), Department of of Germany Biochemistry, Biosciences Division, KarlCOLJA LAANE (20), Unilever Research LaMarx-University, DDR-7010 Leipzig, boratorium, 3130 AC Vlaardingen, The German Democratic Republic Netherlands VIOLETA G. JANOLINO (39), Department of Food Science, North Carolina State Uni- KEITH J. LAIDLER (6), Department of Chemistry, Faculty of Science, University versity at Raleigh, Raleigh, North Caroof Ottawa, Ottawa, Ontario, Canada lina 27695 KIN 6N5 VILLY J. JENSEN (33), Molecular Biology, PER-OLOE LARSSON (21), Pure and Applied Novo Industri A/S, DK-2880 Bagsvaerd, Biochemistry, Chemical Center, UniverDenmark sity of Lund, S-221 O0 Lund, Sweden VOLKER KASCHE (26), Arbeitsbereich M. D. LILLY (13), Department of Chemical Biotechnologie H, Biotransformation and Biochemical Engineering, University und -Sensorik, Technische Universitiit College London, London WC1E 7JE, EnHamburg-Harburg, D-2100 Hamburg 90, gland Federal Republic of Germany A. LOPEZ (23), Departamento de Alimentos ROMAS J. KAZLAUSKAS (25), Research and DEPg, Facultad de Quimica, Mexico Development Center, General Electric D.F., Mexico Company, Schenectady, New York 12301

CONTRIBUTORS TO VOLUME 136

xi

P. L. LuIsI (19), lnstitut fiir Polymere,

SHOGO NOJIMA (35), Catalysts and Chemi-

Swiss Federal Institute of Technology, EidgenOssische Technische Hochschule (ETH)-Zentrum, CH-8092 Zurich, Switzerland HIDEKATSU MAEDA (3), Fermentation Research Institute, Agency of Industrial Science and Technology, Ministry of International Trade and Industry, Ibaraki-ken 305, Japan MATS-OLLE MXNSSON (1, 10, 11), Pure and Applied Biochemistry, Chemical Center, University of Lund, S-221 O0 Lund, Sweden M. ABDUL MAZID (6), Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2 P. MONSAN (23), BioEurope, F-31400 Toulouse, France KAZUYUKI MORIHAgA (16), Kyoto Research Laboratories, Toho Pharmaceutical Industry Company, Limited, Kyoto Prefecture 617, Japan KLAUS MOSBACH (1, 10, 11, 21, 32), Pure and Applied Biochemistry, Chemical Center, University of Lund, S-221 O0 Lund, Sweden RYONOSUKE MUNEYUKI (16), Research Division, S. T. Company, Limited, Shimoochiai, Shinjuku-ku, Tokyo 161, Japan T. MURAYAMA(49), Oga Operation Office, Akita Oil Storage Co., Ltd., Funagawa, Funagawa-ko, Oga City, Akita Prefecture 010-05, Japan MINORU NAGASHIMA (36), Technical Research Laboratories, Kyowa Hakko Kogyo Co., Ltd., Hofu Plant, Hofu-shi, Yamaguchi Prefecture 747, Japan T. J. NAI~NDgANATHAN (13), Wellcome Biotechnology Ltd., South Eden Park, Beckenham, Kent BR3 3BS, England STEFAN NEUMANN (28), Consortium far Elektrochemische, Industrie GmbH, D8000 Miinchen, Federal Republic of Germany SADAO NOGUCHI (36), Technical Research Laboratories, Kyowa Hakko Kogyo Co., Ltd., Hofu Plant, Hofu-shi, Yamaguchi Prefecture 747, Japan

cals, Industry Co., Ltd., Chiyoda-ku, Tokyo I00, Japan TATSUSHI OKA (16), Shionogi Research Laboratories, Shionogi and Company, Limited, Osaka 553, Japan HIROSUKE OKADA (4), Department ofFermentation Technology, Faculty of Engineering, Osaka University, Osaka 565, Japan KIYOTAKAOYAMA(46), Chemical Research Center, Toyo Soda Manufacturing Co., Ltd., Yamaguchi-ken 746, Japan F. PAUL (23), BioEurope, F-31400 Toulouse, France URSULA PFITZNER (31), Department of Molecular Biology, Massachusetts General Hospital, Boston, Massachusetts 02114 LuTz RIECHMANN(26), Fachbereich Biologie, Universitgit Bremen, D-2800 Bremen 33, Federal Republic of Germany G. P. ROYER (15), Biotechnology Division, Amoco Corporation, Amoco Research Center, Naperville, Illinois 60566 J. DAVID ROZZELL (44), Genetics Institute, Cambridge, Massachusetts 02140 SUSANNERUG8 (33), Process Development, Novo Industri A/S, DK-2880 Bagsvaerd, Denmark T. SAKATA (49), Metal Mining Agency of Japan, Minato-ku, Tokyo 105, Japan HmOTOSHX SAMEJIMA (36), Kyowa Medex Co., Ltd., Chiyoda-ku, Tokyo 100, Japan MERTEN SCHLINGMANN (47), Hoechst Aktiengesellschaft, D-6230 Frankfurt am Main 80, Federal Republic of Germany NILS SIEGBAHN(10), Perstorp Biolytica AB, S-223 70 Lund, Sweden HELMVT SIMON (28), Lehrstuhlfiir Organische Chemie und Biochemie im OrganischChemischen Institut der Technischen Universitiit Miinchen, D-8046 Garching, Federal Republic of Germany PAUL J. SKUDDER (39), Process Development Department, APV International Ltd., West Sussex RHIO 2QB, England B. STEINMANN-HOFMANN(19), Institutfiir Polymere, Swiss Federal Institute of

xii

CONTRIBUTORS TO VOLUME 136

Technology, EidgenOssische Technische CEES VEEGER (20), Department of BiochemHochschule (ETH)-Zentrum, CH-8092 istry, Agricultural University WaZurich, Switzerland geningen, 6703 BC Wageningen, The Netherlands" HIDEO SUZUKI (5), Fermentation Research Institute, Agency of Industrial Science CHRISTIAN WANDREY (2), Institut fiir and Technology, Ministry of International Biotechnologie 2 in der KernforsTrade and Industry, Ibaraki-ken 305, chungsanlage Jiilich GmbH, D-5170 JiiJapan lich 1, Federal Republic of Germany SIGFRID SVENSSON (21), Carbohydrate ICHIRO WATANABE (48), Central Research Chemistry, Chemical Center, University Laboratory, Nitto Chemical Industry Co., of Lund, S-221 O0 Lund, Sweden Ltd., Tsurumi-ku, Yokohama 230, Japan HAROLD E. SWAISGOOD (39), Department GEORGE M. WmTESIDES (25), Department of Food Science, North Carolina State of Chemistry, Harvard University, CamUniversity at Raleigh, Raleigh, North bridge, Massachusetts 02138 Carolina 27695 G. WIENHAUSEN (8), Department of BiolSATORU TAKAMATSU (43), Research Laboogy, University of California, San Diego, ratory of Applied Biochemistry, Tanabe La JoUa, California 92093 Seiyaku Co., Ltd., Yodogawa-ku, Osaka ROLF WOERNLE (47), Hoechst Aktienge532, Japan seUschaft, D-6230 Frankfurt am Main 80, ISAO TAKATA (41), Research Laboratory of Federal Republic of Germany Applied Biochemistry, Tanabe Seiyaku Cm-HuEY WONG (25), Department of Co., Ltd., Yodogawa-ku, Osaka 532, Chemistry, Texas A&M University, ColJapan lege Station, Texas 77843 ATSUO TANAKA (27), Department oflndusL. L. WOOD (45), RhOne.Poulenc Research trial Chemistry, Faculty of Engineering, Center, Savage, Maryland 20763 Kyoto University, Yoshida, Sakyo-ku, TOMIAKI YAMADA (35), Biotechnology ReKyoto 606, Japan search Department, Research and DevelTAKASHI TANAKA (37), Central Research opment Division, JGC Corporation, Laboratories, Ajinomoto Co., Inc., KaMinami-ku, Yokohama 232, Japan wasaki-ku, Kawasaki 210, Japan IORDANIS THANOS (28), Lehrstuhl fiir Or- K o z o YAMAMOTO (41), Research Laboratory of Applied Biochemistry, Tanabe ganische Chemie und Biochemie im OrSeiyaku Co., Ltd., Yodogawa-ku, Osaka ganisch-Chemischen Institut der Technis532, Japan chen Universitiit Miinchen, D-8046 SHIGERU YAMANAKA (37), Central ReGarching, Federal Republic of Germany search Laboratories, Ajinomoto Co., TETSUYA TOSA (41, 43), Research LaboraInc., Kawasaki-ku, Kawasaki 210, Japan tory of Applied Biochemistry, Tanabe Seiyaku Co., Ltd., Yodogawa-ku, Osaka YOSHIMITSU YAMAZAKI (3, 5), Fermenta532, Japan tion Research Institute, Agency of Industrial Science and Technology, Ministry of JOHANNES TRAMPER (24), Department of International T r a d e and Industry, Food Science, Food and Bioengineering lbaraki-ken 305, Japan Group, Agricultural University Wageningen, 6703 BC Wageningen, The A. YARON (17), Department of Biophysics, Netherlands The Weizmann Institute of Science, 76 100 Rehovot, Israel EDWIN F. ULLMAN (9), Syva Company, Palo Alto, California 94304 MEINHART H. ZENK (31), Lehrstuhl PharITARU URAaE (4), Department of Fermentamazeutische Biologie, Universitgit Mantion Technology, Faculty of Engineering, chen, D-8000 Miinchen 2, Federal RepubOsaka University, Osaka 565, Japan lic of Germany

Preface Volumes 135 through 137 of Methods in Enzymology, Immobilized Enzymes and Cells, Parts B through D, include the following sections: (1) Immobilization Techniques for Enzymes; (2) Immobilization Techniques for Cells/Organelles; (3) Application of Immobilized Enzymes/Cells to Fundamental Studies; (4) Multistep Enzyme Systems and Coenzymes; (5) Immobilized Enzymes/Cells in Organic Synthesis; (6) Enzyme Engineering (Enzyme Technology); (7) Analytical Applications with Emphasis on Biosensors; (8) Medical Applications; and (9) Novel Techniques for and Aspects of Immobilized Enzymes and Cells. The first three sections appear in Volume 135, the next three in Volume 136, and the last three in Volume 137. Immobilization techniques for enzymes, Section (1), has already been treated in Volume XLIV of this series. Immobilization techniques for cells/organelles, Section (2), an area which seems to have great potential, especially for the application of immobilized yeast and plant and animal cells, is covered for the first time in these volumes. Sections (3) and (4) have been dealt with previously. Section (5), the use of immobilized enzymes/cells in organic synthesis, has probably not been covered before. It is my firm opinion that in the not too distant future we will see a number of processes employed which are based, in part, on the examples given in this section. Section (6) on industrial uses updates the material presented in Volume XLIV. The examples given are, to the best of my knowledge, in operational use today or, at least, on a pilot plant level. Section (7), analytical applications with emphasis on biosensors, is the subject of a great deal of research at present, and it may very well be that in the not too distant future we will witness a breakthrough, i.e., many applications of a number of such devices. The medical area, covered in Section (8), seems promising, but certainly more research is required to fully exploit any underlying potential. Finally, in Section (9), I have collected a number of contributions that did not seem to fit in any of the other sections, but do address important and novel developments. I would like to note that although major emphasis in these volumes has been placed on immobilization in its strictest sense, preferentially, covalent attachment of enzymes or entrapment of cells, one should not view immobilizedsystems in too limited a manner. In fact, bioreactors confined by ultrafilter membranes or hollow fiber systems belong in this category, and the various systems appear to overlap. Immobilization techniques as applied to affinity chromatography or immunoassays such as ELISA are not included to any extent in these volumes since they have xiii

xiv

PREFACE

been adequately covered in other volumes of this series (e.g., Volumes XXXIV and 104 on affinity techniques). An area that was originally scheduled for inclusion is synzymes or artificial enzymes. These include attempts to create catalysts mimicking enzymes by coupling of functional groups to, for instance, cyclodextrin [e.g., D'Souza et al. (Biochem. Biophys. Res. Commun. 129, 727-732, 1985) and Breslow et al. (J. Am. Chem. Soc. 108, 1969, 1986)], to crown ethers [Cram et al. (J. Am. Chem. Soc. 107, 3645, 1985)], or to solid matrices [Nilsson and Mosbach (J. Solid-Phase Biochem. 4, 271, 1979) and Leonhardt and Mosbach (Reactive Polymers, in press)]. Related to these studies are attempts to create cavities in polymers with substrate-binding properties ]notably by Wulff et al. (e.g., Reactive Polymers 3, 261, 1985; and previous publications by these authors) and Arshady and Mosbach (Makromol. Chem. 182, 687, 1981)]. This exciting area is presently in a rapid state of development, and the methodology involved should soon be made available in a more comprehensive context. Mention should be made of the developments in the utilization of recombinant DNA technology for the immobilization (and affinity purification) of biomolecules. I refer to the reported fusion of "affinity tails" as polyarginine (Smith et al., Gene 32, 321, 1984), of polycysteine [B01ow and Mosbach, Proceedings of the VIII International Conference on Enzyme Engineering, Annals of the New York Academy of Sciences, in press (presented 1985)], or of protein A (Nilsson et al., EMBO J. 4, 1075, 1985) to enzymes facilitating their purification and immobilization. These preparations can be obtained by fusion of the respective groups as "tail" to the NH2 or COOH termini of the enzyme or by site-directed mutagenesis leading to substitution on the enzyme structure. DNA technology can also be usefully employed to create new multienzyme complexes, fusing enzymes acting in sequence to one another (B01ow et al., Bio/Technology 3, 821, 1985) as an alternative to their co-immobilization on supports; similarly, attachment of "tails" allowing reversible coenzyme binding may be accomplished. The same technology has also been used recently in attempts to prepare esterase mimics from the ground up (Biilow and Mosbach, FEBS Lett. 210, 147, 1987). Since this is such a rapidly moving area, I advise the reader, apart from the usual standard books in this area, to read the proceedings of the Enzyme Engineering Conferences 1-8 (Wiley, first conference; Plenum Press, second-sixth conferences; and Annals of the New York Academy of Sciences, seventh and eighth conferences); Biochemical Engineering, Volumes I-III and subsequent volumes; Annals of the New York Academy of Sciences, 1983; the patent book "Enzyme Technology, Recent Advances" (S. Torrey, ed.), Noyes Data Corporation, Park Ridge, New

PREFACE

XV

Jersey, 1983; and Biotechnology Review no. 2. In addition, in the following journals many articles relating to immobilized enzyme and cell research can be found: Biotechnology and Bioengineering (John Wiley & Sons); Trends in Biotechnology (Elsevier, The Netherlands); Bio/Technology (Nature Publishing Co., U.S.); Applied Biochemistry and Biotechnology (The Humana Press, Inc., U.S.); Applied Biochemistry with Special Emphasis on Biotechnology; Biotechnology Letters (Science and Technology Letters, England); Applied Microbiology and Biotechnology (Springer-Verlag, Germany); Enzyme and Microbial Technology (Butterworth Scientific Limited, England); Biosensors (Elsevier Applied Science Publishing Ltd., England). In studies with immobilized systems, sometimes useful, not immediately obvious "by-products" may be obtained. I refer to the finding that immobilized Escherichia coli cells, when kept in media without selection pressure, show improved plasmid stability (de Taxis du PoEt, P., Dhulster, P., Barbotin, J.-N., and Thomas, D., J. Bact. 165, 871, 1986). An additional example would be the improved regeneration of plants using immobilized protoplasts discussed in Section (2). I would like to express the hope that these volumes present an overview of the various areas in which immobilized enzymes and cells are used, act as a stimulus for further research, and provide methodological "know-how." The proper choice of support and/or immobilization technique for a particular application may not always be easily accomplished, but I hope that guidance to do so is found in these volumes. Putting these volumes together has been a time-consuming and, at times, frustrating undertaking. Without the coeditors, Drs. Lars Andersson, Peter Brodelius, Bengt Danielsson, Stina Gestrelius, and Mats-Olle M~nsson, the volumes would not have materialized. Because of the number of coeditors, some heterogeneity in the editing has resulted. Contributors to the various sections are from substantially different disciplines, and again this has contributed to the heterogeneity that can be found. Part of the editing of the three volumes was carried out in Ziirich, where I held a chair in biotechnology at the Swiss Federal Institute of Technology. Without the enormous efforts and skills of the staff of Academic Press, these volumes would never have reached production. I also owe much gratitude to my secretaries, notably lngrid Nilsson, for their highly qualified help. Finally, I would like to thank the contributors for their efforts. These volumes are dedicated to the memory of the late Professors N. O. Kaplan and S. P. Colowick, with whom I had highly fruitful discussions, especially at the beginning of this undertaking.

K L A U S MOSBACH

METHODS IN ENZYMOLOGY EDITORS-IN-CHIEF

Sidney P. Colowick and Nathan O. Kaplan

VOLUME VIII. Complex Carbohydrates

Edited by ELIZABETH F. NEUFELD AND VICTOR GINSBURG VOLUME IX. Carbohydrate Metabolism

Edited by WILLIS A. WOOD VOLUME X. Oxidation and Phosphorylation

Edited by RONALD W. ESTABROOK AND MAYNARD E. PULLMAN VOLUME XI. Enzyme Structure Edited by C. H. W. HIRS VOLUME XlI. Nucleic Acids (Parts A and B)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME XlII. Citric Acid Cycle

Edited by J. M. LOWENSTEIN VOLUME XIV. Lipids

Edited by J. M. LOWENSTEIN VOLUME XV. Steroids and Terpenoids

Edited by RAYMOND B. CLAYTON VOLUME XVI. Fast Reactions

Edited by KENNETH KUSTIN VOLUME XVII. Metabolism of Amino Acids and Amines (Parts A and B)

Edited by HERBERT TABOR AND CELIA WHITE TABOR xix

XX

M E T H O D S IN E N Z Y M O L O G Y

VOLUME XVIII. Vitamins and Coenzymes (Parts A, B, and C)

Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME XIX. Proteolytic Enzymes

Edited by GERTRUDE E. PERLMANN AND LASZLO LORAND VOLUME XX. Nucleic Acids and Protein Synthesis (Part C)

Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN VOLUME XXI. Nucleic Acids (Part D)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME XXII. Enzyme Purification and Related Techniques

Edited by WILLIAM B. JAKOBY VOLUME XXIII. Photosynthesis (Part A)

Edited by ANTHONY SAN PIETRO VOLUME XXIV. Photosynthesis and Nitrogen Fixation (Part B)

Edited by ANTHONY SAN PIETRO VOLUME XXV. Enzyme Structure (Part B) Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XXVI. Enzyme Structure (Part C)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XXVII. Enzyme Structure (Part D)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XXVIII. Complex Carbohydrates (Part B)

Edited by VICTOR GINSBURG VOLUME XXIX. Nucleic Acids and Protein Synthesis (Part E)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME XXX. Nucleic Acids and Protein Synthesis (Part F)

Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN VOLUME XXXI. Biomembranes (Part A)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

METHODS IN ENZYMOLOGY

xxi

VOLUME XXXII. Biomembranes (Part B)

Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME XXXIII. Cumulative Subject Index Volumes I - X X X

Edited by MARTHA G. DENNIS AND EDWARD A. DENNIS VOLUME XXXIV. Affinity Techniques (Enzyme Purification: Part B)

Edited by WILLIAM B. JAKOBY AND MEIR WILCHEK VOLUME XXXV. Lipids (Part B)

Edited by JOHN M. LOWENSTEIN VOLUME XXXVI. Hormone Action (Part A: Steroid Hormones)

Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN VOLUME XXXVII. Hormone Action (Part B: Peptide Hormones) Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN VOLUME XXXVIII. Hormone Action (Part C: Cyclic Nucleotides)

Edited by JOEL G. HARDMAN AND BERT W. O'MALLEY VOLUME XXXIX. Hormone Action (Part D: Isolated Cells, Tissues, and Organ Systems) Edited by JOEL G. HARDMAN AND BERT W. O'MALLEY VOLUME XL. Hormone Action (Part E: Nuclear Structure and Function)

Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN VOLUME XLI. Carbohydrate Metabolism (Part B)

Edited by W. A. WOOD VOLUME XLII. Carbohydrate Metabolism (Part C)

Edited by W. A. WOOD VOLUME XLIII. Antibiotics Edited by JOHN H. HASH VOLUME XLIV. Immobilized Enzymes

Edited by KLAUS MOSBACH VOLUME XLV. Proteolytic Enzymes (Part B)

Edited by LASZLO LORAND

xxii

METHODS IN ENZYMOLOGY

VOLUME XLVI. Affinity Labeling

Edited by WILLIAM B. JAKOBY AND MEIR WILCHEK VOLUME XLVII. Enzyme Structure (Part E)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME XLVIII. Enzyme Structure (Part F)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEEF VOLUME XLIX. Enzyme Structure (Part G) Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME L. Complex Carbohydrates (Part C)

Edited by VICTOR GINSBURG VOLUME LI. Purine and Pyrimidine Nucleotide Metabolism

Edited by PATRICIA A. HOFFEE AND MARY ELLEN JONES VOLUME LII. Biomembranes (Part C: Biological Oxidations) Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME LIII. Biomembranes (Part D" Biological Oxidations)

Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME LIV. Biomembranes (Part E: Biological Oxidations)

Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME LV. Biomembranes (Part F: Bioenergetics)

Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME LVI. Biomembranes (Part G: Bioenergetics)

Edited by SIDNEY FLEISCHER AND LESTER PACKER VOLUME LVII. Bioluminescence and Chemiluminescence

Edited by MARLENE A. DELUCA VOLUME LVIII. Cell Culture

Edited by WILLIAM B. JAKOBY AND IRA PASTAN VOLUME LIX. Nucleic Acids and Protein Synthesis (Part G)

Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN

M E T H O D S IN E N Z Y M O L O G Y

xxiii

VOLUME LX. Nucleic Acids and Protein Synthesis (Part H)

Edited by KIVlE MOLDAVE AND LAWRENCE GROSSMAN VOLUME 61. Enzyme Structure (Part H)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME 62. Vitamins and Coenzymes (Part D) Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME 63. Enzyme Kinetics and Mechanism (Part A: Initial Rate and Inhibitor Methods) Edited by DANIEL L. PURICH VOLUME 64. Enzyme Kinetics and Mechanism (Part B: Isotopic Probes and Complex Enzyme Systems) Edited by DANIEL L. PURICH VOLUME 65. Nucleic Acids (Part I)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME 66. Vitamins and Coenzymes (Part E) Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME 67. Vitamins and Coenzymes (Part F) Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT VOLUME 68. Recombinant DNA

Edited by RAY Wu VOLUME 69. Photosynthesis and Nitrogen Fixation (Part C)

Edited by ANTHONY SAN PIETRO VOLUME 70. Immunochemical Techniques (Part A) Edited by HELEN VAN VUNAKIS AND JOHN J. LANGONE VOLUME 71. Lipids (Part C)

Edited by JOHN M. LOWENSTEIN VOLUME 72. Lipids (Part D)

Edited by JOHN M. LOWENSTEIN

xxiv

METHODS IN ENZYMOLOGY

VOLUME 73. Immunochemical Techniques (Part B) Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS VOLUME 74. Immunochemical Techniques (Part C) Edited by JOHN J. LANGONEAND HELEN VAN VUNAK1S VOLUME 75. Cumulative Subject Index Volumes XXXI, XXXII, XXXIV-LX Edited by EDWARD A. DENNIS AND MARTHAG. DENNIS VOLUME 76. Hemoglobins

Edited by ERALDO ANTONINI, LUIGI ROSSI-BERNARDI, AND EMILIA CHIANCONE

VOLUME 77. Detoxication and Drug Metabolism Edited by WILLIAM B. JAKOBY VOLUME 78. Interferons (Part A)

Edited by SIDNEY PESTKA VOLUME 79. Interferons (Part B)

Edited by SIDNEY PESTKA VOLUME 80. Proteolytic Enzymes (Part C)

Edited by LASZLO LORAND VOLUME 81. Biomembranes (Part H: Visual Pigments and Purple Membranes, I) Edited by LESTER PACKER VOLUME 82. Structural and Contractile Proteins (Part A: Extracellular Matrix) Edited by LEON W. CUNNINGHAM AND DIXIE W. FREDERIKSEN VOLUME 83. Complex Carbohydrates (Part D) Edited by VICTOR GINSBURG VOLUME 84. Immunochemical Techniques (Part D: Selected Immunoassays)

Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS

M E T H O D S IN E N Z Y M O L O G Y

XXV

VOLUME 85. Structural and Contractile Proteins (Part B: The Contractile Apparatus and the Cytoskeleton) Edited by DIXIE W. FREDERIKSEN AND LEON W. CUNN1NGHAM VOLUME 86. Prostaglandins and Arachidonate Metabolites

Edited by WILLIAM E. M. LANDS AND WILLIAM L. SMITH VOLUME 87. Enzyme Kinetics and Mechanism (Part C: Intermediates, Stereochemistry, and Rate Studies) Edited by DANIEL L. PURICH VOLUME 88. Biomembranes (Part I: Visual Pigments and Purple Membranes, II) Edited by LESTER PACKER VOLUME 89. Carbohydrate Metabolism (Part D)

Edited by WILLIS A. WOOD VOLUME 90. Carbohydrate Metabolism (Part E)

Edited by WILLIS A. W o o d VOLUME 91. Enzyme Structure (Part I)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME 92. Immunochemical Techniques (Part E: Monoclonal Antibodies and General Immunoassay Methods) Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS VOLUME 93. Immunochemical Techniques (Part F: Conventional Antibodies, Fc Receptors, and Cytotoxicity) Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS VOLUME 94. Polyamines

Edited by HERBERT TABOR AND CELIA WHITE TABOR VOLUME 95. Cumulative Subject Index Volumes 61-74, 76-80

Edited by EDWARD A. DENNIS AND MARTHA G. DENNIS VOLUME 96. Biomembranes [Part J: Membrane Biogenesis: Assembly and Targeting (General Methods; Eukaryotes)] Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER

xxvi

METHODS IN ENZYMOLOGY

VOLUME 97. Biomembranes [Part K: Membrane Biogenesis: Assembly and Targeting (Prokaryotes, Mitochondria, and Chloroplasts)]

Edited by

SIDNEY FLEISCHER AND BECCA FLEISCHER

VOLUME 98. Biomembranes (Part L: Membrane Biogenesis: Processing and Recycling) Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER VOLUME 99. Hormone Action (Part F: Protein Kinases)

Edited by JACKIE D. CORBIN AND JOEL G. HARDMAN VOLUME 100. Recombinant DNA (Part B)

Edited by RAY Wu, LAWRENCE GROSSMAN, AND KIVIE MOLDAVE VOLUME 101. Recombinant DNA (Part C)

Edited by RAY Wu, LAWRENCE GROSSMAN, AND KIVIE MOLDAVE VOLUME 102. Hormone Action (Part G: Calmodulin and Calcium-Binding Proteins) Edited by ANTHONY R. MEANS AND BERT W. O'MALLEY VOLUME 103. Hormone Action (Part H: Neuroendocrine Peptides)

Edited by P. MICHAEL CONN VOLUME 104. Enzyme Purification and Related Techniques (Part C)

Edited by WILLIAM B. JAKOBY VOLUME 105. Oxygen Radicals in Biological Systems

Edited by LESTER PACKER VOLUME 106. Posttranslational Modifications (Part A)

Edited by FINN WOLD AND KIVIE MOLDAVE VOLUME 107. Posttranslational Modifications (Part B)

Edited by FINN WOLD AND KIVIE MOLDAVE VOLUME 108. Immunochemical Techniques (Part G: Separation and Characterization of Lymphoid Cells) Edited by GIOVANNI DI SABATO, JOHN J. LANGONE, AND HELEN VAN VUNAKIS

METHODSIN ENZYMOLOGY

xxvii

VOLUME 109. Hormone Action (Part I: Peptide Hormones)

Edited by LUTZ BIRNBAUMERAND BERT W. O'MALLEY VOLUME 110. Steroids and Isoprenoids (Part A) Edited by JOHN H. LAW AND HANS C. RILLING VOLUME I 11. Steroids and Isoprenoids (Part B)

Edited by JOHN H. LAW AND HANS C. RILLING VOLUME 112. Drug and Enzyme Targeting (Part A)

Edited by KENNETH J. WIDDER AND RALPH GREEN VOLUME 113. Glutamate, Glutamine, Glutathione, and Related Corn-

pounds

Edited by ALTON MEISTER VOLUME 114. Diffraction Methods for Biological Macromolecules (Part A)

Edited by HAROLD W. WYCKOEE, C. H. W. HIRS, AND SERGE N. TIMASHEFF

VOLUME 115. Diffraction Methods for Biological Macromolecules (Part

B) Edited by HAROLD W. WYCKOEE, C. H. W. HIRS, AND SERGE N. TIMASHEFF

VOLUME 116. Immunochemical Techniques (Part H: Effectors and Mediators of Lymphoid Cell Functions) Edited by GIOVANNI DI SABATO, JOHN J. LANGONE, AND HELEN VAN VUNAKIS

VOLUME 117. Enzyme Structure (Part J)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME 118. Plant Molecular Biology

Edited by ARTHUR WEISSBACHAND HERBERT WEISSBACH VOLUME 119. Interferons (Part C)

Edited by SIDNEY PESTKA VOLUME 120. Cumulative Subject Index Volumes 81-94, 96-101

XXVIn

METHODS IN ENZYMOLOGY

VOLUME 121. Immunochemical Techniques (Part I: Hybridoma Technology and Monoclonal Antibodies) Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS VOLUME 122. Vitamins and Coenzymes (Part G)

Edited by FRANK CHYTIL AND DONALD B. MCCORMICK VOLUME 123. Vitamins and Coenzymes (Part H)

Edited by FRANK CHYTIL AND DONALD B. MCCORMICK VOLUME 124. Hormone Action (Part J: Neuroendocrine Peptides)

Edited by P. MICHAEL CONN VOLUME 125. Biomembranes (Part M: Transport in Bacteria, Mitochondria, and Chloroplasts: General Approaches and Transport Systems) Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER VOLUME 126. Biomembranes (Part N: Transport in Bacteria, Mitochondria, and Chloroplasts: Protonmotive Force) Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER VOLUME 127. Biomembranes (Part O: Protons and Water: Structure and Translocation) Edited by LESTER PACKER VOLUME 128. Plasma Lipoproteins (Part A: Preparation, Structure, and Molecular Biology) Edited by JERE P. SEGREST AND JOHN J. ALBERS VOLUME 129. Plasma Lipoproteins (Part B: Characterization, Cell Biology, and Metabolism) Edited by JOHN J. ALBERS AND JERE P. SEGREST VOLUME 130. Enzyme Structure (Part K)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME 131. Enzyme Structure (Part L)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF VOLUME 132. Immunochemical Techniques (Part J: Phagocytosis and Cell-Mediated Cytotoxicity) Edited by GIOVANNI DI SABATO AND JOHANNES EVERSE

METHODS IN ENZYMOLOGY

xxix

VOLUME 133. Bioluminescence and Chemiluminescence (Part B)

Edited by MARLENE DELUCA AND WILLIAM D. MCELROY VOLUME 134. Structural and Contractile Proteins (Part C: The Contractile Apparatus and the Cytoskeleton) Edited by RICHARD B. VALLEE VOLUME 135. Immobilized Enzymes and Cells (Part B)

Edited by KLAUS MOSBACH VOLUME 136. Immobilized Enzymes and Cells (Part C)

Edited by KLAUS MOSBACH VOLUME 137. Immobilized Enzymes and Cells (Part D) (in preparation)

Edited by KLAUS MOSBACH VOLUME 138. Complex Carbohydrates (Part E)

Edited by VICTOR GINSBURG VOLUME 139. Cellular Regulators (Part A: Calcium- and CalmodulinBinding Proteins) Edited by ANTHONY R. MEANS AND P. MICHAEL CONN VOLUME 140. Cumulative Subject Index Volumes I02-119, 121-134 VOLUME 141. Cellular Regulators (Part B: Calcium and Lipids)

Edited by P. MICHAEL CONN AND ANTHONY R. MEANS VOLUME 142. Metabolism of Aromatic Amino Acids and Amines

Edited by SEYMOUR KAUFMAN VOLUME 143. Sulfur and Sulfur Amino Acids

Edited by WILLIAM B. JAKOBY AND OWEN GRIFFITH VOLUME 144. Structural and Contractile Proteins (Part D: Extracellular Matrix) Edited by LEON W. CUNNINGHAM VOLUME 145. Structural and Contractile Proteins (Part E: Extracellular Matrix) Edited by LEON W. CUNNINGHAM

XXX

M E T H O D S IN E N Z Y M O L O G Y

VOLUME 146. Peptide Growth Factors (Part A) (in preparation)

Edited by DAVID BARNES AND DAVID A. SIRBASKU VOLUME 147. Peptide Growth Factors (Part B) (in preparation)

Edited by DAVID BARNES AND DAVID A. SIRBASKU VOLUME 148. Plant Cell Membranes (in preparation)

Edited by LESTER PACKER AND ROLAND DOUCE VOLUME 149. Drug and Enzyme Targeting (Part B) (in preparation)

Edited by RALPH GREEN AND KENNETH J. WIDDER VOLUME 150. Immunocheraical Techniques (Part K: In Vitro Models of B and T Cell Functions and Lymphoid Cell Receptors) (in preparation) Edited by GIOVANNI DI SABATO VOLUME 151. Molecular Genetics of Mammalian Cells (in preparation)

Edited by MICHAEL M. GOTTESMAN VOLUME 152. Guide to Molecular Cloning Techniques

Edited by SHELBY L. BERGER AND ALAN R. KIMMEL VOLUME 153. Recombinant DNA (Part D) (in preparation)

Edited by RAY Wu AND LAWRENCE GROSSMAN VOLUME 154. Recombinant DNA (Part E) (in preparation)

Edited by RAY Wu AND LAWRENCE GROSSMAN VOLUME 155. Recombinant DNA (Part F) (in preparation)

Edited by RAY Wu

[1]

IMMOBILIZED

ACTIVE COENZYMES

3

[1] I m m o b i l i z e d A c t i v e C o e n z y m e s By MATS-OLLE MANSSON and K t n u s MOSBACH This section includes a number of contributions concerning immobilized multienzyme systems and coenzymes. Other contributions addressing these systems in a different context are also found in Vols. 135 and 136 (articles by Rozzell, Luisi and Steinmann-Hofmann, Scheller et al., Carrea and Cremonesi, Murachi and Tabata, Laane and Veeger, and Chibata et al.). Furthermore, in a previous volume in this series I similar questions and methodologies have been presented. In the following we wish to discuss briefly general aspects especially related to immobilized coenzymes and also to provide a number of additional references. Coenzyme-dependent enzymes are of great potential interest for the synthesis of interesting stereospecific compounds and for analytical applications. Paralleling attempts to use immobilized coenzymes or coenzyme analogs in affinity chromatography 2 the question of their coenzymatic activity in the immobilized state has been addressed. The first immobilization, e.g., of NAD(H), showing coenzymatic activity was reported in 1971.3 Later, synthesis of chemically better defined analogs, several of which can be seen in Table I, was carried out. The main interest has focused on the redox coenzyme NAD, but NADP, ATP, and CoA have also been immobilized as active coenzymes. Preparation of Immobilized Active Coenzymes Coenzymes usually have to be modified to allow proper immobilization and regeneration. This modification of a coenzyme can be accomplished in basically two ways. One is the preassembly approach and the other has been called the solid-phase modular approach. In the preassembly approach the coenzyme is first modified and then assembled with a spacer molecule to increase the steric availability for the coenzyme of the enzyme and subsequently coupled to a support. In the solid-phase modular approach the coenzyme or a modified coenzyme is coupled to a matrix previously substituted with spacer molecules carrying reactive functional groups. 1 K. Mosbach (ed.), this series, Vol. 44. 2 K. Mosbach, in "Advances in Enzymology" (A. Meister, ed.), Vol. 46, p. 205. Wiley, New York, 1978. 3 p. O. Larsson and K. Mosbach, Biotechnol. Bioeng. 13, 393 (1971).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

4

[1]

M U L T I S T E P E N Z Y M E SYSTEMS A N D C O E N Z Y M E S

TABLE I SUBST1TUENTS USED FOR MODIFICATION OF N A D AND A T P IN THE

Compound NAD

Substituent

N-6 POSITION Reference

CH2COOH CH2~NH(CH2)6NH2

a

O CH2~HCH2COOH OH (CH2)2NH2 (CH2)6NH2 (CH2)2~NHCH(CH2)4NH~CH~-~-CH2 O

C~O

c,d e f

O

I

OCH 3 CH2CHCH2OCCH~CH2

I

II

II

II

OH O CH2CNHCH2CNHCH2CHCH2NHCC~CH2

I

III

O O OH CH2CNH(CH2)6NHCC~CH2

II

OCH3

III

O OCHs (CH2)2CNH(CH2)2CNH(CH2)2NHCC~---CH2

II

II

III

O O OCH3 CH2CNH(CH2)2CNH(CH2)2NHCC---~CH2

II

II

O O (CH2)2COOH (CH2)2~NH(CH2)2NH2 ATP

O CH2COOH CH2~NH(CH2)6NH2

III

OCH3

k k

O CNH(CH2)6NH2

II

O CHzCNH(CH2)6NHCC~CH2

II

O (CH2)6NH2

III

OCH3

M. Lindberg, P. O. Larsson, and K. Mosbach, Eur. J. Biochem. 40, 187 (1973). b p. Zappelli, A. Rossodivita, and L. Re, Eur. J. Biochem. 54, 475 (1975). c H. L. Schmidt and G. Grenner, Eur. J. Biochem. 67, 295 (1976). a A. F. Buckmann, M.-R. Kula, R. Wichmann, and C. Wandrey, J. Appl. Biochem. 3, 301 (1981). e D. B. Craven, M. J. Harvey, and P, D. G. Dean, FEBS Lett. 38, 320 (1974). a

[1]

IMMOBILIZED ACTIVE COENZYMES

5

In the authors' opinion the preassembly approach has significant advantages. First, chemically well-defined preparations can be obtained and second, the preassembled coenzyme-spacer molecule can be characterized prior to coupling and studied in homogeneous solution, yielding useful information in advance concerning kinetic constants such as Kdi~s and Km. The different variations of the solid-phase modular approach all suffer from the drawback that heterogeneous preparations containing excess spacers on the matrix may be obtained which can give rise to unwanted nonspecific interactions with the enzyme. The syntheses of NAD(P) and ATP analogs have been described in a previous volume of this series.l The probably most widely applied synthetic route involves alkylation of the N-1 nitrogen of the adenine nucleus [a part of both NAD(P) and ATP] by reagents such as iodoacetic acid, 4 3,4-epoxybutanoic acid, 5 or ethyleneimine. 6,7 After reduction of the alkaline-labile NAD to the stable NADH, a Dimroth rearrangement is carried out to convert the N-1 derivative to the corresponding desired N-6 analog. This N-6 analog is then used as a starting material for the synthesis of analogs with different spacers carrying different reactive groups (Table I). For an immobilized coenzyme to be used as an active coenzyme in an enzyme reactor or as a component of a biosensor, special requirements have to be fulfilled. Water-soluble polymers should be used as they do not cause as much diffusional hindrance as a water-insoluble matrix. In addition the degree of substitution is of importance since it has been reported that a high degree of substitution will lead to inhibition of the enzyme. 8 Widely used water-soluble polymers are dextran and polyethylene glycol 4 M. Lindberg, P. O. Larsson, and K. Mosbach, Eur. J. Biochem. 40, 187 (1973). 5 p. Zappelli, A. Rossodivita, and L. Re, Eur. J. Biochem. 54, 475 (1975). 6 H. L. Schmidt and G. Grenner, Eur. J. Biochem. 67, 295 (1976). 7 A. F. Btickmann, M.-R. Kula, R. Wichmann, and C. Wandrey, J. Appl. Biochem. 3, 301 (1981). s S. Furukawa, I. Urabe, and H. Okada, Eur. J. Biochem. 114, 101 (1981).

I M . Muramatsu, I. Urabe, Y. Yamada, and H. Okada, Eur. J. Biochem. 80, 111 (1977). g F. LeGoffic, S. Sicsic, and C. Vincent, in "Enzyme Engineering" (W. H. Weetall and G. P. Royer, eds.), Vol. 5, p. 127. Plenum, New York, 1980. h y. Yamazaki and H. Maeda, Agric. Biol. Chem. 45, 2277 (1981). i y. Yamazaki, H. Maeda, A. Satoh, and K. Hiromi, Biochem. J. 95, 109 (1984). J S. Adachi, M. Ogata, H. Tobita, and K. Hasimoto, Enzyme Microb. TechnoL 6, 259 (1984). , M. Lindberg and K. Mosbach, Eur. J. Biochem. 53, 481 (1975). I H. Suzuki and Y. Yamazaki, this volume [5]. my. Yamazaki and H. Maeda, Agric. Biol. Chem. 45, 2091 (1981). " W. Berke, M. Morr, C. Wandrey, and M.-R. Kula, Ann. N.Y. Acad. Sci. 437, 257 (1984).

6

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[1]

(PEG), but polymers such as polyethyleneimine (PEI) have also been used. Polymerizable coenzymes such as the acrylic derivatives listed in Table I are immobilized in a polymerization reaction together with other monomers, e.g., acrylamide. The coenzyme content of the final polymer will then depend on the ratio of coenzyme monomer to the other monomers added. The polymeric derivatives obtained in this way are heterogeneous mixtures of various polymer chains. The "coenzymatic activity" of such a coenzyme moiety placed on a polymer chain may depend on its position in the polymer and on the size of the polymer. 9 Properties of Immobilized Active Coenzymes A coenzyme analog synthesized according to the preassembly method has, as mentioned before, the advantage that it can be thoroughly studied, both chemically and enzymatically prior to immobilization. One aspect of great importance is whether the modification has influenced on the interactions between enzyme and coenzyme. Generally, the N-6 position of NAD is the point of choice for substitution since it has been shown that for many dehydrogenases the exocyclic amino group of the parent NAD molecule points out of the active site. 1° Thus, spacers attached at this position also will project out from the active site of the enzyme, and interference with coenzyme binding will be minimized or nonexistent. However in some instances the C-8 position has also proved useful, at least in affinity chromatography (for an overview, see Ref. 11). The easiest way to determine the amount of immobilized coenzyme on a water-soluble polymer is to record the UV spectrum for the coenzyme and, based on the absorption peak, determine its concentration. 1 It is important that the reference cell during recording of the UV spectrum carries the equivalent amount of nonsubstituted polymer. The degree to which an immobilized coenzyme can be enzymatically reduced or oxidized is also of great interest. The kinetic properties of immobilized NAD derivatives vary depending on both what enzyme and what coenzyme derivative are being used. In general the Km values are higher and the Vmaxvalues are lower for the immobilized NAD than for the free NAD. In a comparison of coenzymatic activity between free NAD, N-6-substituted NAD, dextran-NAD,

9 S. Adachi, M. Ogata, H. Tobita, and K. Hashimoto, Enzyme Microb. Technol. 6, 259 (1984). J0 C. I. Br~inden, H. JOrnvall, H. Eklund, and H. Furugren, in "The Enzymes" (P. D. Boyer, ed.), Vol. 11, 3rd Ed., p. 103. Academic Press, New York, 1975. " M. O. M~nsson and K. Mosbach, in "Coenzymes and Cofactors" (D. Dolphin, R. Poulson, and O. Avramovic, eds.). Wiley, New York, in press.

[1]

IMMOBILIZED ACTIVE COENZYMES

7

and agarose-bound NAD, the relative initial rates toward alcohol dehydrogenase were found to be 100, 61, 16, and 0.7, respectively. ~2 Apart from the different number of interactions possible between coenzyme and support that are dependent on the size of the latter, there are two factors which influence these initial reaction rates when compared to the free coenzyme. First, the modification of the coenzyme can alter its binding to the active site. Also, the polymer to which the coenzyme is anchored may influence the binding of the coenzyme to the active site, e.g., because of its hydrophilic/hydrophobic character. In different reports, the effect of polymer molecular weight, 9 NAD density on the polymer, 8 and length of the spacer ~3 on the enzymatic reaction have been discussed. Applications of Immobilized Active Coenzymes Immobilized active coenzymes such as NAD have found use in analytical systems as well as in enzyme reactors. The reactor of choice is the membrane reactor in which a semipermeable membrane retains the high molecular weight immobilized coenzyme together with the enzymes. 14 The design of the reactor will of course be different if the enzyme-coenzyme system is of a totally "self-sustained" form where both the enzyme and the coenzyme are integrated parts of the same polymer. ~5Alternative systems that allow continuous processes to be carried out have also been studied. Following one approach, an NAD analog has been coupled directly onto an enzyme, allowing it to act as a prosthetic group~6,~7; alternatively the enzyme is coimmobilized with an NAD analog to a support in a predetermined configuration. 16,~8 An important issue both in the reactor systems and for analytical systems is the requirement for regeneration of the coenzyme because of its high cost and the fact that it is needed in stoichiometric amounts relative to the product formed. Regeneration can be accomplished chemically, electrochemically, or enzymatically. At present, enzymatic regeneration has the most advantages, especially because of its high specificity in regeneration. However, electrochemical procedures are gaining in importance (in this context see Ref. 18a). Table II summarizes several appli12 p. O. Larsson and K. Mosbach, FEBS Lett. 46, 119 (1974). J3 y . Yamazaki, H. Maeda, A. Satoh, and K. Hiromi, Biochem. J. 95, 109 (1984). ~4 M.-R. Kula and C. Wandrey, this volume [2]. ~5Y. Yamazaki and H. Maeda, this volume [3]. 16 M. O. Mhnsson, P. O. Larsson, and K. Mosbach, this series, Vol. 89, p. 457. x7 C. Woenckhaus, R. Koob, A. Burkhard, and H. G. Schaefer, Bioorg. Chem. 12, 45 (1983). ~s K. J. Laidler and M. A. Mazid, this volume [6]. ~8~I. Thanos, J. Bader, H. Giinther, S. Neumann, F. Krauss, and H. Simon, this volume [28].

8

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[1]

TABLE II APPLICATIONS OF IMMOBILIZED N A D ANALOGSa

Product or compound analyzed Enzyme reactors with Alanine Alanine Leucine Benzaldehyde Lactate Lactate Lactate Lactate Lactate Alanine Malate Malate Propanediol Propanediol Lactate Phenylalanine Analytical systems Glutamate Pyruvate Alcohol Alcohol Lactate Alcohol Alcohol Lactate Glutamate Lactate Glutamate

Coenzyme

Enzymesb

coupled enzyme regeneration Dextran-NAD Dextran-NAD PEG-NAD Polyacrylate-NAD PEG-NAD Dextran-NAD PEI-NAD Polylysine-NAD Dextran-NAD PEI-NAD Polyacrylate-NAD Dextran-NAD Coimmobilized NAD Enzyme-bound NAD Enzyme-bound NAD PEG-NAD

GalDH + AlaDH AIaDH + LDH FDH + LeuDH ADH (*) ADH + LDH ADH + LDH ADH + LDH ADH + LDH ADH + LDH AIaDH + LDH FDH + MDH ADH + MDH ADH (*) ADH (*) ADH + LDH FDH + PheDH

Dextran-NAD Dextran-NAD Polyacrylate-NAD Coimmobilized NAD Electrode-bound NAD Electrode-bound NAD Dextran-NAD Dextran-NAD Dextran-NAD Dextran-NAD Dextran-NAD

GDH + LDH GDH + LDH ADH + diaphorase ADH LDH ADH ADH LDH GDH LDH GDH

Reference

c e : e h J k m

n.o n P q q r s s t t ' "

GalDH, Galactose dehydrogenase; AIaDH, alanine dehydrogenase; FDH, formate dehydrogenase; LeuDH, leucine dehydrogenase; ADH, alcohol dehydrogenase, LDH, lactate dehydrogenase; MDH, malate dehydrogenase; GDH, glutamate dehydrogenase; and PheDH, phenylalanine dehydrogenase. b (,) Indicates coupled-substrate enzymatic regeneration. c p. Davies and K. Mosbach, Biochim. Biophys. Acta 370, 329 (1974). a M.-R. Kula and C. Wandrey, this volume [2]. e C. W. Fuller, J. R. Rubin, and H. J. Bright, Eur. J. Biochem. 103, 421 (1980). z H. Okada and I. Urabe, this volume [4]. g P. O. Larsson and K. Mosbach, FEBS Lett. 46, 119 (1974). h j. R. Wykes, P. Dunill, and M. D. Lilly, Biotechnol. Bioeng. 17, 151 (1975). i y . Yamazaki, H. Maeda, and H. Suzuki, Biotechnol. Bioeng. 18, 1761 (1976). J Y. Morikawa, I. Karube, and S. Suzuki, Biochim. Biophys. Acta 523, 263 (1978). k W. Marconi, G. Prosperi, S. Giovenco, and F. Morisi, J. Mol. Catal. 1, 111 (19751976).

[2]

CONTINUOUS ENZYMATIC TRANSFORMATION

9

cations of immobilized active N A D in analysis and as a component in enzyme reactors. In addition a polyethylene glycol-bound ATP ~9 and N A D P 2° have recently been used in enzyme reactor studies. It should be kept in mind that native coenzymes for a number of applications remain an alternative (see, e.g., the contribution on ATP recycling). 2~ In this context a recent publication on "an affinity chromatographic reactor for highly efficient turnover of dissociable cofactors" applying free NAD in an ultrafiltration hollow fiber tube should be mentioned. 22 19 W. Berke, M. Morr, C. Wandrey, and M.-R. Kula, Ann. N. Y. Acad. Sci. 437, 257 (1984). 20 K. Okuda, I. Urabe, and H. Okada, Eur. J. Biochem. 151, 33 (1985). 2~ D. C. Crans, R. J. Kazlauskas, B. L. Hirschbein, C.-H. Wong, O. Abril, and G. M. Whitesides, this volume [25]. 22 O. Miyawaki, N. Osato, and T. Yano, Agric. Biol. Chem. 49, 2063 (1985).

I y . Yamazaki and H. Maeda, this volume [3]. m j. Grunwald and T. M. S. Chang, J. Mol. Catal. 11, 83 (1981). n M. O. MAnsson, P. O. Larsson, and K. Mosbach, this series, Vol. 89, p. 457, o S. Gestrelius, M. O. MAnsson, and K. Mosbach, Eur. J. Biochem. 57, 529 (1975). p E. Schmidt, E. Fiolitakis, and C. Wandrey, Enzyme Eng. 8 (in press). q H. Suzuki and Y. Yamazaki, this volume [5]. • K. J. Laidler and M. A. Mazid, this volume [6]. s T. Yao and S. Musha, Anal. Chim. Acta 110, 203 (1979). t A. Malinauskas and J. Kulys, Anal. Chim. Acta 98, 31 (1978). " Y. Sakaguchi, M. Sugahara, J. Endo, and T. Murachi, J. Appl. Biochem. 3, 32 (1981).

[2] C o n t i n u o u s E n z y m a t i c T r a n s f o r m a t i o n in an Enzyme-Membrane Reactor with Simultaneous NADH Regeneration B y M A R I A - R E G I N A K U L A a n d CHRISTIAN WANDREY

Enzymes as catalysts are especially noted for their high stereo- and regiospecificity. To accomplish the synthesis of chiral compounds often not only an enzyme is needed as catalyst but in addition a low molecular weight coenzyme which participates in the reaction. For a large number of redox processes NAD(H) or NADP(H) serves as coenzymes for dehydrogenases and is utilized in stoichiometric amounts in the course of the reaction. These coenzymes are readily dissociable from the enzyme and require a separate second reaction for regeneration. I S. S. Wang and C. K. King, Adv. Biochem. Eng. 12, 119 (1979). METHODSIN ENZYMOLOGY,VOL. 136

Copyright© 1987by AcademicPress, Inc. All rightsof reproductionin any form reserved.

[2]

CONTINUOUS ENZYMATIC TRANSFORMATION

9

cations of immobilized active N A D in analysis and as a component in enzyme reactors. In addition a polyethylene glycol-bound ATP ~9 and N A D P 2° have recently been used in enzyme reactor studies. It should be kept in mind that native coenzymes for a number of applications remain an alternative (see, e.g., the contribution on ATP recycling). 2~ In this context a recent publication on "an affinity chromatographic reactor for highly efficient turnover of dissociable cofactors" applying free NAD in an ultrafiltration hollow fiber tube should be mentioned. 22 19 W. Berke, M. Morr, C. Wandrey, and M.-R. Kula, Ann. N. Y. Acad. Sci. 437, 257 (1984). 20 K. Okuda, I. Urabe, and H. Okada, Eur. J. Biochem. 151, 33 (1985). 2~ D. C. Crans, R. J. Kazlauskas, B. L. Hirschbein, C.-H. Wong, O. Abril, and G. M. Whitesides, this volume [25]. 22 O. Miyawaki, N. Osato, and T. Yano, Agric. Biol. Chem. 49, 2063 (1985).

I y . Yamazaki and H. Maeda, this volume [3]. m j. Grunwald and T. M. S. Chang, J. Mol. Catal. 11, 83 (1981). n M. O. MAnsson, P. O. Larsson, and K. Mosbach, this series, Vol. 89, p. 457, o S. Gestrelius, M. O. MAnsson, and K. Mosbach, Eur. J. Biochem. 57, 529 (1975). p E. Schmidt, E. Fiolitakis, and C. Wandrey, Enzyme Eng. 8 (in press). q H. Suzuki and Y. Yamazaki, this volume [5]. • K. J. Laidler and M. A. Mazid, this volume [6]. s T. Yao and S. Musha, Anal. Chim. Acta 110, 203 (1979). t A. Malinauskas and J. Kulys, Anal. Chim. Acta 98, 31 (1978). " Y. Sakaguchi, M. Sugahara, J. Endo, and T. Murachi, J. Appl. Biochem. 3, 32 (1981).

[2] C o n t i n u o u s E n z y m a t i c T r a n s f o r m a t i o n in an Enzyme-Membrane Reactor with Simultaneous NADH Regeneration B y M A R I A - R E G I N A K U L A a n d CHRISTIAN WANDREY

Enzymes as catalysts are especially noted for their high stereo- and regiospecificity. To accomplish the synthesis of chiral compounds often not only an enzyme is needed as catalyst but in addition a low molecular weight coenzyme which participates in the reaction. For a large number of redox processes NAD(H) or NADP(H) serves as coenzymes for dehydrogenases and is utilized in stoichiometric amounts in the course of the reaction. These coenzymes are readily dissociable from the enzyme and require a separate second reaction for regeneration. I S. S. Wang and C. K. King, Adv. Biochem. Eng. 12, 119 (1979). METHODSIN ENZYMOLOGY,VOL. 136

Copyright© 1987by AcademicPress, Inc. All rightsof reproductionin any form reserved.

l0

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[2]

Since NADH and NADPH are complex and rather labile organic chemicals they are quite expensive. For the application of dehydrogenases in the enzyme-catalyzed synthesis of chiral compounds an efficient coenzyme regeneration is therefore needed in order to make such a process economically viable. A simple comparison of cost of I mol of NADH (665 g/mol, value -$1000) with 1 mol of chiral a-hydroxy or a-amino acid (-100 g/mol, value -$2) shows that separation, external regeneration, and recycling of the coenzyme would require an overall yield of 99.98% per pass to keep the coenzyme cost at 10% of the product value. Simultaneous regeneration offers a chance to overcome the stoichiometric relation between product and coenzyme by an efficient internal recycling. In principle several experimental approaches are possible. ~-6 Here we concentrate on an enzymatic regeneration of NADH by a second enzyme-catalyzed reaction. Our approach follows the strategy developed in cellular metabolism in which NADH is produced and consumed by a series of balanced reactions. 6-~° Considering ultimately a technical application it should be noted that besides the anticipated conversion, a second substrate is now needed in a stoichiometric amount and two products are formed in equimolar concentration. Therefore the cosubstrate should be cheap enough and neither cosubstrate or coproduct should be obnoxious to the enzymes involved. In the authors' laboratories formate dehydrogenase is preferred as a general NADH-regenerating enzyme for several reasons: the enzyme is readily available in large amounts from Candida boidinii grown on methanolH,~2; the equilibrium of the reaction catalyzed lies far on the side of CO2 and therefore NADH formationS1; and the 2 H. Simon, J. Bader, H. Giinther, S. Neumann, and J. Thanos, Ann. N. Y. Acad. Sci. 434, 171 (1984). 3 Z. Shaked, J. J. Barber, and G. M. Whitesides, J. Org. Chem. 46, 4100 (1981). 4 M. O. M~nsson, P. O. Larsson, and K. Mosbach, this series, Vol. 89, p. 457. M. Aizawa, R. W. Coughlin, and M. Charles, Biotechnol. Bioeng. 18, 209 (1976). 6 R. Wichmann, C. Wandrey, A. F. Biickmann, and M.-R. Kula, Biotechnol. Bioeng. 23, 2789 (1981). 7 C. Wandrey, R. Wichmann, W. Leuchtenberger, A. B/ickmann, and M.-R. Kula, U.S. Patent 4,326,031, European Patent 80,104,346.4; U.S. Patent 4,304,858, European Patent 80,104,345.6. s W. Leuchtenberger, C. Wandrey, and M.-R. Kula, DOS 3,307,094.6, Spanish Patent 530,155. 9 A. F. Bfickmann, U.S. Patent 4,443,594, European Patent 79,102,954.9. ~ C- Wandrey and R. Wichmann, in "Application of Isolated Enzymes and Immobilized Cells to Biotechnology" (A. Laskin, ed.), p. 177. 1985. H H. Schtitte, J. Flossdorf, H. Sahm, and M.-R. Kula, Eur. J. Bioehem. 62, 151 (1976). ~2 K. H. Kroner, H. Schiitte, W. Stach, and M,-R. Kula, J. Chem. Technol. Bioteehnol., Biotechnol. 32B, 130 (1982).

[2]

CONTINUOUS ENZYMATIC TRANSFORMATION

HCO0-

reduced subsfrate

u.~

PEG2oooo-NAD"

/z~

•

PE52oooo-NAOH k0H

~A

C0z

1l

oxidized subsfrafe

FIG. 1. Reaction scheme. FDH, Formate dehydrogenase; E, enzyme; PEG 20000-NAD + NADH, coenzyme derivatives.

coproduct CO2 is easily separable. The overall reaction scheme is outlined in Fig. 1. It is apparent that for fast NADH recycling, ready access to both enzymes becomes crucial and reactor systems with inherent mass transfer resistances should be avoided. Therefore an enzyme-membrane reactor was chosen. Here the reaction will proceed in homogenous solution. 6-1° The enzymes are retained in the reactor by virtue of their high molecular weight, utilizing an ultrafiltration membrane as a selective barrier. However, the difference in molecular weight between product and coenzyme is not sufficient to achieve satisfactory results by a membrane separation. Therefore we attempted to increase the molecular weight of the coenzyme by attaching it covalently to a water-soluble polymer such as polyethylene glycol (PEG). 9 This way the retention of the coenzyme by an ultrafiltration membrane could be accomplished together with separation of the coenzyme from the product stream. Provided the coenzyme activity is not lost upon modification, the basic design of the reactor should allow continuous enzymatic conversion and regeneration of NADH to proceed. Coenzyme D e r i v a t i v e s The synthesis of the PEG-NADH is outlined in Fig. 2 and has been described in detail by Bfickmann et al. 9,13 The modified coenzyme is readily soluble in water, and its concentration can be determined from absorbance at 340 or 259 nm, respectively. Assuming that the molar absorption coefficients are identical with the native coenzyme the following values have been used for calculation: 6220 M -~ cm -I for NADH at 340 nm, and 18,000 M -~ cm -~ for NAD at 259 nm. A number of dehydrogenases has been tested with regard to their ability to accept the modified coenzyme. u A. F. Biickmann, M.-R. Kula, R. Wichmann, and C, Wandrey, J. Appl. Biochem. 3, 301 (1981).

12

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[2]

NH2 ~NH~ "~N/~"~ N

LRib-- p--p

\N /

"k

2H*

--.,bJ

PEG--COO-

L.ib-- p-- p--Rin /

CQrbodiimide NH2

reduction

h R i b - - p - - P--Rib ~

Dimroth

H

PEG--

C - - N --CH~--CH~~'- NH

H

H

in reorrQngement

O

o

N

N

NH2

LRib-- p--p--Rib

-]

FIG. 2. General synthesis route to modified coenzymes. Rib, Ribose; P, phosphate; PEG, polyethylene glycol.

TABLE I KINETIC PARAMETERS OF DEYHYDROGENASES FOR NATIVE NADH AND P E G - N A D H a

Conditions

Native NADH

PEG-NADH

Temperature

Vmax

Km

Vmax

Krn

Enzyme b

Substrate

pH

(°C)

(U/rag)

(mM)

(U/mg)

(raM)

AIaDH LeuDH PheDH L-LDH D-LDH L-HicDH b-HicHD FDH

Pyruvate Ketoleucine Phenylpyruvate Pyruvate Pyruvate Ketoleucine Ketomethionine Formate

9 8 8.5 9 9 8.5 8 8

25 25 25 25 25 25 25 40

20.9 13.2 37.2 308.0 37.0 511.0 259.0 1.2

0.022 0.033 0.047 0.001 0.289 0.044 0.170 0.300

10.3 11.5 17.7 150.0 4.3 214.0 240.0 2.6

1.290 0.029 0.099 0.080 0.060 0.144 0.240 0.330

a Activity per milligram protein in the enzyme preparation. b AlaDH, Alanine dehydrogenase; LeuDH, leucine dehydrogenase; PheDH, phenylalanine dehydrogenase; L-LDH, L-lactate dehydrogenase; D-LDH, D-lactate dehydrogenase; L-HicDH, L-hydroxyisocaproate dehydrogenase; D-HicDH, D-hydroxyisocaproate dehydrogenase; FDH, formate dehydrogenase.

[2]

13

CONTINUOUS ENZYMATIC TRANSFORMATION

The results are listed in Table I. Often the Vmaxvalue obtained with the derivative is in the range 50-100% in comparison with the native coenzyme, while the Km value may change by an order of magnitude. It should be noted that formate dehydrogenase exhibits an even higher initial reaction velocity with the modified coenzyme than with the native. 6 Besides the examples listed in Table I the following enzymes are known to accept P E G - N A D ( H ) as coenzyme: alcohol dehydrogenase, isopropanol dehydrogenase, malate dehydrogenase, glutamate dehydrogenase. So far, glucose dehydrogenase from Bacillus megaterium is the only dehydrogenase detected that does not accept P E G - N A D H as a substrate. Retention

The retention of the coenzyme in the reactor becomes very important for continuous operation. In Fig. 3 the loss of coenzyme by incomplete retention in the reactor is plotted as a function of the operating time. The data show that retentions better than 99.9% are necessary to avoid excessive loss of coenzyme at a typical mean residence time of 1 hr. The high

\

" " " ........ \

8.8

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

~

O

\ e. 4

\

.........

PEG_20000_NAD(H) R = 9 9 93 kELU = 1 7 X/d

.........

PEG-1000g-NADCH)~""~ R

\\

= gg82

kELU =

\,

\,

4 3

X

X/d

N ""

0.2 R

~'". 2 4 . 0 ~/cl "'"--.....,...

= 99.00

kELU I 8

2

I 4

I I~

I 8

I 18

I 12

I 14

1 10

I 18

I

I

2a

22

24

TIME l D Flo. 3. Elution of P E G - N A D ( H ) across a YM5 (Amicon) membrane in a continuously operated membrane reactor; residence time Z = l hr. R, Retention (R = 99.00% for comparison); ke~u, elution loss (due to incomplete retention); PEG 10000 and PEG 20000, molecular weight 10,000 and 20,000.

14

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[2]

retention is a very stringent requirement when choosing the right membrane and depends on the average molecular weight of the coenzyme derivative and also on the pore size distribution determining the actual cutoff behavior of a given membrane. The performance of a membrane selected should be checked by experiments lasting for several days in order to allow assessment of the data with sufficient confidence. Retention R is defined by Eq. (1): R = (Cr-

Cf)/Cr

(1)

where Cr and Cf are the concentration of the compound of interest in the retentate and filtrate, respectively. The washout can be described as a function of time by Eq. (2): Cr/Cro = e -[~1 - R)/~-]t

(2)

Cro is the retentate concentration at time zero, t is the elapsed time, and r the mean residence time. The apparent retention R can be calculated according to Eq. (2) from a plot of l n ( C r / C r o ) versus operating time. The elution loss Kelu is then defined by Eq. (3): Kelu = (1 - R ) / r

(3)

Even for rather high retention, washout becomes critical as residence times go down, since it depends on residence times as well as retention. In Fig. 4 the retention of PEG, carboxylated PEG, and the final preparation of the P E G - N A D H are represented. Experiments were performed within the membrane reactor as described below. Concentration of PEG was measured by refractive index; P E G - N A D H was determined by UV absorbance at 260 nm employing a cuvette with a 1-mm light path for the retentate and a 10-mm light path for the filtrate to compensate partly for the large difference in concentration of the different streams. It should be noted, however, that the apparent retention depends on the hydrodynamics of the system employed. Considering its importance for coenzymedependent processes the apparent retention should be reexamined if other reactor configurations are evaluated or hydrodynamic conditions are drastically changed. The absolute loss of coenzyme is also a function of the coenzyme concentration, which should therefore be as small as possible in the reactor. The lower limits of coenzyme concentration are defined by the necessity to saturate the enzymes involved in order to operate with their maximal catalytic efficiency. Table I shows that P E G - N A D H derivatives exhibit comparatively low Kmvalues, so that in many cases stationary coenzyme concentrations below 0.5 mM are sufficient for operation. Since the enzymes employed in the membrane reactor are much larger

[2]

15

CONTINUOUS ENZYMATIC TRANSFORMATION t

8.6

8.4

A D ¢

8.2 o

8

PE8-28888 PEG-28088-COOH PEG-28888-NADH

-

0

I

I

I

I

I

I

I

2

3

4

5

6

TIME

7

/ D

FIG. 4. Elution of P E G and P E G derivatives across a YM5 (Amicon) m e m b r a n e ; residence time ~" = 1 hr.

than the coenzyme derivative, elution losses of the catalyst itself are negligible. Experimental Setup of an Enzyme-Membrane Reactor The ultimate goal of an experiment determines the degree of sophistication the experimental setup of an enzyme-membrane reactor has to meet. A simple homemade version utilizing commercially available equipment is illustrated in Fig. 5. It has been successfully used for the determination of the stereospecificity of enzymes in crude extracts during a screening procedure.~4 A similar version could be employed if the preparation of small amounts (1-100 mmol) of a certain chiral compound is of prime interest. However, if the enzyme-membrane reactor is operated to evaluate coenzyme utilization and to use the data for feasibility studies or the design of a large-scale process several, safety and control features as well as precise measuring devices have to be installed. 14 W. H u m m e l , H. Schiitte, and M.-R. Kula, Ann. N . Y. Acad. Sci. 434, 194 (1984).

16

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[2]

p-

i

i . . . . . . . . . . . . . . .

FIG. 5. Simple version of a laboratory enzyme-membrane reactor. The glass vessel (1) contains the substrate solution in 0.3 M ammonium formate or sodium formate, pH 7.0-7.5. In the thermostated (25 ° water bath) reactor (2), 0.5 ~mol/ml PEG-NADH, 3 U/ml formate dehydrogenase, and 3-5 U/ml of the second dehydrogenase are placed in 10 ml substrate solution containing 0.3 M formate. The reaction mixture is circulated by means of a peristaltic pump (4) passing through an ultrafiltration device (3). A CEC unit from Amicon or a similar device fitted with a YM5 membrane (Amicon) can be conveniently employed. With a CEC unit the circulating rate is set to approximately 30 ml/hr. This will create a small pressure differential and lead to a filtrate flux of approximately 3 ml/hr and a residence time of 3 hr. Filtrate flux is controlled and readjusted manually about twice a day if necessary. Filtrate is collected in bulk or with a fraction collector. The reactor (2) is closed airtight and the liquid level kept constant by the pneumatic overhead arrangement (1). The reactor is stirred using a magnetic stirrer to ensure proper mixing.

Figure 6 gives a flow diagram of a continuous process including all measuring points. Figure 7 shows a laboratory reactor which has been assembled in the mechanics workshop of the authors. The guiding principles in the design were to achieve a high ratio of membrane surface to internal retentate volume, to incorporate temperature control, and to allow easy access to the membrane for setup and changes between experiments. Polarization control is effected by operating the membrane reactor

~

enzyme ~~c ° e no z y mI e ~ ~

~J I product

go substrate I 2 S

METERINfi PUMP STERILE FILTER ENZYME MEMBRANE REACTOR

4 5 6

PHOTOMETER POLARIMETER RECORDER

F

FIG. 6. Flow diagram of a laboratory-scale membrane reactor with magnetic stirrer and thermostating jacket.

[2]

17

CONTINUOUS ENZYMATIC TRANSFORMATION

lid

se mc se sti

bc )strate sa

FIG. 7. Details of a laboratory membrane reactor (the magnetic bar is inserted in a plastic disk to minimize the retentate volume).

on a magnetic stirrer and inserting a magnetic bar underneath the membrane. In general the speed of the stirrer is set to 200 rpm. For long-time operations special care has to be taken to achieve and maintain sterile conditions in the reactor. After complete assembly a solution containing a suitable disinfectant such as 0.1% peroxyacetic acid is pumped through the reactor, and at the same time all tubing and filters are flushed. The disinfectant is replaced after several hours, at least 5, by sterile water followed by substrate solution. At this point the integrity of all seals is also tested. Finally the selected enzymes and PEG-NADH are added aseptically or pumped into the reactor and concentrated on the retentate side. Reaction starts at zero conversion and the time course until steady state is obtained can be followed and utilized to test the kinetic model based on measurement of initial rates, which has previously been developed to describe the reactor performance. 6,1° Because chiral compounds are produced, polarimetry is the method of choice to follow conversion in the reactor effluent. A port to the retentate side, closed by a septum, serves to draw periodically samples from the reactor using sterile syringes and to measure coenzyme content and enzyme activities. If desired, on-line control of enzyme activity and coenzyme levels is possible, but this requires for each single case a detailed

18

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[2]

100

80 g,g 80

z cl H OC bJ Z 0 (J

40

SPACE-TIME-YIEL = D214 G/(LND) 100 MMOL/L KETDISOCAPROATE 400 MMOL/L NH4-FORMATE pH = 8 , 0 , T = 2S * • = 7S MIN

20

0

0

|

I

I

I

}

1

I

I

10

20

3B

48

SO

50

70

80

TIME /

90

D

FIG. 8. L-Leucine production with F D H , L e u O H , and PEG 2 0 0 0 0 - N A D H .

298 U/KG PRODUCT

i~ rl D -I i~r~

O

(3 b.

FDH I

I

I

0

20

30

I

40 TIME

/

ADDITION

1

I

I

i

50

80

70

80

D

FIG. 9. L-Leucine production and deactivation of F D H .

90

[2]

19

CONTINUOUS ENZYMATIC TRANSFORMATION

261U/KGPRODUCT1 ._1 \

I w 1

t LEUDH

o 0

-

I

ADDITION

I

I

1

I

1

I

I

I

10

20

30

40

SO

60

70

80

90

TIME / D FIG. 10. L-Leucine production and deactivation of LeuDH.

kinetic model in order to differentiate the activity of the enzymes involved. 15 The two enzymes and the coenzymes utilized in the coupled reaction will have different rates of decomposition. The component limiting the performance of the reactor can be identified and added with the substrate stream in order to restore conversion and performance. After sufficiently long operating times the consumption of enzyme activity and coenzyme loss per unit weight of product can be determined with confidence and put into an economic model of the reaction in order to optimize the process studied. For calculation and design it is important to maintain a constant flow. A pulse-free piston pump (Reichelt Chemietechnik, Heidelberg) was found satisfactory for prolonged operation with the low flow rates required for the laboratory reactor. Performance of Reactor Figures 8-11 demonstrate 3 months' performance of a laboratory enzyme-membrane reactor for the production of L-leucine. A cycle number of 80,000 was obtained for the coenzyme. The cycle number is defined 15 R. Wichmann and C. Wandrey,

Enzyme Eng. 6,

311 (1982).

20

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[2]

I

100000

O.B

00000

86

60808

0.4

48000 ~ \

0.2

20000

d

0

0 0

10

20

38

40 TZME /

50

60

70

80

90

D

FIG. 11. L-Leucine production, deactivation of PEG 20000-NADH, and development of cycle number.

here as moles of product obtained per moles of coenzyme lost in the process (for whatever reasons). Figure 11 shows that the cycle number increases from the start of the experiment and levels off after - 2 0 days. Figures 9 and 10 demonstrate that the reactor was operated at rather low stationary enzyme concentrations. Increasing the enzyme content in the reactor would give higher space-time yields. Productivities up to 1 kg/ liter-day have been realized, which is much higher than reported values for fixed bed reactors. The optimal ratio of the enzyme activities depends on the kinetic constants, and concentrations are not necessary equal. It is one of the advantages of the enzyme-membrane reactor that the concentration of the enzymes involved in the process can be chosen and maintained at will. The system is amenable for modeling based on kinetic data and can therefore be optimized on rational grounds. Enzymes isolated from different organisms can be mixed and applied together in homogeneous phase. For this reason the enzyme-membrane reactor appears to have advantages compared to the utilization of whole cells, as long as the number of single reactions taking place remains fairly small and the enzymes employed are sufficiently stable in solution during operation. Results shown in Figs. 9 and 10 indicate that these conditions can be fulfilled. The average enzyme

[3]

COIMMOBILIZATION OF

NAD

AND DEHYDROGENASES

21

activity consumption with respect to both enzymes was less than 300 U/ kg product over a period of 3 months of continuous operation. Acknowledgment The coenzyme preparation was from A. F. Biickmann, and the enzymes were isolated by H. Schiitte and K. H. Kroner. The chemical engineering measurements at continuous operation were carried out by Mrs. U. Mackfeld. The engagement of all these co-workers is gratefully acknowledged.

[3] C o i m m o b i l i z e d S y s t e m o f N A D w i t h D e h y d r o g e n a s e s B y YOSHIMITSU YAMAZAKI a n d HIDEKATSU M A E D A

Recycling of NAD is a prerequisite for the industrial application of dehydrogenases. The most effective way of realizing this is to maintain both NAD and the coupled dehydrogenases in one reactor system (bioreactor) and recycle NAD in situ. Several methods have been developed for this purpose: NAD bound to a water-soluble polymer and the dehydrogenases are either placed all together in an ultrafiltration apparatusf1-3 enclosed with a semipermeable membrane, 4,5 microencapsulated, 6 or immobilized in a collagen membrane. 7 Solid-phase coimmobilization of NAD and a dehydrogenase ~,9 and the covalent binding of an NAD derivative in or near the active site of dehydrogenases ~°,~ have also been reported.

l y . Yamazaki and H. Maeda, Agric. Biol. Chem. 50, 3213 (1986). S. Furukawa, N. Katayama, T. Iizuka, I. Urabe, and H. Okada, FEBS Lett. 121, 239 (1980). 3 R. Wichmann, C. Wandrey, A. F. Btickmann, and M.-R. Kula, Biotechnol. Bioeng. 23, 2789 (1981). 4 p. Davies and K. Mosbach, Biochim. Biophys. Acta 370, 329 (1974). 5 A. Malinauskas and J. Kulys, Anal. Chim. Acta 98, 31 (1978). 6 j. Grunwald and T. M. S. Chang, J. Appl, Biochem. 1, 104 (1979). 7 Y. Morikawa, I. Karube, and S. Suzuki, Biochim. Biophys. Acta 523, 263 (1978). 8 S. Gestrelius, M.-O. M~nsson, and K. Mosbach, Fur. J. Biochem. 57, 529 (1975). 9 M. A. Mazid and K. J. Laidler, Biotechnol. Bioeng. 24, 2087 (1982). 10 M.-O. Mhnsson, P.-O. Larsson, and K. Mosbach, Eur. J. Biochem. 86, 455 (1978). I~ M.-O. M~nsson, N. Siegbahn, and K. Mosbach, Proc. Natl. Acad. Sci. U.S.A. 80, 1487 (1983).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[3]

COIMMOBILIZATION OF

NAD

AND DEHYDROGENASES

21

activity consumption with respect to both enzymes was less than 300 U/ kg product over a period of 3 months of continuous operation. Acknowledgment The coenzyme preparation was from A. F. Biickmann, and the enzymes were isolated by H. Schiitte and K. H. Kroner. The chemical engineering measurements at continuous operation were carried out by Mrs. U. Mackfeld. The engagement of all these co-workers is gratefully acknowledged.

[3] C o i m m o b i l i z e d S y s t e m o f N A D w i t h D e h y d r o g e n a s e s B y YOSHIMITSU YAMAZAKI a n d HIDEKATSU M A E D A

Recycling of NAD is a prerequisite for the industrial application of dehydrogenases. The most effective way of realizing this is to maintain both NAD and the coupled dehydrogenases in one reactor system (bioreactor) and recycle NAD in situ. Several methods have been developed for this purpose: NAD bound to a water-soluble polymer and the dehydrogenases are either placed all together in an ultrafiltration apparatusf1-3 enclosed with a semipermeable membrane, 4,5 microencapsulated, 6 or immobilized in a collagen membrane. 7 Solid-phase coimmobilization of NAD and a dehydrogenase ~,9 and the covalent binding of an NAD derivative in or near the active site of dehydrogenases ~°,~ have also been reported.

l y . Yamazaki and H. Maeda, Agric. Biol. Chem. 50, 3213 (1986). S. Furukawa, N. Katayama, T. Iizuka, I. Urabe, and H. Okada, FEBS Lett. 121, 239 (1980). 3 R. Wichmann, C. Wandrey, A. F. Btickmann, and M.-R. Kula, Biotechnol. Bioeng. 23, 2789 (1981). 4 p. Davies and K. Mosbach, Biochim. Biophys. Acta 370, 329 (1974). 5 A. Malinauskas and J. Kulys, Anal. Chim. Acta 98, 31 (1978). 6 j. Grunwald and T. M. S. Chang, J. Appl, Biochem. 1, 104 (1979). 7 Y. Morikawa, I. Karube, and S. Suzuki, Biochim. Biophys. Acta 523, 263 (1978). 8 S. Gestrelius, M.-O. M~nsson, and K. Mosbach, Fur. J. Biochem. 57, 529 (1975). 9 M. A. Mazid and K. J. Laidler, Biotechnol. Bioeng. 24, 2087 (1982). 10 M.-O. Mhnsson, P.-O. Larsson, and K. Mosbach, Eur. J. Biochem. 86, 455 (1978). I~ M.-O. M~nsson, N. Siegbahn, and K. Mosbach, Proc. Natl. Acad. Sci. U.S.A. 80, 1487 (1983).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

22

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[3]

We report still another method in which the dehydrogenases and a polymerizable NAD derivative are coentrapped in a polyacrylamide gel. The gel is prepared by radical copolymerization, and consequently NAD is bound on the matrix which encloses the enzymes. The purpose of this article is to outline the preparation of the coimmobilized gels of NAD and dehydrogenases and to discuss their application to bioreactors for synthesis and analysis. Synthetic Procedures Preparation o f Polymerizable N A D Derivatives

The principle used is the carbodiimide-mediated coupling of N6-carboxymethyl-NAD (1) with amines containing a methacryloyl group. The amines 2 and 3 (see Fig. 1) were synthesized as described below. Muramatsu et al., ~2 Le Goffic et al., ~3 and Yamazaki et al. 14 also reported methods for synthesis of polymerizable NAD derivatives. 6-(Tritylamino)hexylamine dihydrochloride (4). ~5 To a solution containing 30.6 g of 1,6-diaminohexane and 39.2 g of triethylamine in 200 ml of isopropanol is slowly added a solution of trityl chloride (88 g in 100 ml of isopropanol plus 150 ml of chloroform). The reaction mixture is stirred at room temperature for 3 hr and then the solvents and the remaining triethylamine are removed under reduced pressure. The residual solid is mixed with benzene (1.2 liters) and water (1.2 liters) and vigorously shaken in a separatory funnel. After separation of the two phases, the aqueous layer is discarded and the organic layer is washed successively with water (1 liter) and water containing 2.5% KH2PO4 (0.7 liters x 2). Centrifugation (1800 rpm for 10 min) is necessary to promote the phase separation. As part of the product is contained in the aqueous layers, it is recovered by extraction with benzene (0.5 liters each time). All organic layers are combined and evaporated under reduced pressure. The oily ~2M. Muramatsu, 1. Urabe, Y. Yamada, and H. Okada, Eur. J. Biochem. 80, 111 (1977). 13 F. Le Goffic, S. Sicsic, and C. Vincent, Eur. J. Biochem. 108, 143 (1980). 14 y . Yamazaki, H. Maeda, A. Satoh, and K. Hiromi, J. Biochem. (Tokyo) 95, 109 (1984). In this work, stopped-flow reaction kinetics was investigated on the binding of several species of polymer-bound NAD derivatives to yeast alcohol dehydrogenase in the presence of an inhibitor pyrazole. The result showed that the binding of the polymer derivatives of NAD to the enzyme was not essentially weaker and slower than that of native NAD, but was even faster in some cases in spite of an expected large steric hindrance between the polymer derivatives and the enzyme. ~5This section is reprinted with permission from Y. Yamazaki and H. Maeda, Agric. Biol. Chem. 45, 2091 (1981).

COIMMOBILIZATION OF NAD AND DEHYDROGENASES

[3] H2N 7 ~ f ~ . N H

2 ~ 3 c'CI

.~,~'cooet*

)

~ ) 3 c _ ~ N H

~-4

> ¢, ~C-N~COOEt

CI-

3 DCC

H 3 N ~

~'-7;"°"

>

~'-

5

N

2

__ ,~c-N.'ff"'J'-/""=

23

o

~coo,.occ

6 N

N

~/A¢OH

T,.o,

+ ,c ,.~;o

%~ _"~ TosO-

)

N,~L,,./N

7

3

....--.(R 2or3 NAD-> "> N6-carboxymethyI-NAO ~ I

tk~ ~

o

'%-'%-'

~

: r-o+o-~-o-7 .,o~ I I~Ho~ o- o.

8

oH OH u

9

R

=

o

H

I0 R

OH H /

o H/

: N~7-~.~N~

0

Fro. I. Synthetic scheme of polymerizable NAD derivatives. (b3C-, Trityl group; DCC, dicyclohexylcarbodiimide; TosOH, p-toluenesulfonic acid; AcOH, acetic acid; EDC, ethyldimethylaminopropylcarbodiimide hydrochloride.

residue is dissolved in 1 liter of chloroform. This solution is divided into 4 portions and each is extracted 3 times with 0.5 liters of ice-cold 0.1 M HCI. Phase separation is promoted by centrifugation. The aqueous layers are combined, adjusted to a pH of about 13 with 2 M NaOH, and saturated with NaCI. The aqueous solutions (l.5 liters × 4) are respectively extracted with chloroform (0.5 and 0.3 liters). The chloroform layers are combined, dried over Na2SO4, and evaporated under reduced pressure. The residue is dissolved in 400 ml of ether. Into the solution is bubbled dry HCI gas at - 2 0 ° to give a white precipitate. The gas bubbling is stopped when no more precipitate appears. The precipitate is collected by suction filtration, dried in vacuum, and crystallized from ethanol-ether to give 28 g (25% yield) of 4 as prisms, dec. 207 °. 6-Methacrylamidohexylammonium chloride (2).15 Preparation 4, 27.6 g, is dissolved in a mixture of 40 ml of dimethylformamide, 150 ml of chloroform, and 13 ml of triethylamine. To the ice-cold solution is slowly added a chloroform solution (30 ml) containing 22.4 g of methacrylic acid and a small amount of hydroquinone, and then another chloroform solution (40 ml) containing 28.2 g of dicyclohexylcarbodiimide is added under

24

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[3]

stirring. The reaction mixture is stirred at 4 ° for 15 min and then at room temperature for 1 hr. The formed dicyclohexylurea is filtered off and the filtrate is concentrated to about 50 ml in an evaporator at 40°. To the concentrate is added 800 ml of benzene. The insoluble materials are filtered off and the filtrate is washed successively with alkaline water (pH 13 with NaOH, 300 ml × 2), water (200 ml), and 2.5% KH2PO4 (200 ml x 2). After drying over Na2SO4, the benzene layer is evaporated under reduced pressure. The oily residue, N-[6-(tritylamino)hexyl]methacrylamide, shows a substantially single spot in TLC [Rf = 0.66 on silica gel F254 developed with benzene/methanol (9 : 1), detected under UV light and by the ninhydrin spray method], but crystallization does not occur. So, the total amount of the residue is dissolved in a mixture of acetic acid (70 ml) and water (30 ml) to remove the trityl group. The solution is maintained at 100° for 10 min and then cooled. To the ice-cold solution is added I00 ml of water. The precipitate of triphenylcarbinol is filtered off. A small amount of hydroquinone is added to the filtrate, which is concentrated to about 10 ml in an evaporator at 30°. The addition of water (100 ml), filtration, and concentration are performed again. Then, 100 ml of toluene is added to the concentrate to remove the remaining acetic acid by azeotropic distillation. After concentration under reduced pressure, 70 ml of 1 M HC1 is added to the oily residue. This mixture is concentrated in an evaporator and then the remaining water and acetic acid are removed under reduced pressure as azeotropic mixtures with ethanol and toluene, respectively. The final residue is crystallized from ethanol-ether to give 10 g (69% yield) of 2 as needles, mp 113-114 °. Ethyl 2-(tritylamino)ethanoate (5). Commercially available glycine ethyl ester hydrochloride, 15 g, is dissolved in 160 ml of chloroform. To this solution was added 24 g of triethylamine and 32 g of trityl chloride. The mixture is stirred at room temperature for 6 hr and then 160 ml of benzene is added. The insoluble salt is filtered off and the filtrate concentrated to give a syrupy residue, which is crystallized from methanol. Thus 36 g (98% yield) of 5 is obtained as prisms, mp 116°. N-(3-Amino-2-hydroxypropyl)-2-(tritylamino)acetamide (6). 15 A solution containing 12.5 g of 5, 70 g of 1,3-diamino-2-propanol (Tokyo Chemical Industries Ltd., Tokyo), and 20 g of ethylene glycol in 40 ml of isopropanol is refluxed for 4 hr. The solvents and the unreacted diamine are removed in vacuum, and the resultant glassy residue is dissolved in 200 ml of chloroform. After washing with water (200 ml), the solution is dried over NaESO4 and concentrated under reduced pressure. The residue is finally precipitated from ethanol-ether-hexane to give 6 as a crystalline solid (13.5 g, 96% yield, dec. 138-141°).

[3]

COIMMOBILIZAT1ON OF NAD AND DEHYDROGENASES

25

N - [ 2 - Hydroxy - 3 - [ 2 - ( tritylamino ) acetamido ] propyl ] methacrylamide (7). 15 Preparation 6, 14.3 g, is treated with methacrylic acid and dicyclohexylcarbodiimide in a similar manner to that of 4, but dichloromethane is used instead of chloroform and triethylamine is not used. The reaction mixture is similarly worked up. Then, the washed, benzene layer is divided into 4 portions. Each portion (200 ml) is extracted three times with 0.05 M HCI (ice-cold solutions of 800,500, and 500 ml). The aqueous layers are combined, adjusted to a pH of about 13 with 2 M NaOH, saturated with NaCl, and extracted three times with chloroform [600,400, and 200 ml for each combined solution (1.8 liters)]. All the chloroform layers (1.2 liters x 4) are combined, dried over Na2SO4, and evaporated under reduced pressure. The viscous residue is crystallized from chloroform-benzene to give 7 as fine needles (11.14 g, 66% yield, dec. 155157°). [N-(2-Hydroxy-3-methacrylamidopropyl)carbamoyl ]methylammonium p-toluene sulfonate (3). 15 The trityl group of 7 (11.1 g) is removed with acetic acid in the same way as described for 2. In the purification stage, p-toluenesulfonic acid (4.62 g, monohydrate, dissolved in 10 ml of ethanol) is used instead of HCI since the corresponding hydrochloride is very hygroscopic. The product is precipitated from ethanol-ether as a crystalline solid (7.7 g, 82% yield, dec. 110-115°). N6-[ N-[ N - ( 2 - H y d r o x y - 3 - m e t h a c r y l a m i d o p r o p y l ) c a r b a m o y l m e t h y l ] c a r b a m o y l m e t h y l ] - N A D (9) and N6-[N-(6-methacrylamidohexyl)carba m o y l m e t h y l ] - N A D (10). 16 N6-Carboxymethyl-NAD (1) is prepared according to the method of Mosbach et al. 17 Then 1.2 g o f l , 2.7 g of 3, and 50 mg of hydroquinone are dissolved in 14 ml of water and the pH is adjusted to 4.8 with 2 M LiOH. To the solution is added 0.53 g of ethyldimethylaminopropylcarbodiimide hydrochloride (8, Fluka) with continuous stirring at room temperature. The pH is maintained at 4.8 with 2 M LiOH and 1 M HC1. After l hr, another portion of 8 (0.28 g) is added and the stirring is continued. After completion of the coupling [within 3 hr from the initial addition of 8; monitored by TLC on PEI-ceUulose F (Merck) developed with 0.1 M LiCI], the reaction mixture is applied to a gel permeation column (BioGel P-2, 2 cm x 1.5 m; elution with water). The nucleotides are eluted in effluent 130-220 ml separately from other compounds (the unreacted amines, etc.). The main fractions are pooled and concentrated to about 10 ml in an evaporator at 30°. The concentrate is applied to a Dowex l-X2 column (1.5 x 50 cm, HCOO- form) and the elution is performed by a linear gradient made of 500 ml of 0.2 M formic i6 y . Yamazaki and H. Maeda, Agric. Biol. Chem. 45, 2277 (1981). 17 K. Mosbach, P.-O. Larsson, and C. Lowe, this series, Vol. 44, p. 859.

26

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[3]

acid and 500 ml of 2 M formic acid. Fractions of the main peak (effluent 480-600 ml) are combined and concentrated in an evaporator at 30°. An excess amount of chilled acetone is added to the concentrate. The white precipitate is collected by centrifugation, washed with acetone, and dried in vacuum to give 1.05 g of 9 as white powder; 68% yield from 1, based on UV absorption measurement ('~266 = 23.0 × 103 M 1 c m - 1 ) 16 Similarly, 10 is prepared by the coupling of 1 with 2 in 27% yield. These NAD derivatives show a single spot in various TLC systems with the following Rf values: 0.53 (9), 0.74 (10), and 0.46 (NAD) on cellulose F (Merck) developed with isobutyric acid/1 M aqueous NH3 (5/3, v/v) saturated with EDTA; 0.08 (9), 0.03 (10), and 0.16 (NAD) on cellulose F developed with 0.1 M phosphate buffer (pH 6.8)/ammonium sulfate/n-propanol (100/60/2, v/w/v); and 0.57 (9), 0.57 (10), and 0.28 (NAD) on PEI-cellulose F developed with 20 mM LiC1. The spots are located by viewing under UV light (254 nm).

Preparation of Coimmobilized Gel of NAD and Dehydrogenases Enzymes. Yeast formate dehydrogenase (FDH; lyophilized powder), horse liver alcohol dehydrogenase (LADH; suspension in 20 mM phosphate buffer, pH 7.0, containing 10% ethanol), and pig heart diaphorase (dihydrolipoamide dehydrogenase; suspension in 3.2 M ammonium sulfate) were purchased from Boehringer-Mannheim. Thermus thermophilus malate dehydrogenase (MDH; solution in 50% glycerol) was given to us by Dr. N. Miwa of Mitsubishi Petrochemical Co., Ltd. (Ibaraki), whom we wish to thank for his generosity. Standard Assay Procedures. 18 A Hitachi 124 spectrophotometer with a thermostated cell holder is used to follow the absorbance change at 340 nm in the dehydrogenase assays. The following substrate solutions are used: LADH--0.56 M ethanol, 2 mM NAD, 75 mM semicarbazide-HC1, 9.8 mM glutathione, and 62 mM sodium pyrophosphate-18 mM glycine buffer (pH 8.7) in a total volume of 3.00 ml; FDH--35 mM sodium formate, 1 mM NAD, and 87 mM potassium phosphate buffer (pH 7.5) in a total volume of 3.00 ml; MDH--0.52 mM potassium oxaloacetate, 0.2 mM NADH, 98 mM potassium phosphate buffer (pH 7.5) in a total volume of 2.98 ml; and diaphorase--5 mM sodium olAipoate, 0.2 mM NADH, 0.3 mM NAD, 0.94 mM EDTA, 0.69 mg/ml bovine serum albumin (Miles), and 80 mM potassium phosphate buffer (pH 5.9) in a total volume of 2.88 ml. The reaction is started by addition of the enzyme solution (20 p.l) and is This section is reprinted with permission from Y. Yamazaki and H. Maeda, Agric. Biol. Chem. 46, 1571 (1982).

[3]

COIMMOB|LIZATION OF

NAD

AND DEHYDROGENASES

27

the temperature is maintained at 30° in all cases. Proteins are determined by Lowry's method with bovine serum albumin as standard. ~9 F D H - M D H - N A D Gel. 2° The gel containing FDH, MDH, and NAD is prepared as follows: 15 mg (14.7/xmol) of the polymerizable NAD derivative 9 is dissolved in 0.15 ml of 0.05 M sodium pyrophosphate buffer (pH 8.5) and neutralized with 0.1 M NaOH. To the solution are added 0.69 g of acrylamide (recrystallized from benzene), 36.3 mg of methylenebisacrylamide, and 1.1 ml of the pyrophosphate buffer. Nitrogen gas is bubbled into the solution for 5 min under cooling with ice. Then, the following solutions are added successively to the above solution: 0.25 ml of MDH solution (in 50% glycerol; containing 342 U and 3.08 mg protein), 0.3 ml of FDH solution (in 0.1 M phosphate buffer, pH 7.5; containing 7.02 U and 2.93 mg protein), and two 0.15-ml solutions in the pyrophosphate buffer containing 4.5 mg of K2S208 and 11.25 mg of dimethylaminopropionitrile. The final solution is removed from the ice bath and N2 gas is bubbled into it at room temperature for about 5 min until viscosity increases. Then the solution is kept in the ice bath for 20 min. The formed gel is crushed in a blender with 50 ml of ice-cold 0.1 M phosphate buffer, pH 7.5, and thoroughly washed with the buffer on a Btichner funnel. The total weight of the gel (in a wet state) was 3.51 g. The particle size (average of the major and minor diameters) of the crushed gel was in the range of about 0.01-0.8 mm. The protein content, enzyme activity, and NAD content of the gel and their immobilization yield or recovery were as follows: protein content, 1.38 mg/g (81%); FDH, 0.16 U/g (8%); MDH, 31 U/g (32%); NAD, 3.6 /zmol/g (86%). Assay procedures for the gel are described below. L A D H - d i a p h o r a s e - N A D Gel. ~8 This gel is prepared as described above, but using the following amounts of reagents: 60 mg (56.4/zmol) of the polymerizable NAD derivative 10 in 0.2 ml of the pyrophosphate buffer (pH 8.5), 1.2 g of acrylamide, 66 mg of methylenebisacrylamide, 6 mg of bovine serum albumin, 1.5 ml of the buffer, 0.6 ml of LADH suspension (20.1 U and 6.60 mg protein), 0.6 ml of diaphorase suspension (137 U and 6.42 mg protein), 3 mg K2S208 in 0.15 ml of the buffer, and 7.5 mg dimethylaminopropionitrile in 0.15 ml of the buffer. This gel showed the following properties: protein, 2.43 mg/g (59% recovery); LADH, 0.23 U/g (5%); diaphorase, 3.7 U/g (12%); NAD, 7.5/~mol/g (61%). 19 O. H. Lowry, N. J. Rosebrough, A. L. Farr, and R. J. Randall, J. Biol. Chem. 193, 265 (1951). 2o This section is reprinted with permission from Y. Yamazaki, H. Maeda, and K. Kamibayashi, Biotechnol. Bioeng. 24, 1915 (1982).

28

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[3]

Estimation o f Immobilized Enzyme Activity, Protein, and N A D 18 The activities of the immobilized dehydrogenases were assayed by the flow method using a substrate solution of the same composition as in the standard assay procedure. The apparatus consisted of a test tube held in a water bath (30°) and a flow system with a flow cuvette in the spectrophotometer and a peristaltic pump. A net was attached at the inlet of the flow system to filter the gels. At first, the flow system was filled with the substrate solution (total volume of 1.3 ml) and then another substrate solution (three times the volume in the standard procedure) was added to the test tube. After equilibration for 3 min at 30°, the reaction was started by addition of the gel to the test tube with a constant circulation of the solution through the whole system at a flow rate of 2.3 ml/min. The reaction mixture in the test tube was also stirred by a magnetic stirrer. The absorbance change at 340 nm was continuously monitored. The experiment was repeated with different amounts of the gel and the free enzyme. The reaction rate (A340/min) was proportional to the number of milligrams of the wet gel as well as the number of units of the free enzyme determined by the standard method. From these data the enzyme activity (U/g wet gel) was calculated. The immobilized proteins were estimated by amino acid analysis (with a Hitachi 835 amino acid analyzer with an autosampler) after acid hydrolysis. A sample of 100-400 mg of the gel was mixed with 15 ml of 6 M HC1. The mixture (in a sealed vial) was kept at 120° for 30 hr and then the insoluble materials were removed by filtration. A 13-ml portion of the filtrate was evaporated to dryness at 60°. The residue was redissolved in 1.5 ml of 0.06 M HC1 and applied to the analyzer. The protein content (mg/g wet gel) was calculated from the combined peak area of 11 species of amino acids (Asp, Thr, Set, Glu, Pro, Ala, Val, Ile, Leu, Phe, and Lys; other amino acids had decomposed or overlapped with unknown compounds on the chromatograms) using the corresponding calibration curves which were prepared with the hydrolyzates of the free proteins (FDH + MDH and LADH + diaphorase + albumin, mixed in the same ratios as in the immobilization). The combined peak area was proportional to the used amount of the gels as well as that of the free proteins. The coexistence of polyacrylamide gels had no effect on this protein analysis as shown by a comparative experiment with albumin plus protein-free gels and albumin only. The NAD content was estimated from phosphate determinationsz~ after exhaustive washing of the gel with 1 M NaCl. 21 G. R. Bartlett, J. Biol. Chem. 234, 466 (1959).

[3]

COIMMOBILIZATION OF NAD AND DEHYDROGENASES

29

Applications

Continuous Production of L-Malate by the Coimmobilized F D H - M D H - N A D Gel 2° The principle is the coupled reaction: formate

matrix-bound FDH /

CO2

~ matrix-bound ) NADH

t.-malate MDH ~ oxaloacetate

The gel (1.4 g) was packed in a column tube and a solution containing 0.2 M sodium formate and 2 mM sodium oxaloacetate in 0.1 M phosphate buffer (pH 7.5) was passed upward through the column at a flow rate of 2.4 ml/hr. The temperature was kept at 30°. The substrate solution was saturated with chloroform before use for preventing bacterial growth in either the solution or the column. It was further kept in an ice bath and changed to a freshly prepared one, usually every 12 hr. At the points indicated by arrows A and E in Fig. 2, the substrate solution was changed to a solution lacking only formate. It was again changed to the original substrate solution as indicated by arrows B and F. A similar experiment was run for removal and addition of oxaloacetate at the points C and D. The eluent from the column was divided into fractions of 1-hr duration and then held at 100° for 5 min to decompose the unreacted oxaloacetate. The L-malate in the fractions was determined by the method using hydrazine, NAD, and pig heart malate dehydrogenase (Boehringer-Mannheim). 22 The result is shown in Fig. 2. The L-malate production by the F D H - M D H - N A D gel was continued for 3 weeks. When formate was removed from the substrate solution (as indicated by arrow A), the L-malate production ceased. It was, however, restored to the original level by the addition of formate to the solution. This result clearly shows that L-malate was produced with recycling of the matrix-bound NAD. The L-malate production was, of course, stopped by the removal of oxaloacetate from the substrate solution (as indicated by arrow C) and resumed by the addition of the substrate. The concentration of the produced L-malate gradually decreased with time, but even at the final stage of the operation about 60% the initial value was observed and 22O. H. Lowry and J. V. Passonneau, in "A Flexible System of Enzymatic Analysis," p. 203. Academic Press, New York, 1972.

30

MULTISTEP ENZYME SYSTEMS AND COENZYMES

3E E

[3]

LO

O

0

J

I I00

~ i

L

I

r

,

3

,

~

5

-~'B , , 7

200 .--

400

300

_ 1 ~ * _ F r o c t l o n NO. h (.,,/U, , , , 9 II 13

,

,

15

L

, 17

500 ,E ~" 19

'~,'F ,

i 21

Days

FI6.2. Continuous production of L-malate in a column reactor packed with the F D H M D H - N A D gel. 2° The arrows indicate the removal of formate (A and E) and oxaloacetate (C) from the substrate solution and the addition of the removed substrate to the solutions (B, D, and F).

the recycling of the matrix-bound NAD was again confirmed by the second experiment concerning the removal and addition of formate at the points E and F. The total amount of L-malate produced during the operation was about 1300/zmol. Therefore the turnover number for each NAD molecule was 260 cycles, assuming that the total number of bound NAD molecules were available to the enzymes. The gel was recovered from the column and the contents of protein, the enzymes, and NAD were determined. The data and the residual percentages were as follows: protein, 0.77 mg/g (56%); FDH, 0.16 U/g (100%); MDH, 1.4 U/g (5%); NAD, 3.6 /xmol/g (100%). The value of NAD was obtained from phosphate analysis and the residual percentage of enzymatically active form could not be determined. In spite of this ambiguity, it can be concluded that long-term production of L-malate was achieved by the stable supply of the matrix-bound NADH. This is because the rate-limiting step of the coupled reaction is considered to be the reaction of FDH since the content of FDH was remarkably less than that of MDH, and moreover, no loss of FDH was found for the recovered gel. Therefore, the gradual decrease in the L-malate production by the F D H M D H - N A D gel must be due to the loss of MDH. This enzyme is very

[3]

COIMMOBILIZATION OF NAD AND DEHYDROGENASES

31

Sample solution

r~x ,.~,~jlnjector I

~

~

.~-" ~._r~ '~ J

.~

~

Gels of coimmobilized

/

[~I I

~a[~]l I

LADH-Diaphorase-NAD spectrophotometel E x 560nm

i~ ~i;_~ ~'~ Em580nm I~---~ ;--'L~L____~ FlOwcell ~,

(+~-Lipoatesolution j / ~

Resazur n so ut on

FIG. 3. Illustration of the ethanol analyzer using the coimmobilized LADH-diaphoraseNAD gel. ~sThe solutions in the two reservoirs are mixed at an equal flow rate by a peristaltic pump and sent to the column via the injector (Rheodyne, Model 7125). The column is packed with the L A D H - d i a p h o r a s e - N A D gel (5 x 110 mm). The eluent is led to the flow cell (14 p.l) in the fluorescence spectrophotometer (Hitachi 204), which is operated at an excitation wavelength of 560 nm and an emission wavelength of 580 nm. When a solution containing ethanol is added by the injector, resorufin is formed in the column by the coupled reaction of LADH and diaphorase, and the amount thereof (detected as fluorescence intensity) is continuously monitored.

stable as it maintains the whole activity in the incubation at 70° for 3 hr. 23 The reduced protein content of the recovered gel suggests a leakage of MDH from the gel. This was supported by the following experiment. A gel containing only MDH was similarly prepared, packed in a column, and continuously washed with the above substrate solution for 3 weeks in the same way as in the operation of the reactor column. The residual percentages of enzyme activity and protein content of the recovered gel were 4 and 15%, respectively. It was therefore concluded that MDH was removed with time from the reactor column by leakage, but that since its content was large compared with the content of FDH, the loss of MDH was not sharply reflected in the rate of L-malate production.

Ethanol Analyzer Using the Coimmobilized LADH-Diaphorase-NAD Gel 18 The L A D H - d i a p h o r a s e - N A D gel was applied to an ethanol analyzer composed of a column reactor and a fluorescence spectrophotometer (Fig. 3). The system is based on the enzymatic oxidation of NADH with 23 N. Miwa, personal communication (1981).

32

MULTISTEP ENZYMESYSTEMSAND COENZYMES

[3]

"G

i

_=

\

B

1T

[ EtOH] I0 (rnM)

25

50

I0

25

50

I0

25

5O

FIG. 4. Reproducibility of the ethanol analyzer/8 To the system shown in Fig. 3 were applied 0.12-ml solutions of ethanol in 0.05 M phosphate buffer (pH 7.5). The ethanol concentrations are given in the figure. This figure shows the trace of the fluorescence intensity. The operational conditions were as follows: buffer, the same as above; DLlipoate, 5 mM in the reservoir; resazurin, 32 tzM in another reservoir; temperature, 25°; flow rate, 0.95 ml/min; column size, 5 x 110 mm; sensitivity of the photometer, 6 x 10; recorder range, 5 mV. Injection points are indicated by arrows. resazurin, 24 but practically, lipoate (a natural substrate of diaphorase) was

also used to promote the oxidation. The response of this analyzer (the ethanol ~

/matriN-bDUnd ~x~

LADH

/

~N/~"~/"O~o~~

O

~hor~ (fluorescent)

acetaldehyde

0~.~.~

matrix-bound NADH

0 ....~jO

-

N O resazurin sum of fluorescences represented by peak areas) was reproducible with regard to the applied amount of ethanol as shown in Fig. 4. The response increased with increasing ethanol concentration but seemed to saturate at high concentration. The fluorescence intensity reached maximum at 4.5 min after the sample injection at a flow rate of 0.95 ml/min and then gradually returned to its original level. The time required for the return and the response intensity were affected by many factors: pH, temperature, concentrations of lipoate and resazurin, column size, and flow rate. From an investigation on these factors, :8 the optimal conditions were chosen as given in the legend to Fig. 4. The long-term use of the ethanol 24G. G. Guilbault and D. N. Kramer, Anal.

Chem.

37, 1219 (1965).

[3]

COIMMOBILIZATION OF N A D AND DEHYDROGENASES

33

analyzer was tested in the same way as in Fig. 4, but the temperature was maintained at 15° during the operation. Between 9 and 12 samples were applied each day, and then the column was washed with the lipoate solution only and stored in a refrigerator until the next operation. Thus, this analyzer was used 84 times in 1 week for a total operation time of 41 hr. The responses gradually decreased with time, but about 50% of the initial value was still obtained at the final stage. The residual contents (percentage relative to the initial value) of the enzymes and NAD in the recovered gel were as follows: protein, 87%; LADH activity, 41%; diaphorase activity, 21%; and NAD, 100% as phosphate recovery. In a stability experiment it was found that both the free enzymes maintained 100% of the initial activity for 43 hr at 15° in a solution of the same composition as in the analyzer column. The decrease of the analyzer response is due to the loss of the enzymes and it must have been caused by leakage from the gel. Repeated use of NAD as an analytical reagent has also been studied by Davies and Mosbach 4 and Malinauskas and Kulys 5 based on a potentiometric or amperometric procedure. Their systems use a semipermeable membrane to retain a water-soluble polymer-bound NAD along with enzymes. Concluding Remarks As this article shows, the entrapment of dehydrogenases in polyacrylamide gels prepared with the usual monomers plus a radically polymerizable NAD derivative is a useful technique for the immobilization of coupled enzyme systems including the coenzyme. The advantages of the present coimmobilization method are (1) the simplicity, in that NAD and the enzymes are simultaneously immobilized in a single step, and (2) the applicability of the resultant gel to simple column-type reactors, without any need for an additional material such as ultrafiltration membrane. The principal disadvantage is that the reaction mixture cannot be supplemented with the enzyme even when it has been removed from the gel and the bound NAD still remains intact. The ultrafiltration method has an advantage in this connection. Thus, it is essential in the present method to use stable enzymes. The second important problem of this method is the leakage of enzymes from the gels. The molecular weight of Thermus MDH has been found to be relatively low (6 × 104).23 A cross-linking pretreatment to increase its molecular size or preparation of the gel at higher monomer concentrations (to reduce the pore size of the gel 25,z6) z5 C. J. O. R. Morris, in "Protides of the Biological Fluids" (H. Peeters, ed.), p. 543. Elsevier, Amsterdam, 1%6. 26 j. S. Fawcett and C. J. O. R. Morris, Sep. Sci. 1, 9 (•966).

34

MULTISTEP

ENZYME

SYSTEMS AND COENZYMES

[4]

would be effective to prevent the leakage. In contrast, FDH was well retained in the gel even though this enzyme is also not such a large protein (MW = 7.4 x I04, information from the maker). These facts suggest that the degree of leakage of proteins from the gel is dependent not only on the molecular size of the proteins but also on other factors, for instance, their shape and affinity to the matrix. It may be possible to increase the effect on enzyme retention of the affinity factor by selecting more suitable gels and/or chemically modifying the surface of the proteins. Since the present work has shown that long-term recycling of matrix-bound NAD by the simultaneously immobilized enzymes is possible, this method can be expected to find wide use once the above problem of the leakage has been solved.

[4] P o l y m e r i z a b l e N A D D e r i v a t i v e a n d M o d e l E n z y m e Reactor with Recycling of Polyethylene Glycol-Bound NAD By HIROSUKE OKADA and ITARU URABE

Many enzymes require the participation of readily dissociable cofactors such as NAD for their catalytic activities. The continuous utilization of these enzymes requires the retention and regeneration of the coenzymes. For this purpose, several methods of covalently attaching NAD to water-soluble polymers have been reported. 1,2These NAD derivatives are also useful for the study of the interaction between enzymes and NAD. In this chapter, methods for preparing a polymerizable NAD derivative, N A D - N 6-[N-( N-acrylo yl- l -metho x ycarbon yl- 5-aminopentyl )propio amide], and a polyethylene glycol-bound NAD derivative (PEG-NAD) are described (Fig. 1). The polymerizable NAD is a unique derivative, which easily copolymerizes with other vinyl monomers such as acrylamide to give macromolecular NAD derivatives (polymeric NAD). 3 Furthermore, various kinds of polymeric NAD can be obtained by copolymerization with different vinyl monomers in different molar ratios: .~ From investigations into the coenzymatic properties of the polymeric NAD, it has been suggested that NAD derivatives of smaller molecular 1 K. Mosbach, P.-O. Larsson, and C, Lowe, this series, Vol. 44, p. 859. 2 C. R. Lowe, in "Topics in Enzyme and Fermentation Biotechnology" (A. Wiseman, ed.), Vol. 5, p. 13. Horwood, Chichester, England, 1981. 3 M. Muramatsu, I. Urabe, Y. Yamada, and H. Okada, Eur. J. Biochem. 80, 111 (1977). 4 S. Furukawa, Y. Sugimoto, I. Urabe, and H. Okada, Biochimie 62, 629 (1980). 5 S. Furukawa, I. Urabe, and H. Okada, Eur. J. Biochern. 114, 101 (1981).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

34

MULTISTEP

ENZYME

SYSTEMS AND COENZYMES

[4]

would be effective to prevent the leakage. In contrast, FDH was well retained in the gel even though this enzyme is also not such a large protein (MW = 7.4 x I04, information from the maker). These facts suggest that the degree of leakage of proteins from the gel is dependent not only on the molecular size of the proteins but also on other factors, for instance, their shape and affinity to the matrix. It may be possible to increase the effect on enzyme retention of the affinity factor by selecting more suitable gels and/or chemically modifying the surface of the proteins. Since the present work has shown that long-term recycling of matrix-bound NAD by the simultaneously immobilized enzymes is possible, this method can be expected to find wide use once the above problem of the leakage has been solved.

[4] P o l y m e r i z a b l e N A D D e r i v a t i v e a n d M o d e l E n z y m e Reactor with Recycling of Polyethylene Glycol-Bound NAD By HIROSUKE OKADA and ITARU URABE

Many enzymes require the participation of readily dissociable cofactors such as NAD for their catalytic activities. The continuous utilization of these enzymes requires the retention and regeneration of the coenzymes. For this purpose, several methods of covalently attaching NAD to water-soluble polymers have been reported. 1,2These NAD derivatives are also useful for the study of the interaction between enzymes and NAD. In this chapter, methods for preparing a polymerizable NAD derivative, N A D - N 6-[N-( N-acrylo yl- l -metho x ycarbon yl- 5-aminopentyl )propio amide], and a polyethylene glycol-bound NAD derivative (PEG-NAD) are described (Fig. 1). The polymerizable NAD is a unique derivative, which easily copolymerizes with other vinyl monomers such as acrylamide to give macromolecular NAD derivatives (polymeric NAD). 3 Furthermore, various kinds of polymeric NAD can be obtained by copolymerization with different vinyl monomers in different molar ratios: .~ From investigations into the coenzymatic properties of the polymeric NAD, it has been suggested that NAD derivatives of smaller molecular 1 K. Mosbach, P.-O. Larsson, and C, Lowe, this series, Vol. 44, p. 859. 2 C. R. Lowe, in "Topics in Enzyme and Fermentation Biotechnology" (A. Wiseman, ed.), Vol. 5, p. 13. Horwood, Chichester, England, 1981. 3 M. Muramatsu, I. Urabe, Y. Yamada, and H. Okada, Eur. J. Biochem. 80, 111 (1977). 4 S. Furukawa, Y. Sugimoto, I. Urabe, and H. Okada, Biochimie 62, 629 (1980). 5 S. Furukawa, I. Urabe, and H. Okada, Eur. J. Biochern. 114, 101 (1981).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

['fir]

POLYMERIZABLE N A D DERIVATIVE AND P E G - N A D

L- Lysine 1) CHz=CHCOCI 2) HCIICH3OH C.OOCH3 H2NCH(CH2)4NHCOCH=CH 2 i

F

I

;oocm

35

NAD Polyethylene glycol 1) Propiolactone I SOCt2 ~v 2) Rearrangement CI(CHzCHtO)nCHtCHzCI NHCH2CH2COOH I NH~ J. ~CONH 2 ~' N" j'~I---N ~ ~t" H2N(CH2CH20)nCH2CH2N H2 /

I,"-C2<..bo.yot,yl)-N,ol

NHCHiCHiCONHCH(CHi)4NHCOCH=CH2 L~N~t~1 N N ~ CONH2

R--P--P--I~

IPolymerizable NADI

I Copotymerization

I C.rbod,,r.,de

I

J ~ NHCHtCH2CONH(CHzCHzO)nCH2CH2N Hz L~N I ~ N iI N N ~'~cONH2 !

i

R--P--P--R

I Polyethylene glycol -bound NADI

IPotym.eric NAD] FIG. 1. Synthesis of a polymerizable NAD derivative and a polyethylene glycol-bound NAD derivative. R, Ribose; P, phosphate. size and lower NAD content in the polymer chain have higher cofactor activity. Therefore, we have prepared P E G - N A D , which has only one N A D moiety located at a terminal o f the linear, flexible, and hydrophilic chain o f polyethylene glycol. P E G - N A D has good cofactor activity and is applicable in a continuous e n z y m e reactor. 6,7 To use these macromolecular NAD derivatives in an e n z y m e reactor, it is necessary to understand the behavior o f the system in which the reactions of dehydrogenases are coupled by the recycling of the NAD derivative. Methods for a theoretical analysis and the experimental operation of a model e n z y m e reactor containing lactate and alcohol dehydrogenases and P E G - N A D are also described. 7 Preparation and Polymerization of Polymerizable NAD Derivative N6-(2-Carboxyethyl ) - N A D

N A D (4.0 g, 5.6 mmol) is dissolved in a small volume of water (about 10 ml), and the pH of the solution is adjusted to 6 with 2 M LiOH. 6 s. Furukawa, N. Katayama, T. Iizuka, 1. Urabe, and H. Okada, FEBS Lett. 121, 239 (1980). 7 N. Katayama, I. Urabe, and H. Okada, Eur. J. Biochem. 132, 403 (1983).

36

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[4l

3-Propiolactone (1 ml, 16 mmol) is added to the solution, and the reaction mixture is stirred in the dark at room temperature, its pH being kept at 6 with a pH-stat (Hiranuma, PS-11) charged with 2 M LiOH. A substantial amount of propiolactone is hydrolyzed to hydroxypropionic acid and consumes LiOH. When the rate of LiOH addition slows down, propiolactone (1 ml) is further added. The alkylation of NAD is monitored by thin-layer chromatography on silica gel (solvent system A: isobutyric acid/water/ 28% aqueous NH3 = 66/33/1.7, by volume), and about 80% of NAD is alkylated after 2 days with the addition of a total of 8 ml of propiolactone. The rate of the alkylation is almost constant within the pH range of 5-7, but the rate of the hydrolysis of propiolactone increases with the increase in pH. The reaction mixture is then acidified to pH 3 with 6 M HCI and poured slowly into 20 volumes of vigorously stirred cold acetone/ethanol (1/1). The precipitate is collected by centrifugation, washed with acetone/ ethanol (1/1), then acetone, and dried in a vacuum. The crude 1-(2-carboxyethyl)-NAD (4.8 g) is dissolved in 200 ml of 1.3% NaHCO3 (pH 8.5). The solution is deaerated with N2 gas, kept at 45 °, and Na2S204 (2 g) is added. The vigorous bubbling of N2 gas is continued for 5 min, and the ratio of the absorbances at 340/260 nm reaches a value between 4 and 5. The pH of the solution is readjusted to 8.5, the reduction is terminated by bubbling air through the solution for 10 min at room temperature, and the pH is then raised to 11 with 2 M LiOH. The solution is incubated for 2 hr at 70° with the pH kept at 11, and 1-(2carboxyethyl)-NADH is rearranged to N6-(2-carboxyethyl)-NADH. The rearrangement is monitored by the change in the ultraviolet spectrum, of which the absorption maximum shifts from 260 to 266 nm and the shoulder at 300 nm disappears. The solution is cooled to room temperature, pH is adjusted to 7.3 with 2 M HCI, and 80% acetaldehyde (2 ml) and yeast alcohol dehydrogenase (1000 U) are added. The pH of the solution is kept at 7.3 with the pH-stat charged with 2 M HCI. The enzymatic oxidation is followed spectrophotometrically until the absorbance at 340 nm almost disappears (overnight). The solution is acidified to pH 3 with 2 M HCI, concentrated at 30° under reduced pressure, and poured into 20 volumes of vigorously stirred cold acetone/ethanol (1/1). The precipitate is collected by centrifugation, washed with acetone/ethanol (1/1) and acetone, and dried. The crude N6-(2-carboxyethyl)-NAD is dissolved in water and chromatographed over a DEAE-Sephadex A-25 (CI-) column (2.5 x 30 cm) with a linear gradient of 0-250 mM LiCI. The ultraviolet spectra in water and in 1 M KCN 8 of the effluent are monitored, and the fractions contains S. P. Colowick, N. O. Kaplan, and M. M. Ciotti, J. Biol. Chem. 191, 447 (1951).

[4]

POLYMERIZABLE NAD DERIVATIVEAND PEG-NAD

37

ing the desired product (eluted at about 130 mM LiC1) are combined and concentrated. Precipitation with acetone/ethanol (1/1) as above gives N 6(2-carboxyethyl)-NAD in a yield of about 30% (1.3 g). N6-(2-Carboxyethyl)-NAD shows an ultraviolet spectrum in water with an absorption maximum at 266 nm (e = 18,600 M -j cm -1) and a nuclear magnetic resonance spectrum in 2H20 with aliphatic side-chain proton signals at 8 = 2.56 (2H, t, CH2COO) and 3.65 ppm (2H, t, N6-CH2) in addition to the NAD proton pattern. The Rr values of thin-layer chromatography on silica gel in solvent system A of NAD, 1-(2-carboxyethyl)NAD, and N6-(2-carboxyethyl)-NAD are 0.40, 0.14, and 0.37, respectively.

N~-Acryloyl-t-lysine Methyl Ester Acryloyl chloride is prepared according to the procedure of Stempel et L-Lysine monohydrochloride (27.6 g, 0.15 mol) is dissolved in hot water (70 ml), and basic cupric carbonate (15 g) is added to the solution. After addition of 4 M NaOH (37.5 ml) and 2,6-bis(tert-butyl)-p-cresol (50 mg), the solution is chilled to - 5 °, and both acryloyl chloride (13.7 g, 0.15 mol) and 4 M NaOH (37.5 ml) are added dropwise to the solution at - 5 ° over 1.5 hr with vigorous stirring. The pH of the reaction mixture is adjusted to 6 with 5 M HC1, and the precipitate formed is collected by filtration and washed with water. The precipitate is suspended in water (150 ml), dissolved by lowering the pH of the solution to 1.5 with 5 M HCI, and reprecipitated by raising the pH to 6 with Na2CO3. The precipitate is collected by filtration and washed with water to yield cupric N ~acryloyl-L-lysinate. The product is suspended in water (150 ml) and dissolved by acidification with 5 M HCI. Hydrogen sulfide is introduced into the solution, and cupric sulfide formed is filtered off. The filtrate is evaporated at 30 ° under reduced pressure. To the residue ethanol is added and evaporated to remove water. N~-Acryloyl-L-lysine hydrochloride is obtained in crystalline form. The product is converted to its methyl ester by treatment with 21% methanolic hydrogen chloride (200 ml) for 48 hr at room temperature. N~-Acryloyl-L-lysine methyl ester hydrochloride is obtained as brown syrup in a yield of about 50% (20.5 g). The nuclear magnetic resonance spectrum of the product in ZHEO shows these proton signals: 6 = 1.3-1.7 (4H, m, "y,~-CH2), 1.92 (2H, m, flCH2), 3.26 (2H, m, e-CH2), 3.77 (3H, s, C O O C H 3 ) , 4.09 (1H, t, a-CH), 5.66 (IH, dd, COC--~-~CH), 6.07-6.23 ppm (2H, m, C O C H ~ C H ) . The Rf values of thin-layer chromatography on silica gel in propanol/ethyl aceal. 9

9 G. H. Stempel, R. P. Cross, and R. P. Mariella, J. Am. Chem. Soc. 72, 2299 (1950).

38

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[4]

tate/water/28% aqueous NH3 (6/1/3/1, by volume, solvent system B) of L-lysine. HC1, cupric N~-acryloyl-L-lysinate, N~-acryloyl-L-lysine • HCI, and N~-acryloyl-L-lysine methyl ester. HC1 are 0.03, 0.82, 0.65, and 0.84, respectively. NAD-Nr-[N-(N-Acryloyl-l-methoxycarbonyl-5aminopentyl )propionamide]

N~-Acryloyl-L-lysine methyl ester hydrochloride (2.5 g, 10 mmol) is dissolved in water (10 ml), and the pH is adjusted to 4.5 with 2 M LiOH. To the solution are added N6-(2-carboxyethyl)-NAD (0.5 g, 0.7 mmol), hydroquinone (10 mg), and aqueous solution (1 ml) of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (1.34 g, 7 mmol). The reaction mixture is stirred for 20 hr at room temperature, and its pH is kept at 4.5. NAD-N6-[N-(N-Acryloyl-l-methoxycarbonyl-5-aminopentyl)pro pionamide] (polymerizable NAD derivative) is obtained by the precipitation with acetone/ethanol (1/1) and by the chromatography on DEAESephadex A-25 column as described above. The yield is about 90% (0.4 g), the overall yield from NAD being approximately 25%. The polymerizable NAD derivative shows an Rr value of 0.56 on thinlayer chromatography on silica gel in solvent system A, an ultraviolet spectrum in water with an absorption maximum at 266 nm (e = 22700 M-1 cm-1), and a nuclear magnetic resonance spectrum in 2HzO with characteristic side-chain proton signals at 8 = 1.1-1.5 (4H, m, 3,4-CH2), 1.6-1.8 (2H, m, 2-CH2), 3.11 (2H, m, 5-CH2), and 3.62 ppm (3H, s, COOCH3) for the 1-methoxycarbonyl-5-aminopentyl group, 8 = 2.75 ppm (2H, m, CH2CO) for the propionamide group, and 8 = 5.63 (1H, dd, C O C ~ C H ) and 6.0-6.2 ppm (4H, m, C O C H = C H combined with two anomeric protons of the ribose) for the vinyl group. The NAD derivative has high cofactor activities (70-90%) relative to native NAD for alcohol (horse liver and yeast), lactate (rabbit muscle), and malate (pig heart) dehydrogenases. Polymeric NAD Derivative

Copolymerization is carried out in an aqueous reaction mixture containing 1 mM NAD-N6-[N-(N-acryloyl-l-methoxycarbonyl-5-aminopen tyl)propionamide], 300 mM acrylamide, 0.8% 3-(dimethylamino)propionitrile (pH 7.5), and 0.2% potassium persulfate. N2 gas is bubbled through the solution for 3 min, and the solution is kept in an ice bath for 2 hr. The reaction mixture is then dialyzed thoroughly against water, and a solution of polymeric NAD derivative is obtained in a yield of more than 90%. The polymeric NAD derivative shows an identical ultraviolet spectrum to the polymerizable NAD derivative, and is reduced completely

[4]

POLYMERIZABLE N A D DERIVATIVE AND P E G - N A D

39

with yeast alcohol dehydrogenase. The relative molecular mass is about 5 x l05. The relative cofactor activities (NAD -- 100) for alcohol and malate dehydrogenases are 15-35%, but for lactate dehydrogenase, the NAD derivative is a competitive inhibitor with respect to NAD. The conditions of the polymerization described above are very flexible. For example, the molar ratio of the polymerizable NAD derivative and acrylamide can freely be changed to give polymeric NAD derivatives with different NAD densities. The comonomer, acrylamide, can also be replaced with other compounds having an acryloyl group. It should be pointed out that the polymeric NAD derivatives thus obtained are heterogeneous mixtures of various polymer chains especially with respect to their molecular size. The cofactor activity of an NAD moiety on a polymer chain may depend on its position in the polymer chain and on the size of the polymer. These heterogeneities should be borne in mind when interpreting the properties of macromolecular NAD derivatives. Preparation of Polyethylene Glycol-Bound NAD(H)

a,co-Dichloropoly(ethylene glycol) Polyethylene glycol (Mr 3000-3700, 30 g) is dissolved in CHCI3 (55 ml), and dry pyridine (0.2 ml) is added. Thionyl chloride (30 ml) is added dropwise to the solution at 30° with stirring. After refluxing for 8 hr, the reaction mixture is chilled in an ice bath, and water (100 ml) is added to decompose excess thionyl chloride. The water layer is extracted with CH2C12, and the organic layers are combined and dried over anhydrous Na2SO4, and the solvent is evaporated. The product is precipitated from the residue ( - l 0 ml) by addition of ether (200 ml), harvested by filtration, and dried. Yield: 29 g; content of Cl: 2.6%.

a,to-Diaminopoly(ethylene glycol) a,to-Dichloropoly(ethylene glycol) (30 g) is converted into a,codiaminopoly(ethylene glycol) by heating in ethanol saturated with ammonia as described by Biickmann et al.l° After evaporation, ether is added and the precipitate is harvested by filtration and dried. The precipitate is dissolved in 5% NaHCO3 (-50 ml), extracted with CH2C12,concentrated, and precipitated with ether as described above. The crude c~,co-diaminopoly(ethylene glycol) is dissolved in water and chromatographed over a SP-Sephadex C-25 (H +) column (4 x 80 cm) with a linear gradient of 0-40 mM NaCl. The concentration of amino group in the effluent is measured 10 A. F. Biickmann, M. Morr, and G. Johansson, Makromol. Chem. 182, 1379 (1981).

40

M U L T I S T E P E N Z Y M E SYSTEMS AND COENZYMES

[4]

using 2,4,6-trinitrobenzenesulfonic acid," and the fractions containing the desired product are combined, concentrated, and extracted with CH2CI2. The organic layer is dried over anhydrous NaESO4, evaporated to remove the solvent, and tx,o-diaminopoly(ethylene glycol) is obtained in a yield of about 70% (22 g).

Polyethylene Glycol-Bound NAD a,to-Diaminopoly(ethylene glycol) (20 g) is dissolved in water (30 ml), the pH is adjusted to 4.5 with 2 M HCI, and N6-(2-carboxyethyl)-NAD (1 g) and l-ethyl-3-(3-dimethylaminopropyl)carbodiimide.HCI (4.6 g) are added to the solution. The reaction mixture is stirred for 24 hr at room temperature with the pH kept at 4.5, dialyzed against water by use of a Spectrapor 6 dialysis membrane (Mr cutoff 1000, Spectrum Medical Industries Inc.), and passed through a SP-Sephadex C-25 (H +) column (4 × 60 cm), a DEAE-Sephadex A-25 (C1-) column (1.5 × 20 cm), and then the SP-Sephadex C-25 column. The eluted fractions are combined, concentrated, and lyophilized to yield polyethylene glycol-bound NAD (PEGNAD, 2.4 g). The preparation of PEG-NAD contains about 0.2 g of a,todiaminopoly(ethylene glycol), which is removed by repeated passage through the SP-Sephadex C-25 column. P E G - N A D shows an ultraviolet spectrum in water with an absorption maximum at 266 nm (e = 23,500 M 1 cm-l), and good cofactor activity for various kinds of dehydrogenases.

Polyethylene Glycol-Bound NADH P E G - N A D is reduced to PEG-NADH with yeast alcohol dehydrogenase in a reaction mixture (10 ml) containing 50 mM Tris-HCl buffer (pH 9.0), 100 mM ethanol, 50 mM semicarbazide • HC1, 5 mM PEG-NAD, and the enzyme (200 U). After 3 hr at 37° the reaction mixture is heated in a boiling-water bath for 2 min to inactivate the enzyme, and dialyzed against 5 mM NazCO3 (pH 10.0) by using Spectrapor 6 dialysis membrane. Steady-State Analysis of Coupled Two-Enzyme Reactor

Theoretical Analysis Consider the following two reactions catalyzed by dehydrogenases, E1 and E2: El " PI + C,,~

(1)

S., + Co~ - " p,_ + C~

(2)

Si + Crd E,

11 R. F i e l d s , Biochem. J. 124, 581 (1971).

[4]

POLYMERIZABLE N A D DERIVATIVE AND P E G - N A D

41

where Crd and Cox are the reduced and oxidized forms of an NAD derivative, respectively. When the two reactions are coupled by the recycling of the coenzyme, the overall reaction is SI + $2 -

El, E, -" Pi + P_,

(3)

In a continuous reaction system, the enzymes and the coenzyme are kept in a reactor, the substrates (S~ and $2) are fed continuously at a fixed flow rate, and the products (P~ and P2) and nonconverted substrates leave the reactor at the same flow rate. The material balances for the reaction system at steady state can be written as follows if it is assumed that the reaction mixture in the reactor is completely mixed. v=

v= [Sl]0 = [$2]0 = [Pt] =

-vl

= vz

[Pl]/r [S1] + [P1] [$2] + [P2] [P2]

(4) (5) (6) (7) (8)

where v is the rate of P1 production, v~ and v2 are the rates of reduction of the coenzyme by E~ and Ez, respectively, r is mean residence time, and [ ]0 and [ ] show the concentration in the eluent and in the reactor, respectively. The reaction rates of the enzymes (vl and v2) are expressed by the following equations: Vf,l Vra[E1]([Cox][Pl] - [Crd][Sl]/Keq,l) Ol = Or,lKicox,lKpl + Vr,lKpl[Cox] + Vr,zKcox,l[P1]

(9) + Vr, l[Cox][P1] + Vf, l(gcrd,l[S1] + g s l [ C r d ]

+ [Crd][S~])/Keq,~ + Vf,~Kcrd,~[Cox][S~]/Ko.~ Kicox,~ + Vr, l Kcox,~[Crd][P~]/Ki.d,~ Vf,2 Vr,2[E2]([Cox][S2] - [Crd][Pz]/Keq,2) V2 = Vr,2Kicox,2Ks2 + Vr,2Ksz[Cox] + Vr,2Kcox,2[S2]

(10) + Vr,2[Cox][S2] + Vf,2(Kcrd,2[P2] + Kp2[Crd] + [Crd][P2])/Keq,2 + Vf,2Kcrd,2[Cox][P2]/Keq,2Kicox,Z

+ Vr,2Kcox,2[Crd][S2]/Kicrd,2 where Vf,l, Vr,1, Vf,2, and Vr,2 are maximal reaction velocities (V) of the reduction (f) and oxidation (r) of the coenzyme by the enzymes (El and E2); K~oxa, K~rd,l, K~ox,2, and Kerd,2 are the limiting Michaelis constants for the coenzymes; Kp~, Ks~, Kp2, and Ks2 are the limiting Michaelis constants for the substrates; Ki~ox,l, Ki~rd,~, Kicox,2, and Kicrd,Z are the dissociation constants for the enzyme-coenzyme complex. These equations are de-

42

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[4]

rived from an ordered Bi-Bi mechanism with negligible steady-state concentration of ternary complex (Theorell-Chance mechanism). ~2 The kinetic constants are obtained from initial rate studies of each enzyme reaction. Keq,~ and Keq,2 are the equilibrium constants of these enzyme reactions and are calculated by the Haldane equations:

Vf, f gicrd,l Ksl

(11)

Keq,l = Vr, l Kicox,l Kpl

Vf'2 Kicrd'2gP2 Ke,,2 =

(12)

Vr,2Kicox,2Ks2

The yield of P~ from Sl, Y, the total enzyme concentration, [E]t, the mole fraction of El, Ef, the total coenzyme concentration, [C]t, and the mole fraction of Crd, Cf, are defined as follows, by assuming that the fraction of the coenzyme bound to the enzymes is negligible: Y= [E]t-Ef = [C]t = Cf--

[PI]/[S1]0 × 100 [El] + [EL] [E1]/[E]t [Cox] + [Cr0] [Crd]/[C]t

(13) (14) (15) (16) (17)

At one fixed set of values of [E]t, El, [C]t, [Sl]0, [$2]0, and Y, values of Cf, v, and ~" are calculated from Eqs. (4)-(17). By changing the values of Y, the relationship between ~" and Y is obtained. The steady-state values of the enzyme reactor can thus be calculated for any operational conditions.

Example For a model enzyme reactor, the following system is used: El, lactate dehydrogenase (rabbit muscle); EL, alcohol dehydrogenase (horse liver); Cox, P E G - N A D ; Crd, P E G - N A D H ; S~, pyruvate; P~, L-lactate; $2, ethanol; P2, acetaldehyde. Table I shows the kinetic constants of the enzymes. These values were obtained from initial rate studies using P E G NAD(H) as coenzyme, and were used for the calculation of steady-state values. Figure 2 shows the theoretical relationship between ~- and Y obtained as described above. With different sets of operational conditions (in Fig. 2, [ethanol]0 is changed), different curves are obtained. The method of continuous operation of the model enzyme reactor is as follows. The enzymes and P E G - N A D were kept in a model 8MC ultrafiltration apparatus (Amicon Corp., Lexington, MA) fitted with a UM2 i2 I. H. Segel,

in " E n z y m e

Kinetics," p. 593. Wiley, New York, 1975.

[4]

POLYMERIZABLE N A D DERIVATIVE AND P E G - N A D

43

TABLE I KINETIC CONSTANTS OF LACTATE DEHYDROGENASE (RABBIT MUSCLE) AND ALCOHOL DEHYDROGENASE (HORSE LIVER)a Lactate dehydrogenase Vfj K~oxj

22.6 50,6 2330 542 413 21.2 501 2.79

Kpl Kicox.i Vrj K,,~j Ks~ K~¢rdj

--- 0.7 sec -j ± 6,5 ~tM 4- 5 9 0 / t M ± 166/zM ± 12 sec 1 ± 1.1 /xM -+ 27/.tM ± 0,50/zM

Alcohol dehydrogenase Vf,2 Kcox,2 Ks2 Kicox.2 Vr.2 Kcrd.2 Kp2 K~rd,2

0.591 18.7 1950 22.0 10.8 33.1 192 8.74

- 0.023 sec -~ ± 1.5/xM ± 210/xM -+ 3.4/.tM ± 0.2 sec -~ ---- 0.9/zM ± 9 ~M ± 1.05 IzM

Initial rate studies were carried out at 30° in 30 mM sodium phosphate buffer, pH 7.5. Reaction mixture contained the following components. For coenzyme reduction by lactate dehydrogenase: 5-100 mM DL-lactate, 50-400 /.tM PEG-NAD, 30 mM hydrazine, and 3.95 nM enzyme; for coenzyme oxidation by lactate dehydrogenase: 0.57-5.7 mM pyruvate, 5 - 5 0 / z M P E G NADH, and 2.36 nM enzyme; for coenzyme reduction by alcohol dehydrogenase: 1-10 mM ethanol, 6.25-50/zM PEG-NAD, 10 mM semicarbazide' HC1, and 0.584/xM enzyme; for coenzyme oxidation by alcohol dehydrogenase: 0.15-1.5 mM acetaldehyde, 6.25-50/.tM PEG-NADH, and 0.117/zM enzyme. The enzyme concentrations are expressed as subunit concentration. Initial rates were measured spectrophotometrically at 340 nm. Results are expressed -+ standard error.

r Et hanoi ]o 100 I

lO0 mM

f

80

60 rnM

60

30 mM

>" 40 /

10 mM

20

0

I

I

I

1

2

3

-c(h) FIG. 2. Theoretical relationship between residence time (r) and the yield of L-lactate (Y) at steady state with different ethanol concentrations in the eluent ([ethanol]0). Conditions for calculation are [lactate dehydrogenase] = 0.12/xM, [alcohol dehydrogenase] = 2.88/zM, [C]~ = lmM, [pyruvate]0 = 5 mM.

44

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[4]

100

8O

60

> 40

20 I

0

I

I

50 100 150 [ E t h a n o L ] o (raM)

200

FIG. 3. Effect of ethanol concentration in the eluent ([ethanol]0) on the yield of L-lactate (Y) at steady state. (0) Y determined experimentally; error bars indicate fluctuation range; (O) Y calculated for each set of experimental conditions; ( ) Y calculated with the average conditions of these experiments. The average conditions: [lactate dehydrogenase] = 0.11/,~M, [alcohol dehydrogenase] = 2.64/zM, [C]t = 0.749 mM, [pyruvate]0 = 5.0 mM, ~- = 2.61 hr.

ultrafiltration membrane, and the apparatus was placed in a temperaturecontrolled chamber (Tabai LN-110). The continuous enzyme reaction was carried out at 30° in 30 mM sodium phosphate buffer, pH 7.5, containing 0.1% streptomycin and 0.02% NaN3 by continuous ultrafiltration with substrate solution containing pyruvate and ethanol in the buffer as eluent under N2 pressure, and the volume of the reaction mixture was kept at 6.5 ml. The filtrate was collected in fractions at 1-hr intervals, and the Llactate concentration and volume of each fraction were measured. LLactate was assayed with lactate dehydrogenase and NAD.13 The concentration of L-lactate became constant 10-15 hr after the start of the reaction. The steady-state values of L-lactate concentration and residence time 0") were obtained as average values during the steady state over 10 hr (10 measurements). Their fluctuation range was less than 10% of the average value. The concentration of PEG-NAD and the activities of the enzymes in the reactor decreased gradually during the continuous operation. The quasi-steady-state values were estimated from initial and final values by assuming a first-order rate of decrease. The difference between the initial and final values of a steady state was less than 20% of the average value. Figure 3 shows the effects of ethanol concentration in the eluent at a t3 I. Gutman and A. W. Wahlefeld, in "Methods of Enzymatic Analysis" (H. U. Bergmeyer, ed.), 2nd Ed., p. 1464. Academic Press, New York, 1974.

[5]

E N Z Y M E REACTOR BASED ON a q T W O - P H A S E SYSTEMS

45

fixed value of z. The solid and open circles show experimental and theoretical values, respectively. The solid line was calculated with the average conditions of fixed variables in the figure. Deviation of an open circle from the solid line is due to the deviation of the fixed variables of the experiment from their average values. Generally, the steady-state values obtained by continuous operation of the enzyme reactor are lower than those obtained by calculation based on the kinetic model. These differences are thought to be due to the simplifications made for the kinetic model, in which we assumed a simplified ordered Bi-Bi mechanism and neglected the effects of substrate inhibition and so on. However, the differences are not large in the region of Ef -> 0.03 and [C]t -> 0.7 raM, and the steady-state behavior of the enzyme reactor can be explained and predicted semiquantitatively by the simple kinetic model.

[5] A d e n o s i n e 5 ' - T r i p h o s p h a t e R e c y c l i n g in a n E n z y m e Reactor Based on Aqueous Two-Phase Systems B y HIDEO SUZUKI a n d YOSHIMITSU YAMAZAKI

The retention and in situ regeneration of coenzymes (e.g., NAD or ATP) in a reactor are prerequisites for the industrial use of dehydrogenases, kinases, or ligases. One possible approach to this is to bind the coenzymes to water-soluble polymers and use them with the enzymes in an ultrafiltration apparatus. J-3 However, this method has the disadvantage that enzymes as well as the polymers carrying coenzymes tend to be adsorbed on the ultrafiltration membrane and thus lost from the reaction mixtures. J The need for developing a better method for separating polymers from low molecules reminded us of countercurrent partitioning using aqueous two-phase systems. These systems were developed by Albertsson 4 and l y. Yamazaki, H. Maeda, and H. Suzuki, Biotechnol. Bioeng. 18, 1761 (1976). 2 S. Furukawa, N. Katayama, T. Iizuka, I. Urabe, and H. Okada, FEBS Lett. 121, 239 (1980). 3 R. Wichmann, C. Wandrey, A. F. B0ckmann, and M.-R. Kula, Biotechnol. Bioeng. 23, 2789 (1981). 4 P.-A. Albertsson, "Partition of Cell Particles and Macromolecules," 2nd Ed. Wiley, New York, 1971.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[5]

E N Z Y M E REACTOR BASED ON a q T W O - P H A S E SYSTEMS

45

fixed value of z. The solid and open circles show experimental and theoretical values, respectively. The solid line was calculated with the average conditions of fixed variables in the figure. Deviation of an open circle from the solid line is due to the deviation of the fixed variables of the experiment from their average values. Generally, the steady-state values obtained by continuous operation of the enzyme reactor are lower than those obtained by calculation based on the kinetic model. These differences are thought to be due to the simplifications made for the kinetic model, in which we assumed a simplified ordered Bi-Bi mechanism and neglected the effects of substrate inhibition and so on. However, the differences are not large in the region of Ef -> 0.03 and [C]t -> 0.7 raM, and the steady-state behavior of the enzyme reactor can be explained and predicted semiquantitatively by the simple kinetic model.

[5] A d e n o s i n e 5 ' - T r i p h o s p h a t e R e c y c l i n g in a n E n z y m e Reactor Based on Aqueous Two-Phase Systems B y HIDEO SUZUKI a n d YOSHIMITSU YAMAZAKI

The retention and in situ regeneration of coenzymes (e.g., NAD or ATP) in a reactor are prerequisites for the industrial use of dehydrogenases, kinases, or ligases. One possible approach to this is to bind the coenzymes to water-soluble polymers and use them with the enzymes in an ultrafiltration apparatus. J-3 However, this method has the disadvantage that enzymes as well as the polymers carrying coenzymes tend to be adsorbed on the ultrafiltration membrane and thus lost from the reaction mixtures. J The need for developing a better method for separating polymers from low molecules reminded us of countercurrent partitioning using aqueous two-phase systems. These systems were developed by Albertsson 4 and l y. Yamazaki, H. Maeda, and H. Suzuki, Biotechnol. Bioeng. 18, 1761 (1976). 2 S. Furukawa, N. Katayama, T. Iizuka, I. Urabe, and H. Okada, FEBS Lett. 121, 239 (1980). 3 R. Wichmann, C. Wandrey, A. F. B0ckmann, and M.-R. Kula, Biotechnol. Bioeng. 23, 2789 (1981). 4 P.-A. Albertsson, "Partition of Cell Particles and Macromolecules," 2nd Ed. Wiley, New York, 1971.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

46

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[5]

are useful in collecting cell particles 5 and biopolymers. 6'7 They are prepared by mixing aqueous solutions of different polymers. For instance, when a solution of dextran in water is mixed with a solution of polyethylene glycol (PEG) in water, the mixture becomes turbid and after a while it separates into two phases. If proteins or other polymers are added to the mixture, they are often asymmetrically partitioned to the upper and the lower phases, depending on the experimental conditions. Low molecular weight molecules are usually almost symmetrically partitioned into both phases. Thus one can see the possibility of an enzyme reactor based on the principle that enzymes and the polymer-bound coenzymes are retained in the lower phase while the low molecular weight products are extracted to the upper phase. The purpose of this article 8 is to present an outline of such a reactor in the case of kinases, especially for the coupled system of hexokinase and acetate kinase. Design of the System In constructing a reactor of this type, the first step is to find an aqueous two-phase system in which all catalysts (here, acetate kinase, hexokinase, and polymer-bound ATP) are mostly partitioned to either the upper or the lower phase, and in which phase the enzyme reaction can actually proceed. There are many factors affecting aqueous two-phase systems: (1) the species of polymers and their ratios, (2) the species of buffers, their pH, and concentrations, (3) the species of salts in the media and their concentrations, (4) temperature, and so on. An additional factor in the present case is the structure of the ATP carrier, which must be carefully selected with consideration given to what polymers are used for the aqueous two-phase system. We chose dextran as the ATP carrier and a combination of dextran and PEG for the basic phase system) In this system, the upper phase is rich in PEG and the lower phase is rich in dextran; therefore dextran-bound ATP goes to the lower phase by the tendency of similar materials to meet. In contrast, Johansson ~° reported 5 A.-L. Smeds, A. Veide, and S.-O. Enfors, Enzyme Microb. Technol. 5, 33 (1983). 6 K. H. Kroner, H. Hustedt, and M.-R. Kula, Biotechnol. Bioeng. 24, 1015 (1982). 7 A. Veide, A.-L. Smeds, and S.-O. Enfors, Biotechnol. Bioeng. 25, 1789 (1983). 8 The original publication is Y. Yamazaki and H. Suzuki, Biseibutsu Kogyo Gijutsu Kenkyusho Kenkyu Hokoku 52, 33 (1979). 9 A recent review of the syntheses of polymer-bound ATP is by Y. Yamazaki and H. Maeda, Yuki Gosei Kagaku Kyokaishi 41, 1088 (1983). to G. Johansson, Biochim. Biophys. Acta 222, 381 (1970).

[5]

ENZYME REACTOR BASED ON aq TWO-PHASE SYSTEMS

47

that acidic proteins were mostly partitioned in the lower phase when a charged PEG derivative, PEG sulfonate, was used in preparing the system of PEG and dextran. This can possibly be accounted for by electric repulsion. Dextran-bound ATP is negatively charged in neutral solutions owing to the phosphate groups. Thus, a more favorable partition of dextran-bound ATP to the lower phase can be expected in a phase system supplemented with a negatively charged PEG. Hexokinase (pI 4.5-4.81~) and serum albumin (pI 4.7-4.9, j2 used as enzyme stabilizer) are also expected from their isoelectric point to show a higher affinity to the lower phase in such a system. Johansson used PEG sulfonate ( - O 3 S - C H z C H z O (CH2CHzO)n-CHzCH2-SO3-), but we used a more easily preparable polymer, PEG sulfate (-03S-O-CH2CH20-(CH2CH20),-CH2CHz-O-S03-). Synthetic procedures of dextran-bound ATP and PEG sulfate are described below. Preparation of Dextran-Bound A TP 13

The principle used is the introduction of a spacer chain having a terminal amino group to the N-6 site of adenine nucleus and the succeeding coupling to BrCN-activated dextran using the amino group. In the reaction shown in Fig. 1 ADP is used as the starting material, but the ATP derivative, N6-[N-(6-aminohexyl)carbamoyl]-ATP, which is formed probably from the disproportionation (2ADP = ATP + AMP), is purified by column chromatography. One gram of ADP (free acid, Boehringer-Mannheim) is dissolved in 50 ml of hexamethylphosphoramide with stirring at 70°, and then 20 ml of freshly distilled hexamethylene diisocyanate is added. After incubation at 75 ° for 2 hr, the reaction mixture is cooled to 0 ° and slowly poured into an ice-cold two-phase mixture of chloroform (750 ml) and acidic water (750 ml, pH 1 with HC1), under vigorous stirring. The resultant emulsion is transferred into a separatory funnel and allowed to stand overnight at room temperature. The aqueous layer becomes clear and is separated from the chloroform layer, which is reextracted with 750 ml of acidic water (pH 1). The combined aqueous layer is neutralized with 1 M LiOH and concentrated on a rotary evaporator at 40 ° to approximately 100 ml. To the concentrate is added 1 liter of an ice-cold mixture made of ethanol u M. Kunitz and M. R. McDonald, J. Gen. Physiol. 29, 393 (1946). t2 E. G. Young, in "Comprehensive Biochemistry" (M. Florkin and E. H. Stotz, eds.), Vol. 7, p. 22. Elsevier, New York, 1963. u This section is reprinted with permission from Y. Yamazaki, H. Maeda, and H. Suzuki, Eur. J. Biochem. 77, 511 (1977),

48

MULTISTEP ENZYME SYSTEMS AND COENZYMES

H N . , ~ N . " % / " k ~ NC0 ADP

OCNo (CHI) Q- NCO •

t~j ~ - - ~ "~N'~'N /

[5]

HN . - " ~ N ~ / ~ I )

N.~..,,~ N ~'~NI~"N "~

OH OH

H3 CB-

OH OH

Chromatographlc Separation I Dextran-bound ATP

<

I

Br CN-activated dextran FIG. 1. Synthetic scheme of dextran-bound ATP (HMPA, hexamethylphosphoramide; (~),, mono-, di-, or triphosphate).

and acetone (1/I). The precipitate is collected by centrifugation and dried in vacuum. The resultant white powder is dissolved in water, adjusted to pH 7, and adsorbed on a Dowex I-X2 (C1-) column (1.4 × 55 cm). The column is washed with 150 ml of 50 mM LiCI (pH 5.0). A linear gradient of LiCI is then applied. The mixing vessel contains 50 mM LiCI (1.2 liters, pH 5.0) and the reservoir 350 mM LiC1 (1.2 liters, pH 2.0). The effluent is collected as 15-ml fractions. The nucleotides are eluted into three major peaks (fractions 400-600, 900-1100, and 1400-1600 ml). These peaks correspond to the AMP derivative, ADP derivative, and ATP derivative, respectively. The fractions in the latter major peak are pooled, adjusted to pH 7 with 1 M LiOH, and concentrated on a rotary evaporator at 30° to approximately 30 ml. To the concentrate is added 300 ml of an ice-cold mixture of ethanol-acetone (I/1). The precipitate is collected by centrifugation and dried in vacuum. This crude product is again purified by Dowex l-X2 column chromatography in a similar manner as described above. The precipitation procedure is repeated twice and the final precipitates are dried for 24 hr over P205 in vacuum. Thus, 193 mg of N6-[N-(6 aminohexyl)carbamoyl]-ATP is obtained as white powder (12% yield as tetralithium salt). Coupling of the ATP derivative to soluble dextran is carried out essentially according to the method by Mosbach e t al. 14 Dextran T40 (2 g, Pharmacia) is dissolved in 20 ml of distilled water and activated at pH 11.0 by addition of 100 mg cyanogen bromide in 2 ml of water and continuous 14 K. Mosbach, P.-O. Larsson, and C. Lowe, this series, Vol. 44, p. 859.

[5]

ENZYME REACTOR BASED ON a q TWO-PHASE SYSTEMS

49

titration with I M NaOH at room temperature. After 5 min the consumption of NaOH ceases, and then the pH is lowered to 9.5 with 0.1 M HCI. To the activated dextran solution is added 600 mg of the above ATP derivative dissolved in 5 ml of 0.1 M NaHCO3, pH 8.5. After the reaction mixture has been stirred overnight at room temperature, 4 ml of 0.8 M ethanolamine-HCl buffer, pH 8.0, is added to quench any residual active groups. The reaction mixture is allowed to stand for an hour at room temperature and then applied to a Sephadex G-50 column (2.5 x 112 cm) equilibrated with 0.1 M LiCI. Elution is performed with 0.1 M LiCI, and the effluent is collected as 15-ml fractions. Dextran-bound ATP is eluted in fractions 12-20 and the uncoupled derivative in fractions 21-40. The fractions containing the polymer are pooled, concentrated 5-fold by ultrafiltration (Amicon PM10 membrane) and then dialyzed against 0.1 M triethanolamine-HCl buffer, pH 7.6, for 2 days. The buffered solution of dextran-bound ATP is used in all subsequent studies. Determination of phosphate ~5 and total hexoses ~6 has indicated a nucleotide content of 43 /xmol/g dry dextran and a coupling yield of 15%.

Preparation of PEG Sulfate 8 PEG 4000 (10 g, Wako Pure Chemical Industries Ltd., Osaka) is dissolved in 20 ml of freshly distilled chloroform. To this ice-cold solution is added 0.84 ml of chlorosulfonic acid. The mixture is gently shaken for 10 min under cooling with ice and then stirred at room temperature for 1 hr. The reaction mixture is again cooled with ice and mixed with about 20 g of ice and 20 ml of water containing 1 g of NaOH. This mixture is vigorously stirred at room temperature for 10 min. The emulsion is evaporated under reduced pressure at 35° to remove chloroform. The residual syrup is mixed with 100 ml of water and 100 ml of ethanol. The solution (pH 2) is neutralized with 2 M NaOH and then concentrated in vacuum at 65°. The remaining water is removed as the ethanol azeotrope. The residue is dissolved in 50 ml of hot ethanol. Insoluble salts are filtered off and washed with ethanol (I0 ml, 2x). The combined filtrate and washings are allowed to stand at - 2 0 ° overnight. The precipitate is collected by filtration and then dissolved in 100 ml of chloroform. After addition of ether (400 ml), the solution is cooled to -20 ° for 2 hr. The precipitate is collected by suction filtration and dried in vacuum at 80° for 1 hr. Thus 7.83 g of PEG sulfate was obtained as white solids, mp 48 °. Analysis: Found: S, 1.2%. Calcd. for NaSO4(CH2CHzO)70SO3Na: S, 1.9%; the degree of polymerization 70 is an assumption based on the ~5 G. R. Bartlett, J. Biol. Chem. 234, 466 (1959).

16 j. H. Roe, J. Biol. Chem. 212, 335 (1955).

50

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[5]

average molecular weight (3000, information from the manufacturer) of PEG 4000. PEG sulfate migrated faster (5.7 cm/hr) to the anode than PEG (4.2 cm/hr) in zone electrophoresis [on a cellulose-coated glass plate (Avicel SF, 5 x l0 cm, 0.2-0.25 mm thick) using 10 mM phosphate buffer, pH 6.8, and applying a potential of 600 V for 30 min; spots were located in 12 vapor]. Investigation of Optimal Partition The reagents used for making the system are (1) enzymes and polymers to be partitioned: hexokinase, acetate kinase, bovine serum albumin, and dextran-bound ATP; (2) polymers for the preparation of twophase systems: dextran T500, PEG 4000, and PEG sulfate; (3) substrates for the enzymes: glucose and acetyl phosphate. Other factors to be selected are species of buffer and additional salts, pH, and temperature. However, the ranges within which these can be selected is narrowly limited for the enzyme reaction. We used triethanolamine-HC1 buffer (pH 7.6). The optimal conditions for partitioning all of the proteins and dextran-bound ATP into the lower phase were investigated by changing the concentrations of dextran, PEG, PEG sulfate, and the buffer. Determination o f Partition Coefficients ~7

A dextran solution (20%, w/w) is prepared by dissolving l0 g of dextran T500 (Pharmacia) into 40 ml of water. PEG + PEG sulfate solutions (20%, w/w; ratio of PEG sulfate/PEG = 0/1, 1/3, or I/l) are prepared by dissolving l0 g of PEG 4000 or mixtures of PEG 4000 and PEG sulfate into 40 ml of water. Phase systems containing 8% (w/w) PEG + PEG sulfate and 8% (w/w) dextran are prepared by mixing the 20% solutions (0.8 g each) and 0.4 ml of triethanolamine-HCl buffer (10, 50, or 150 mM), pH 7.6, containing the materials to be tested. The amounts of the materials are given in the legends to Figs. 2 and 3. Yeast hexokinase (BoehringerMannheim) and E. coli acetate kinase (Boehringer-Mannheim) are dialyzed against 2 mM triethanolamine-HCl buffer, pH 7.6, overnight before use. The systems are mixed by gentle shaking and allowed to separate by centrifugation (3000 rpm, for l0 min) at about 20°. Samples (0.2 ml each) are removed from the upper phase and then from the lower phase using hole pipettes. The partition coefficient (K, concentration in the upper phase divided by concentration in the lower phase) for each material is ~v R e p r o d u c e d from Y a m a z a k i and Suzuki 8 with permission.

[5]

51

ENZYME REACTOR BASED ON aq TWO-PHASE SYSTEMS

calculated from the following values of the samples for each phase: absorbance at 269 and 280 nm for dextran-bound ATP and albumin, respectively, and activity for enzymes. The absorbances are measured, following adequate dilution of the samples with water, using a blank prepared from the corresponding system without the materials. The enzymes are assayed from the initial velocities of absorbance change (A340/min) in the following reaction mixtures at 25°: for hexokinase, in a total volume of 3.03 or 3.10 ml containing 242/zmol triethanolamine-HC1 buffer (pH 7.6), 20/zmol MgSO4, 667 ~mol glucose, 8.2 /zmol ATP, 2.2/zmol NADP, 1.4 U yeast glucose-6-phosphate dehydrogenase (Boehringer-Mannheim), and hexokinase; and for acetate kinase, in a total volume of 3.00 or 3.19 ml containing 204/~mol triethanolamineHCI buffer (pH 7.6), 4/zmol MgSO4, 1000/zmol acetate, 3.3/zmol phosphoenolpyruvate, 16.2/zmol ATP, 0.96/zmol NADH, 15 U rabbit muscle pyruvate kinase (Boehringer-Mannheim), 40 U beef heart lactate dehydrogenase (Miles), and acetate kinase. Figure 2 shows that the proteins and dextran-bound ATP were all partitioned more than 90% in the lower phase when PEG contained 25% PEG sulfate. This fact confirms the validity of our prediction regarding the design for a two-phase system. In addition, there was a tendency for the lower buffer concentration to give a higher affinity to the lower phase. The system, however, did not contain the substrates for the enzymes. The partition coefficients of the three catalysts were thus also determined in the same aqueous two-phase system as the above except that all the necessary substances for the coupled reaction of hexokinase and acetate kinase were added (Fig. 3). Dextran-bound ATP %PEG sulfate 0

25

50

Hexoklnase %PEG sulfate 0

25

50

o£

oT

oT

Acetate kinase %PEG sulfate 0

o

:}5

o

50

oT

Albumin %PEG sulfate 2S 50

o-

oT

-- ~

"It

-2 -

-2

-

FI6. 2. Partition of dextran-bound ATP, hexokinase, acetate kinase, and albumin in aqueous two-phase systems containing dextran T500 and PEG plus PEG sulfate, t7 Quantities in the 0.4 ml of triethanolamine-HC1 buffer: dextran-bound ATP, 7 mg; hexokinase, 10 U; acetate kinase, 1 U; bovine serum albumin (Miles), 2 mg; the buffer: 2 mM (O), 10 mM (A), or 30 mM ([~). The buffer concentrations were calculated by assuming homogeneous dispersion and a total volume of 2 ml. % PEG sulfate = [PEG sulfate/(PEG sulfate + PEG)] x 100. K is the partition coefficient.

52

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[5]

%PEG sulfate 0 0

25

-

0

)

- -

m

-

1.6

- 1.6

FIG. 3. Partition of dextran-bound ATP, hexokinase, and acetate kinase in aqueous twophase systems containing dextran T500, PEG plus PEG sulfate, and the substrates for the enzymes.17Quantities in the 0.4 ml of triethanolamine-HCl buffer (10 mM): dextran-bound ATP, 3.5 rag; hexokinase, 10 U; acetate kinase, 1 U; albumin, 2 mg; MgSO4, 14/xmol;acetyl phosphate, 15/xmol;and glucose, 180 tzmol. (O) Dextran-bound ATP; (A) hexokinase; (V1) acetate kinase. When the substrates and Mg 2+ ion were present in quantity in the system, the partition coefficients of all three catalysts were found to increase as compared to those obtained in the former system lacking the substances. The reduced affinity of the catalysts to the lower phase can possibly be interpreted as an electric repulsion between them and the acetyl phosphate present in the lower phase, or as the formation of ionic bonds between them and the PEG sulfate through the Mg 2+ ion. The results in Fig. 3 are somewhat undesirable for our purpose. Nevertheless, the three catalysts were still partitioned 6-20 times more in the lower phase when P E G sulfate was added to the system. We surmised that the catalysts could be retained in a reactor using the aqueous two-phase system by carrying out countercurrent partitioning several times. Application to Repeated E n z y m e Reaction

Repeated Production of Glucose 6-Phosphate 17 The principle is the coupled reaction: glucose

/-

dextran-bound ATP

/"

acetate

hexokinase acetate kinase d/ ~ dextran-boundJ ~,. glucose 6-phosphate ADP - - acetyl phosphate

F o u r test tubes are used (see Fig. 4). A phase system is prepared by mixing 0.2 g of 20% (w/w) (75% P E G + 25% P E G sulfate) solution, 0.2 g

[5]

ENZYME REACTOR BASED ON aq TWO-PHASE SYSTEMS

tube

A B substrate/'~'~ Ition~ t

C

53

D ~'~ IO0*C,5rain t ~)determi natioofn

glucose 6-phosphate

,

|Dextran

[ ~

~PEG+PEG sulfate [Substrates, Buffer

Oextran PEG+PEG sulfate Enzymes Dextran-bound ATP Substrates,

[Dextran

~PEG +PEGsulfate tBuffer

Buffer

FIG. 4. Illustration of the procedure for repeatedly producing glucose 6-phosphate in an aqueous two-phase system) 7

of 20% (w/w) dextran T500 solution, and 0.1 ml of 10 mM triethanolamine-HCl buffer, pH 7.6, containing 1 mg of dextran-bound ATP, 3.5 /.~mol MgSO4, 3.8 /~mol acetyl phosphate (Boehringer-Mannheim), 45 /~mol glucose, 0.5 mg albumin, 2.8 U hexokinase, and 0.08 U acetate kinase. This mixture is placed in test tube A. In tubes B, C, and D are placed the lower phases (0.25 ml each; the same volume as that of the lower phase in tube A) from the corresponding system minus all of dextran-bound ATP, the enzymes, albumin, and the substrates. In addition, a substrate solution is prepared which is the upper phase of the system having the same composition as the above, but without the three catalysts and albumin. The initial content in tube A is incubated at 25 ° for 30 rain with shaking and then allowed to separate by centrifugation for 10 rain. The upper phase is transferred to tube B with a pipette and a new substrate solution (0.25 ml; the same volume as that of the upper phase) is added to tube A. The tubes A and B are incubated at 25 ° for 30 min with shaking. After centrifugation, the upper phase in tube B is transferred to tube C and that in tube A to tube B. Then, a new substrate solution (0.25 ml) is added to tube A. In this manner, the upper phases in tube A are extracted four times. Immediately after the final extraction, the upper phases are heated to 100° for 5 min and then subjected to the enzymatic determination of glucose 6-phosphate. The cooled sample is mixed with 2 ml of 0.1 M triethanolamine-HCl buffer (pH 7.6) and centrifuged to remove the denatured proteins. One volume of the supernatant (1 ml) is mixed with 1.5 ml of the buffer, 0.1 ml of 0.1 M MgSO4, and 0.2 ml of NADP solution (10 mg/ml). Absorbance increment at 340 nm after addition of yeast glucose6-phosphate dehydrogenase (10/xl, 3.5 U) and incubation at 28 ° for 30 min

54

M U L T I S T E P E N Z Y M E SYSTEMS A N D C O E N Z Y M E S

[5]

is read, from which the concentration of glucose 6-phosphate is calculated using the value of ~340 = 6.2 × 10 3 M -1 cm 1. The four times repeated extraction procedure was carried out to reextract some part of the catalysts partitioned in the upper phases into the lower phases. In this way, glucose 6-phosphate was repeatedly produced in the final upper phases (Fig. 5) even though the catalysts were added only one time in tube A. The maximum amount of glucose 6-phosphate production was 1.9/xmol (in phase 9). This value was comparable to that obtained at the equilibrium of the coupled reaction in one batch experiment (see the original workS), and therefore the time (at phase 9) was considered to be the start of a steady state for this production. The total operation was repeated nine times and stopped, since the substrate solution was used up. Thus, the final upper phases 1-9 had passed the four times extracting procedure before the heating treatment, but the final upper phases I0, 11, and 12 were heated when withdrawn from tubes C, B, and A, respectively. In other words, the final upper phases 11 and 12 had not contacted with the lower phases in tubes C and B, respectively. The fact that the quantities of glucose 6-phosphate in the final upper phases 11 and 12 were less than those of the preceding final upper phases (8-10) suggests that some of the enzymes and dextran-bound ATP, which had been supplied in tube A, was displaced into tubes B and C by moving the upper phases. Conversely, for retaining most of the catalysts as long as possible in the whole system, the upper phases should be reextracted more times by the lower phases. Concluding Remarks The present work provides a prospect for using aqueous two-phase systems to retain enzymes, especially coupled enzyme systems including @

I

2

3

4

5

15

7

8

9

I0

II

12

Number of the Upper Phase

FIG. 5. Repeated enzymatic production of glucose 6-phosphate in an aqueous two-phase system containing dextran T500 and PEG plus PEG sulfate.t7 Quantities of glucose 6-phosphate in the upper phases finally withdrawn out of the whole system are plotted against the numbers of these phases.

[5]

ENZYME REACTOR BASED ON aq TWO-PHASE SYSTEMS

55

coenzymes, in bioreactors. In another experiment, we also found that native ATP was partitioned more in the lower phase (K = 0.6) in a system containing dextran T500, PEG, and PEG sulfate. 8 Pollak and Whitesides 18 found that another important coenzyme, NADP, was partitioned more in the lower phase (K = 0.44) in an aqueous two-phase system composed of Ficoll (polysucrose) and UCON (copolymer of propylene glycol and ethylene glycol). These facts suggest the possibility that the coenzymes might be retained, without being bound to polymers, in the reactor if countercurrent partitioning is repeated several times. Of course, a small portion of the catalysts actually passes into the upper phase and is finally drawn out of the whole system. This loss is intrinsic and cannot be avoided. However, it can be minimized if the partition coefficients are exceedingly high or low. A possible way to realize this is to find a more desirable phase system by changing the materials and/or the conditions for the systems or by chemically modifying the catalysts. In fact, when carboxymethyl dextran was used as the ATP carrier, this ATP derivative showed a remarkably lower partition coefficient (<0.01) than the usual dextran-bound ATP in the systems for Fig. 2 which contained PEG sulfate. 8 The present method will be useful if the required enzymes are too unstable to be immobilized on solid supports or to be employed in usual solutions, although stable in polymer solutions. It is highly possible that such a situation will be encountered, since several enzymes were reported to be stabilized by polymers.19 Furthermore, if the substrates and/or the products are of relatively high molecular weight and thus the usual techniques of enzyme immobilization cannot be employed, the aqueous twophase partition method will be useful. Bioreactors based on aqueous two-phase systems have also been reported for ethanol production 2°,2~and aminoacylase reaction. 22

18 A. Pollak and G. M. Whitesides, J. Am. Chem. Soc. 98, 289 (1976). z9 R. D. Schmid, in "Advances in Biochemical Engineering" (T. K. Ghose, A. Fiechter, and N. Blakebrough, eds.), Vol. 12, p. 64. Springer-Verlag, Berlin, 1979. z0 I. Kfihn, Biotechnol. Bioeng. 22, 2393 (1980). 2~ B. Hahn-H~igerdal, E. Andersson, M. Lrpez-Levia, and B. Mattiasson, Biotechnol. Bioeng. Symp. 11, 651 (1981). 22 W. Kuhlmann, W. Halwachs, and K. SchiJged, Chern.-lng.-Tech. 52, 607 (1980).

56

MULTISTEP

ENZYME

SYSTEMS AND COENZYMES

[6]

[6] A l c o h o l D e h y d r o g e n a s e C o i m m o b i l i z e d w i t h Its C o e n z y m e

By KEITH J. LAIDLER and M. ABDUL MAZID Whereas much work has been done on the separate immobilization of enzymes and coenzymes, little has as yet been done on their coimmobilization. Many enzymes, such as the dehydrogenases, require the participation of easily dissociable coenzymes, and for obvious practical and economic reasons it is useful to be able to immobilize them side by side. Coimmobilization is, however, only successful if certain conditions are satisfied. It is important for the enzyme and coenzyme to be held in a suitable spatial arrangement and for the coenzyme to be easily regenerated during the course of reaction. The problem of spatial arrangement may often be controlled by the introduction of suitable spacing groups; these can be placed between the supporting matrix and the coenzyme, which is thus more flexible. The matter of regeneration of the coenzyme also has to be dealt with by positioning it suitably with respect to the enzyme. Because of these factors the coimmobilization of enzymes and coenzymes has not always been met with complete success. In the case of alcohol dehydrogenase and its coenzyme nicotine adenine dinucleotide (NAD), the recycling of the latter has been achieved for many years, I-9 and derivatives of the coenzyme with flexible side arms have been synthesized~°-~2; there are, however, few reports on their coimmobilization. Immobilization of alcohol dehydrogenase itself has been l N. K. Gupta and W. G. Robinson, Biochirn. Biophys. Acta 118, 429 (1%6). 2 C. L. Woodley and N. K. Gupta, Anal. Biochem. 43, 341 (1971). 3 M. P. Schulman, N. K. Gupta, A. Omachi, G. Hoffman, and W. E. Marshall, Anal. Biochem. 60, 302 (1974). 4 G. E. Glock and P. McLean, Biochem. J. 61, 381 (1955). 5 C. A. Villee, Biochem. J. 83, 191 (1962). 6 T. F. Slater and B. Sawyer, Nature (London) 193, 454 (1962). 7 T. F. Slater, B. Sawyer, and U. Strauli, Arch. Int. Physiol. Biochim. 72, 427 (1964). s j. S. Nisselbaum and S. Green, Anal. Biochem. 27, 212 (1969). 9 C. Bernofsky and M. Swan, Anal. Biochem. 53, 452 (1973). io C. R. Lowe and K. Mosbach, Eur. J. Biochem. 49, 511 (1974). 11 M. Muramatsu, I. Urabe, Y. Yamada, and H. Okada, Eur. J. Biochem. 80, 111 (1977). 12 H. L. Schmidt and G. Grenner, Eur. J. Biochem. 67, 295 (1976).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[6]

ALCOHOL DEHYDROGENASE COIMMOBILIZED WITH NAD

57

reported by a number of workers, 13-18 and the preparation and uses of NAD analogs for affinity adsorbents are well documented in this series of volumes) 9,2° However, it was only in 1975 that Gestrelius et al. 21 first reported the preparation of an alcohol dehydrogenase-NAD(H)Sepharose complex which did not require for its activity the addition of any soluble coenzymes. It involved coimmobilization of horse liver alcohol dehydrogenase and an NAD(H) analog, N6-[(6-aminohexyl)carbamoylmethyl]-NAD(H), to Sepharose 4B under conditions which permitted binary complex formation between the enzyme and the coenzyme derivative prior to their immobilization into the gel. A somewhat similar method for the coimmobilization of alcohol dehydrogenase and NAD on a water-dispersible high molecular weight polymer containing methyl methacrylate and acrylic acid has been reported in a Japanese patent. 22In another approach, Grunwald and Chang23-25described the microencapsulation of a multienzyme system involving yeast alcohol dehydrogenase and malate dehydrogenase with soluble dextran-NAD, for the continuous regeneration of the latter within a collodion-hemoglobin24 and a nylonpolyethyleneimine microcapsule. 25 We have more recently reported the covalent attachment of NAD and yeast alcohol dehydrogenase to the interior surface of a partially hydrolyzed nylon tube. 26 Materials for Immobilization Alcohol dehydrogenase occurs in a variety of plant and animal organisms. The most widely studied alcohol dehydrogenase, a cytoplasmic protein, is usually obtained from bakers' yeast which is also the source of the coenzyme NAD. The other commonly studied alcohol dehydrogenase is that obtained from mammalian liver, particularly equine liver. Both 13 W. E. Hornby, D. J. Inman, and A. McDonald, FEBS Lett. 23, 114 (1972). 14 A. C. Hohansson and K. Mosbach, Biochim. Biophys. Acta 370, 348 (1974). 15 N. Kelly, A. Flynn, and D. B. Johnson, Biotechnol. Bioeng. 19, 1211 (1977). l~ M. J. Brougham and D. B. Johnson, Int. J. Biochem. 9, 283 (1978). 17 S, Barry, T. Griffin, and D. B. Johnson, Int. J. Biochem. 9, 289 (1978). is M. A. Mazid, Ph.D. Thesis, University of Ottawa, Ottawa, Ontario, Canada, 1981. 19 M. J. Harvey, D. B. Craven, C. R. Lowe, and P. D. G. Dean, this series, Vol. 34, p. 242. 20 K. Mosbach, P.-O. Larsson, and C. Lowe, this series, Vol. 44, p. 859. 21 S. Gestrelius, M.-O. Mansson, and K. Mosbach, Eur. J. Biochem. 57, 529 (1975). 22 Nitto Electric Industrial Co. Ltd., Japanese Patent 57,150,387 [82,150,387] (CI. C12NI 1/ 08), September 17, 1982. 23 j. Grunwald and T. M. S. Chang, Biochem. Biophys. Res. Commun. 81, 565 (1978). 24 j. Grunwald and T. M. S. Chang, J. Appl. Biochem. 1, 104 (1979). 25 j. Grunwald and T. M. S. Chang, J. Mol. Catal. 11, 83 (1981). 26 M. A. Mazid and K. J. Laidler, Biotechnol. Bioeng. 24, 2087 (1982).

58

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[6]

alcohol dehydrogenase and its coenzyme are commercially available in different grades from a number of suppliers, such as Sigma Chemical Company in the United States, Boehringer-Mannheim in Germany, and several others. The supporting materials such as Sepharose and dextran are available from Sigma and Pharmacia Fine Chemicals (Uppsala, Sweden), while Bio-Rad Laboratories (Richmond, VA) sells them under the trade name of BioGel A. Nylons can be obtained in different forms from a number of sources. All other materials including various reagents and chemicals, notably 1-cyclohexyl-3(2-morpholinoethyl)carbodiimide metho-p-toluene sulfonate (CMC), a water-soluble condensation agent, are also available commercially. Coimmobilization Methods The various methods of coimmobilization involve somewhat different principles. They all basically use a derivative of the coenzyme such as N 6[(6-aminohexyl)carbamoylmethyl]-NAD which provides some degree of flexibility in binding with the enzyme under restrictions imposed by its immobilization, except in the case of microencapsulation in which the coenzyme would not otherwise be retained. The preparation and uses of NAD derivatives have been reviewed earlier in this series.19,2° Our own method of coimmobilization will be described as an example of covalent attachment to the inner surface of a nylon tube. Gestrelius e t al. 21 coimmobilized horse liver alcohol dehydrogenase and the NAD(H) analog, N6-[(6-aminohexyl)carbamoylmethyl]-NADH, to Sepharose 4B following the commonly used method of cyanogen bromide activation under alkaline conditions. 27 The enzyme and the NADH analog were mixed prior to immobilization to permit initial formation of the binary enzyme-coenzyme complex, thereby allowing them to interact more efficiently after the immobilization. Both thermal and storage stability of the enzyme was increased to some extent, allowing a high recycling rate which resulted from protection of the active site by the coimmobilized coenzyme, the latter being fixed in or near the active site of the enzyme. The coenzyme was used in its reduced form (NADH) because of its lower dissociation constant than that for NAD, 28 which thus favored binary complex formation, while the analog was chosen to provide a "spacer arm" as well as a functional group suitable for covalent binding to the CNBr-activated Sepharose matrix. This method may be compared 27 R. Axen, J. Porath, and S. Ernbach, Nature (London) 214, 1302 (1967). 28 H. Theorell and A. D. Winer, Arch. Biochem. Biophys. 83, 291 (1959).

[6]

ALCOHOL DEHYDROGENASE COIMMOBILIZED WITH N A D

59

with that reported by the Japanese workers, 22 in which the supporting gel consisted of a copolymer of methylmethacrylate and acrylic acid, and in which the polymerization was initiated by potassium sulfate. The coupling of the mixed enzyme and NAD was carried out in a suspension of gel particles in the presence of the water-soluble carbodiimide 1-cyclohexyl3-(2-morpholinoethyl)carbodiimide metho-p-toluene sulfonate (CMC) at a temperature of 5°. Grunwald and Chang 24 described the preparation of immobilized NAD-N6-[N-(6-aminohexyl)acetamide] with soluble dextran T70 (Pharmacia) activated with cyanogen bromide, following the method of Larsson and Mosbach. 29Their procedure for coimmobilization involves preparation of semipermeable microcapsules 23'25 by interfacial polymerization in which an aqueous phase containing yeast alcohol dehydrogenase, malate dehydrogenase, dextran-NAD analog, 1,6-diaminohexane, and polyethyleneimine (used as a microcapsule filler) is emulsified in an organic phase of I : 4 (v/v) chloroform : cyclohexane solution with Span 85, followed by the addition of terephthaloyl chloride. The polymerization reaction is terminated by the addition of extra organic solvent, and the resulting microcapsule is sedimented by centrifugation. Finally, the covalent attachment procedure used by ourselves with nylon tubing 26 involves activation of the inner surface of nylon tubing by partial acidic hydrolysis, followed by condensation of the amino group of N6-[(6-aminohexyl)carbamoylmethyl]-NAD with the liberated carboxyl group, and then coupling of the enzyme to the amino groups via glutaraldehyde. This method leads to a stable enzyme-coenzyme tube and one that is ideally suitable as an open tubular heterogeneous enzyme reactor (OTHER), in which diffusional limitations are unimportant since the enzyme and the coenzyme are both suspended in solution from the interior surface of the tube. However, the activity of such a system is limited by the extent to which the coenzyme molecules are properly oriented with respect to the active site of the enzyme. This method will now be described in some detail. A t t a c h m e n t Procedure Inside a N y l o n Tube

The coimmobilization of the enzyme, yeast alcohol dehydrogenase (233 U/mg protein), and the coenzyme analog, N6-[(6-aminohexyl)carba moylmethyl]-NAD, synthesized by the method of Lindberg et al., 3° is .,9 P.-O. Larsson and K. Mosbach, F E B S Lett. 46, 119 (1974). 3o M. Lindberg, P.-O. Larsson, and K. Mosbach, Eur. J. Biochem. 40, 187 (1973).

60

M U L T I S T E P E N Z Y M E SYSTEMS AND COENZYMES

[6]

summarized in the following scheme: Nylon,'-,~,,,CONH,,,,,,,,~

nylon partial hydrolysis H+/H20

) ~COO-

H3H +,,',,,,,,~

H2N(CH2)6NHCOCH2 + NAD

) " , " , ~ C O H N ( C H 2 ) 6 N H C O C H 2 - N A D + H3N +,'.~,~ CMC, pH 5, 21° glutaraldehyde ' """~COHN(CHz)6NHCOCH2-NAD + OCH(CH2)3CH=N"~"" pH 8.5, 0-5 o H2N-enzyme "~'~COHN(CHz)6NHCOCHz-NAD + enzyme--N=CH(CH2)3CH=N,,,,,,,,~ pH 7.5, 0-5 °

This procedure is essentially the same as that described in a previous paper 26and is a modification of our earlier method for the covalent attachment of the enzyme only31; in that procedure the carboxyl groups of the partially hydrolyzed nylon tubing is blocked by reaction with benzidine. Nylon tubing (type 6, from John Tullis, Scotland) 1 m in length and of 0.068 cm internal radius is used as a tightly wound coil. The tube is first filled with a mixture of 18% (w/v) CaCI2 in a solution of 18% (v/v) water in methanol, and incubated for 30 min at 40 °, followed by rinsing with water at a flow rate of 5 cm3/min for 20 min, to increase the available surface area. Partial hydrolysis of the inner surface is then accomplished at 40 ° by pumping 4 M HCI at a flow rate of 5 cm3/min. This is stopped after 20 min by washing the tube, at the same flow rate, with ice-cold distilled water for 30 min. In the next step, 100 mg of the NAD derivative is dissolved in 10 cm 3 water to which 500 mg of the carbodiimide CMC has been added, and the pH is adjusted to 5. This solution is circulated for 24 hr at a flow rate of 0.5 cm3/min through the tube kept in a water bath at room temperature (-21°). The tube is then rinsed with water at a flow of 5 cm3/min for 1 hr. The enzyme is coupled to the free amino groups, liberated during the acid hydrolysis of the tube, through the bifunctional reagent glutaraldehyde. A solution of 12.5% (v/v) glutaraldehyde in 0.2 M borate buffer, pH 8.5, is passed through the tube in which the NAD analog is already attached, immersed in crushed ice, at a flow rate of 1 cm3/min for 2 hr. Any unreacted glutaraldehyde is removed by washing the tube with 0.1 M phosphate buffer, at pH 7.5, at a flow rate of 5 cm3/min for some time. The last step involving the coupling of the enzyme to the glutaraldehyde residues is carried out by circulating overnight a solution of 10 mg enzyme in 10 cm 3 buffer (0.1 M phosphate, pH 7.5) at a flow rate of 0.5 cma/min, the tube being maintained in a bath of crushed ice. Finally, any unbound enzyme is removed by perfusing the tube with 0.1 M phosphate buffer, pH 7.0, containing 0.5 M NaC1, at a flow rate of 5 cm3/min for 3 hr. 31 M. A. M a z i d a n d K. J. L a i d l e r , Biochim. Biophys. Acta 614, 225 (1980).

[6]

ALCOHOL DEHYDROGENASE COIMMOBILIZED WITH NAD

61

The tube is next filled with 0.1 M phosphate buffer, at pH 7.5, containing 1 mM EDTA and 0.1 mM 2-mercaptoethanol and, when not in use, is stored below 5° in a water bath kept in a refrigerator. Kinetic Studies The reaction usually studied involves the oxidation of ethanol by NAD with the production of acetaldehyde and NADH, the reduced form of the immobilized coenzyme. Unless the latter is recycled efficiently, very little reaction occurs and it is not practical to measure the rate. A number of methods, both enzymatic and nonenzymatic, are available for this purpose. ~-9 We found that the rapid nonenzymatic reduction of 2,6-dichlorophenolindophenol (DCPIP) dye by NADH in the presence of phenazine ethosulfate (PES) is suitable in our tubular flow reactor system. Gestrelius et al. 21 also studied different methods, and following the method of Schulman et al. 3 found that the most successful regeneration was accomplished by using the coupled oxidoreduction between the two alternative substrates, ethanol and lactaldehyde. Our method is similar to that described by Slater and co-workers6.7; the entire reaction scheme is as follows: -reduced PES --. 2,6-dichlorophenol ethanol ](NAD ~(phenazine ethosul-~( indophenol (DCPIP) acetaldehyded~NADH J\fate (PES) */~reduced DCPIP

In this method, PES acts as a mediator for electron transport during the oxidation of NADH, and reduces DCPIP. Phenazine methosulfate has also been used, but PES is chemically more stable. It is to be noted that PES and DCPIP can oxidize 2-mercaptoethanoP 2 which was used for the storage of the tube, and any traces of the latter must therefore be removed before a kinetic run is carried out. Reagents and Methods

Both PES and DCPIP were obtained from the Sigma Chemical Company. They were used in fixed concentrations of 0.25 and 70 mM, respectively, which were sufficiently high to permit measurement of the N A D ethanol reaction rate. The concentration of the substrate ethanol was varied from 5 to 50 mM in the reaction mixture with PES and DCPIP. All solutions were freshly prepared in 0.1 M phosphate buffer, pH 7.5 - 0.05, made up with doubly distilled deionized water. They were protected from 32 C. Bernofsky and K. M. Royal, Biochim. Biophys. Acta 215, 210 (1970).

62

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[6]

exposure to air and light, which reoxidizes DCPIP. The flow-kinetic runs were carried out as described previously, 3~ using a Pye Unicam SP 1800 UV spectrophotometer. Rates were calculated from the decrease in absorbance of DCPIP at 600 nm, under steady-state conditions, in which the absorption coefficient was taken to be 20,000 M -~ cm -~.33 Also, the activity of the tube was monitored periodically by assays at a linear flow rate of 0.69 cm/sec with 20 mM ethanol, 0.25 mM PES, and 70 mM DCPIP, at 25.5 ° and pH 7.5. The tube showed no measurable loss of enzymatic activity over a period of 1 month.

Analysis of Data The methods of analysis of flow-kinetic data are based on the Kobayashi-Laidler theory 34 developed originally for one-substrate reactions catalyzed by immobilized enzymes; the procedure has been described in a number of previous papers 34-4° and recently reviewed. 41 The methods have been found to be equally applicable to two-substrate reactions under the limiting conditions of one substrate in excess while the other is varied) 1,42-44However, the present system is essentially a one-substrate one, since the coenzyme is immobilized and immediately recycled so that its concentration is practically constant. The equations used in the analyses were summarized in a previous paper 26 and are given here very briefly. The apparent Michaelis constant Km,app is related to the inherent Michaelis constant K~ through the mass transfer coefficient kL as follows: Km,app =

Vm

t

K~n + ~ L = Km +

~ 8 (\D2/ rt ~1/3 O f 1/3

(1)

where Vm is the limiting rate for the immobilized enzyme at the flow rate vf, r is the radius and L the length of the tube, and D is the diffusion coefficient for the substrate in solution. 33 D. S. Coffey and L. Hellerman, Biochemistry 3, 394 (1964). 34 T. Kobayashi and K. J. Laidler, Biotechnol. Bioeng. 16, 99 (1974). 35 p. S. Bunting and K. J. Laidler, Biotechnol. Bioeng. 16, 119 (1974). 36 T. T. Ngo and K. J. Laidler, Biochim. Biophys. Acta 377, 317 (1975). 37 D. Narinesingh, T. T. Ngo, and K. J. Laidler, Can. J. Biochem. 53, 1061 (1975). 38 T. T. Ngo, D. Narinesingh, and K. J. Laidler, Biotechnol, Bioeng. 18, 119 (1976). 39 T. T. Ngo and K. J. Laidler, Biochirn. Biophys. Acta 525, 93 (1978). 4o T. T. Ngo, K. J. Laidler, and C. F. Yam, Can. J. Biochem. 57, 1200 (1979). 4~ K. J. Laidler and P. S. Bunting, this series, Vol. 64, p. 227. 42 M. A. Mazid and K. J. Laidler, Biochim. Biophys. Acta 614, 237 (1980). 43 N. J. Daka and K. J. Laidler, Can. J. Biochem. 56, 774 (1978). 44 N. J. Daka and K. J. Laidler, Biochim. Biophys. Acta 612, 305 (1980).

[6]

ALCOHOL DEHYDROGENASE COIMMOBILIZED WITH N A D

63

The mass transfer coefficient kL is a measure of the rate at which the substrate passes through a diffusion boundary layer established at the wall of the tube. There can be full diffusion control at low substrate concentrations and low flow rates, and under these conditions the concentration of the product at the exit of the tube is given by

(DL ] 2/3 [P]eD = 2.56 \r-Tvf/ [S]

(2)

At the other extreme, the exit concentration is given by 2L [p]o

VM = -rof

(3)

where VMis the inherent enzymatic reaction rate as given by an equation of the Michaelis-Menten form. The extent of diffusion control can also be evaluated on the basis of double-logarithmic plot of the dimensionless parameters qb and 19 defined as follows:

( ufr2~2/3 \~--~: p = Km,app/[S]

(4) (5)

Results and Discussion Table I summarizes results obtained from Lineweaver-Burk plots at different flow rates; the plots were linear over the concentration range of 5-50 mM ethanol. The Michaelis parameters, Km.avpand Vm, were obtained from the best fit, using the method of least squares. Included in the table are the extrapolated values of Km and V~ at infinite flow rates [Eq. (1)], i.e., under conditions when there are no diffusional effects. The value of 1.5 mM for the inherent Michaelis constant Km is significantly lower than that of 4.6 raM, found previously3~ with only the enzyme attached to the inner surface of a nylon tube. This means the present system is a more efficient one, which in turn indicates that there are fewer steric restrictions to reaction, and that the coenzyme has been suitably positioned on immobilization. Similar results to ours were obtained by Gestrelius et al. 2~ The activity of our preparation was rather low compared to that for the free enzyme and coenzyme in solution, as reported earlier by Larsson and Mosbach. 45 Our results also show that, on 45 P.-O. Larsson and K. Mosbach,

Biotechnol. Bioeng. 13, 393 (1971).

64

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[6]

TABLE I VALUES OF Kin,apt, (ETHANOL) AND Vm AT VARIOUS FLOW RATES a

Flow rate, of (cm/sec)

Km,app(ethanol) (raM)

V" (10 -9 M sec ~)

Reference

2.14 2.16 2.20 1.99 1.85 1.5 (Kin)

74.6 77.3 82.3 84.5 87.8 - 100

26 26 26 26 26 26

Enzyme and NAD coimmobilized 0.25 0.40 0.69 1.27 2.00 (extrapolated)

Enzyme immobilized; saturation with N A D o¢ (extrapolated)

4.6

~ 1000

31

a At 25.5 ° at pH 7.5.

an absolute basis, the tube with the coimmobilized enzyme and the coenzyme is less active than that used in our previous work, 31 in which only the enzyme was immobilized and the NAD was present in saturating concentrations. This is evident from the V" value of 10 -7 M sec -I which is an order of magnitude lower than that of 10 -6 M sec -1, obtained in our previous work. 3~ Such a result may be hard to improve upon, since it is obviously difficult to immobilize as much coenzyme as can be made available in a saturation experiment. We studied other methods to immobilize the coenzyme 4s-5° but none of them led to a preparation as active or stable as the one we have described. Our procedure follows that of Lindberg e t al. 3° for the synthesis of the coenzyme analog, the attachment being at the N-6 group of the adenine moiety, which in conjunction with the method of attaching the enzyme appears to be quite satisfactory. The success of the method is attributable to the fact that both enzyme and coenzyme are attached covalently to the surface of the tube, and that there are neutralizations of surface charges; the latter have a detrimental effect on enzymatic activity. 43,44,5~ Figure 1 shows a double-logarithmic plot of product concentration at 46 J. R. W y k e s , P. Dunhill, a n d M. D. Lilly, Biochim. Biophys. Acta 286, 260 (1972). 47 M. Guilford, P.-O. Larsson, and K. Mosbach, Chem. Scr. 2, 165 (1972). S. Barry and P. O ' C a r r a , FEBS Left. 37, 134 (1973). 49 R. L a m e d , Y. Levin, and M. Wilcheck, Biochim. Biophys. Acta 304, 231 (1973). 50 M. J. H a r v e y , C. R. L o w e , D. B. Craven, and D. G. Dean, Eur. J. Biochem. 41, 335 (1974). 51 p. S. Bunting and K. J. Laidler, Biochemistry 11, 4477 (1972).

[6]

65

ALCOHOL DEHYDROGENASE COIMMOBILIZED WITH N A D

-4.4

-4.6

-48

-5.0

-% oo ~ - 5 . 2

-54 20raM 10 mM 7OmM 5.0mM - 5"6 j~ _ L _ - 0,60

-045

-0.30 -0.15 IOglo(V f / c r n $'1)

0

0.15

0.30

FIG. I. Double-logarithmic plots of product concentrations at the exit against flow rates for the l-m tube, at various ethanol concentrations as indicated. From Mazid and Laidler -'6 by permission of John Wiley and Sons, Inc.

the exit against flow rate with different concentrations of the substrate ethanol. The slopes of the lines are from - 0 . 9 0 to -0.92. A slope of - 2 / 3 is expected when there is full diffusion control [Eq. (2)] and of - 1 when there is no diffusion control [Eq. (3)]. The fact that the slopes are not far from unity is thus consistent with little diffusion control. This is confirmed by the calculation of the dimensionless parameters [Eqs. (4) and (5)] as shown in Table II, the values of • being obtained using a diffusion coefficient of 4 × 10 -6 c m 2 s e c -1 for ethanol. 3],37 Inspection shows that the points lie well inside the diffusion-free region of the plot of qb against p (compare, for example, Fig. 2 of Ref. 31), since the qb values are very low as a result of the low product concentrations compared to the concentrations of the substrate [Eq. (4)]. Our previous work with the immobilized

66

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[6]

T A B L E II DIMENSIONLESS PARAMETERS qb AND p AT DIFFERENT ETHANOL CONCENTRATIONS AND FLOW RATES Parameter values at three flow rates, uf (cm sec -~) 0.25

0.69

2.00

[C2HsOH] (mM)

10 3 (])

p

10 3 qb

to

10 3 qb

p

5.0 10.0 50.0

8.5 5.0 1.2

0.43 0.21 0.04

6.6 3.9 0.9

0.44 0.22 0.04

5.3 3.0 0.7

0.37 0.19 0.04

enzyme only 31 also showed that there is little diffusion control with respect to ethanol when NAD is present at saturating concentrations, but that the slopes of the double-logarithmic plots (see Fig. l) had anomalously low negative values of -0.57 to -0.60 and -0.40 to -0.45 with different concentrations of saturating NAD. This result was accounted for by the product inhibition, which was more pronounced at lower flow rates when the products remain attached to the surface, as a result of longer residence times, and block further reaction. However, direct tests showed that there was no product inhibition in the coimmobilized system either by NADH or by acetaldehyde, the former being immediately reoxidized in the recycling mechanism and the latter being produced in the reaction in sufficiently low concentrations. Conclusions Alcohol dehydrogenase has been coimmobilized with its coenzyme, NAD. Our procedure for covalent attachment of both enzyme and a coenzyme analog inside a nylon tube showed higher efficiency compared to a similar system with only the enzyme immobilized, indicating that on coimmobilization the coenzyme has been suitably positioned with respect to the active center of the enzyme. The higher efficiency also suggests that the spacer arm through which the coenzyme is attached to the tube is of an appropriate length. On an absolute basis, however, the activity was low, and this was also observed in the coimmobilization by other methods. The activity may be improved by introducing a longer and more flexible spacer arm; any coenzyme immobilized at a distance away from the active center could then be available for enzymatic activity. Our method showed no measurable loss of enzymatic activity over a period of 1 month, and the success of the method was attributed to covalent attach-

[7]

RECYCLINGov NAD(P) By MULTIENZYMESYSTEMS

67

ment as well as to the neutralization of surface charges in the nylon tube. Apart from low activity, however, a system of the kind described would be ideally suitable as an open tubular heterogeneous enzyme reactor (OTHER) in applications such as automated analysis, with no need to add soluble coenzyme.

[7] R e c y c l i n g o f N A D ( P ) b y M u l t i e n z y m e S y s t e m s I m m o b i l i z e d b y M i c r o e n c a p s u l a t i o n in Artificial Cells

By T. M. S. CHANG Semipermeable microcapsules have been used as artificial cells encapsulating enzymes, cells, cell extracts, adsorbents, immunogens, and other biologically active materials. ~-5A number of artificial cell systems are now being applied in the treatment of patients or in biotechnology. 5 The possible applications of microencapsulated enzymes were proposed and demonstrated in animal experiments some time ago using mostly single enzyme systems. TM For more widespread applications, it will be important to have microencapsulated multienzyme systems. There is no problem in enclosing high concentrations of any number of enzyme systems within the artificial cells, since this can be done by dissolving the enzymes in the solution before microencapsulation.l-4 Enzymes immobilized within semipermeable microcapsules are in close proximity and in free solution, and a large number of different enzyme systems can be enclosed at high concentrations. In this way, cofactor in the microcapsules can interact much more freely with the multienzyme systems. Furthermore, the large surface area to volume relationship and the ultrathin membrane are such that substrates and products of the enzyme reaction can equilibrate across the membrane at a rate which is at least 100 times that of standard dialysis membrane systems. This paper is an updated discussion of our studies on the recycling of NAD(P) by multienzyme systems immobilized by microencapsulation within artificial cells. i T. M. S. Chang, Science 146, 524 (1964). 2 T. M. S. Chang, F. C. Macintosh, and S. G. Mason, Can. J. Physiol. Pharmacol. 44, 115 (1966). 3 T. M. S. Chang, "Artificial Cells." Thomas, Springfield, Illinois, 1972. 4 T. M. S. Chang, "Biomedical Applications of Immobilized Enzymes and Proteins," Vols. 1 and 2. Plenum, New York, 1977. T. M. S. Chang, "Microencapsulation including Artificial Cells." Humana, New York, 1984.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[7]

RECYCLINGov NAD(P) By MULTIENZYMESYSTEMS

67

ment as well as to the neutralization of surface charges in the nylon tube. Apart from low activity, however, a system of the kind described would be ideally suitable as an open tubular heterogeneous enzyme reactor (OTHER) in applications such as automated analysis, with no need to add soluble coenzyme.

[7] R e c y c l i n g o f N A D ( P ) b y M u l t i e n z y m e S y s t e m s I m m o b i l i z e d b y M i c r o e n c a p s u l a t i o n in Artificial Cells

By T. M. S. CHANG Semipermeable microcapsules have been used as artificial cells encapsulating enzymes, cells, cell extracts, adsorbents, immunogens, and other biologically active materials. ~-5A number of artificial cell systems are now being applied in the treatment of patients or in biotechnology. 5 The possible applications of microencapsulated enzymes were proposed and demonstrated in animal experiments some time ago using mostly single enzyme systems. TM For more widespread applications, it will be important to have microencapsulated multienzyme systems. There is no problem in enclosing high concentrations of any number of enzyme systems within the artificial cells, since this can be done by dissolving the enzymes in the solution before microencapsulation.l-4 Enzymes immobilized within semipermeable microcapsules are in close proximity and in free solution, and a large number of different enzyme systems can be enclosed at high concentrations. In this way, cofactor in the microcapsules can interact much more freely with the multienzyme systems. Furthermore, the large surface area to volume relationship and the ultrathin membrane are such that substrates and products of the enzyme reaction can equilibrate across the membrane at a rate which is at least 100 times that of standard dialysis membrane systems. This paper is an updated discussion of our studies on the recycling of NAD(P) by multienzyme systems immobilized by microencapsulation within artificial cells. i T. M. S. Chang, Science 146, 524 (1964). 2 T. M. S. Chang, F. C. Macintosh, and S. G. Mason, Can. J. Physiol. Pharmacol. 44, 115 (1966). 3 T. M. S. Chang, "Artificial Cells." Thomas, Springfield, Illinois, 1972. 4 T. M. S. Chang, "Biomedical Applications of Immobilized Enzymes and Proteins," Vols. 1 and 2. Plenum, New York, 1977. T. M. S. Chang, "Microencapsulation including Artificial Cells." Humana, New York, 1984.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

68

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[7]

Basic Methods for Immobilization of Multienzymes by Microencapsulation

Preparation of Cellulose Nitrate Membrane Microcapsules Cellulose nitrate membrane microcapsules are prepared using an updated procedure based on earlier publications. 1-7 Hemoglobin (bovine hemoglobin Type 1, 2× crystallized and dialyzed, and lyophilized, Sigma Co.), 15 g, is dissolved in 100 ml of distilled water and filtered (Whatman No. 42). Multienzymes and other materials to be microencapsulated are dissolved or suspended in 2.5 ml of this hemoglobin solution. The final pH is adjusted to pH 8.5 with Tris buffer and the hemoglobin concentration adjusted to 10 g/dl. This solution, 2.5 ml, is added to a 150-ml glass beaker. Water-saturated ether (analytical-grade ether shaken with distilled water in a separating funnel and water discarded), 25 ml, is added. The mixture is immediately stirred with a Fisher Jambo magnetic stirrer at 1200 rpm (setting of 5) for 5 sec. While stirring is continued, 25 ml of a cellulose nitrate solution is added (this solution is prepared earlier by evaporating 100 ml of USP collodion to a thin sheet and redissolving in 100 ml of a mixture of 82.5 ml ether and 17.5 ml absolute alcohol). Stirring is continued for another 60 sec. The beaker is covered and allowed to stand unstirred at 4° for 45 min. The supernatant is decanted and 30 ml of n-butyl benzoate added. The mixture is stirred for 30 sec at the same magnetic stirrer setting. The beaker is allowed to stand uncovered and unstirred at 4° for 30 rain. Then the butyl benzoate is removed completely after centrifugation at 350 g for 5 min. A Tween 20 (Atlas Powder Co.) solution, 25 ml, at 50% (v/v) concentration, pH 7.0, is added. Stirring is started at a setting of 10 for 30 sec. Water, 25 ml, is added and stirring continued at a setting of 5 for 30 sec; then 200 ml of water is added. The supernatant is removed and the microcapsules are washed 3 more times with 200 ml of a 1% Tween 20 solution, pH 7. The microcapsules are then suspended in a suitable buffer. In properly prepared microcapsules, there should not be leakage of hemoglobin after the preparation. Hemoglobin at a concentration of 10 g/dl is necessary for successful preparation. Furthermore, this high concentration of protein stabilizes the enzymes during the preparation, reaction, and storage. 8 However, crude hemoglobin contains enzymes which inactivate NADH and unless highly purified hemoglobin is used, purification using affinity chromatography on 6 T. M. S. Chang, this series, Vol. 44, p. 201. 7 y . T. Yu and T. M. S. Chang, Int. J. Biomat. Med. Devices Artif. Organs 8, 273 (1980). 8 T. M. S. Chang, Biochem. Biophys. Res. Commun. 44, 1531 (1971).

[7]

RECYCLING OF N A D ( P ) BY MULTIENZYME SYSTEMS

69

an NAD+-Sepharose column is required. 9 The long-term stability of microencapsulated multienzyme activity can be greatly increased by crosslinking with glutaraldehyde, 8 but this is done at the expense of greatly reduced initial enzyme activity.

Preparation of Polyamide Membrane Microcapsules by lnterfacial Polymerization Polyamide membrane microcapsules of 100 /zm mean diameter are prepared using an updated method based on earlier methods. 1-6:0-14 Just before use the following two solutions are prepared. (l) Terephthaloyl chloride (ICN K & K Inc.), 100 mg, is added to 30 ml of an organic solution (chloroform : cyclohexane, 1 : 4) kept in an ice bath. This is covered and stirred with a magnetic stirrer for 4 hr and then filtered with Whatman No. 7 paper. (2) A diamine-polyethyleneimine solution is prepared by dissolving 0.378 g NaHCO3 and 0.46 g 1,6-hexadiamine (J. T. Baker Chemical Co.) in 5 ml distilled water, and adjusting the pH to 9 with 6 N HCI. Then 2 ml 50% polyethyleneimine (ICN K & K Inc.) is added to the diamine solution, the pH readjusted to 9, and the final volume made up to I0 ml with distilled water. The hemoglobin solution, 10 g/dl, is prepared as described above for cellulose nitrate microcapsules. Multienzymes and other materials are added to the hemoglobin solution. Two and one-half milliliters of this hemoglobin solution and 2.5 ml of the diamine-polyethyleneimine solution are mixed for 10 sec in a 150-ml beaker placed in an ice bath. A 0.5% (v/v) Span 85 (Atlas Powder Co.) organic solution (chloroform : cyclohexane, 1 : 4), 25 ml, is added and stirred in the Fisher Jambo magnetic stirrer at speed setting of 2.5 for 60 sec. The terephthaloyl chloride solution prepared earlier is added (25 ml) and the reaction is allowed to proceed for 3 rain with the same stirring speed. The supernatant is discarded and another 25 ml of the terephthaloyl chloride solution is added. The reaction is carried out with stirring for another 3 rain. The supernatant is discarded. Then 50 ml of the 0.5% Span 85 chloroform-cyclohexane solution is added and stirred for 30 sec. The supernatant is discarded. After this the procedure with Tween 20 as describd for cellulose nitrate micro9 j. Grunwald and T. M. S. Chang, J. Appl. Biochem. 1, 104 (1979). 10j. C. W. Ostergaard and S. C. Marting, Biotechnol. Bioeng. 15, 561 (1973). " R. B. Aisina, N. F. Kazanskafa, E. V. Lukasheva, and V. Berezin, Biokhimiya 41, 1656 (1976). 12j. Grunwald and T. M. S. Chang, Biochem. Biophys. Res. Commun. 81, 565 (1978). 13 j. Grunwald and T. M. S. Chang, J. Mol. Catal. 11, 83 (1981). ]4 y. T. Yu and T. M. S. Chang, Enzyme Microb. TechnoL 4, 327 (1982).

70

MULTISTEP ENZYME SYSTEMS AND COENZYMES

/

\

/

ALCOHOL

x

ethanoI-"'-/-',-DEHYDROCENASE

+,

/+

NAD ~'T'~NAD

malate

[7]

\

,.

- acetaldehyde

NAD H "q------'~NADH

~ -.~ DEHYDROGENAS E y

oxaloacetate

%

FIG. 1. High concentrations of yeast alcohol dehydrogenase and malate dehydrogenase immobilized in solution within semipermeable microcapsules. NAD + and NADH can equilibrate rapidly across the membrane to be recycled in the microcapsules.

capsules is used here for the transfer of the microcapsules into the buffer solution. Failure in preparing good microcapsules is frequently due to diamine or diacids which have been stored after they have been opened. A new, unopened bottle will usually solve the problems. Unlike the cellulose nitrate microcapsules, in interfacial polymerization the hemoglobin solution can be replaced by a 10% polyethyleneimine solution adjusted to pH 9.u However, the microcapsules prepared without hemoglobin may not be as sturdy. Cross-linking with glutaraldehyde could also be carried out. 8 As stated before, this markedly decreases the initial enzyme activity, though the long-term stability of the remaining enzyme activity after cross-linking with glutaraldehyde is greatly increased.

Recycling of Free NADH and NADPH Recycling of Free N A D + and N A D H by Microencapsulated Yeast Alcohol Dehydrogenase and Malate Dehydrogenase us'n6 (Fig. I) Yeast alcohol dehydrogenase, 2.8 rag, and 0.22 mg of malate dehydrogenase are dissolved in 2.5 ml of the hemoglobin solution prepared as described under the basic methods, and then immobilized within cellulose nitrate microcapsules by the procedure described above. The final preparation is suspended in 0.1 M phosphate buffer, pH 8.0 at 4 °. Recycling of free NAD + is studied by suspending the microcapsules in the presence of 0.0196 mM NAD +, 0.217 M ethanol, 0.981 mM oxaloacetic acid, and 0.1 M phosphate, pH 8.0. Recycling of free NAD + is determined as the production of malic acid as follows. Reactions are ~5 j. Campbell and T. M. S. Chang, Biochim. Biophys. Acta 397, 101 (1975). t6 j. Campbell and T. M. S. Chang, Biochem. Biophys. Res. Commun. 69, 562 (1976).

[7]

RECYCLING OF

urea

/ /

+/

glucose '

BY MULTIENZYME SYSTEMS

~ UREASE .~

I

i glutamates------

NAD(P)

x. NH3

7

~

GLUTAMATE DEHYDROCENASE 9

/;+

71

NH 3

~ ketoglutarate

,,

t

~ ,-~1 GLUCOSE / ~ DEHYDROCENASE

-\ ,~ / pyruvate~CLUTAMATE-PYRUVAT~ TRANSANINASE ~

\

',glucuronate /

) /

/ ketoglutarate

--

+ alantne

FIG. 2. High concentrations of four enzymes (urease, glutamate dehydrogenase, glucose dehydrogenase, and glutamate pyruvate transaminase) immobilized in solution within semipermeable microcapsules. Substrates and NAD ÷ can equilibrate rapidly across the membrane for the multistep enzyme reaction in the conversion of urea to ammonia to glutamic acid to alanine.

allowed to be carried out and then terminated after preset time intervals by heating the reaction mixture at 100° for 5 min. The reaction mixture is centrifuged, and 0.5 ml of the supernatant is used in the following assay for malic acid. 17 The supernatant, 0.5 ml, is added to 1.5 ml NAD ÷ (2 raM), 0.5 ml 30 mM glutamic acid, and 1.0 ml 0.1 M glycine/NaOH, pH 10.0. The change in absorbance at 340 nm after the addition of 5 ttl glutamate oxaloacetate transaminase (aspartate aminotransferase) and 25 /zl malate dehydrogenase is noted. This procedure is standardized by subjecting known amounts of malic acid to the same assay. Recycling of free NADH is studied in a similar fashion to NAD + except that 0.0196 mM NADH is used instead of NAD +. Recycling of N A D H in the Conversion of Urea and Ammonia into Amino Acid w'-2° (Fig. 2) The hemoglobin solution (3 ml of 10 g/100 ml solution) is prepared as described under the basic method, but in addition containing 1.0 mg of urease (Worthington Biochemical Corp., 60 U/mg), 16 mg of L-glutamate dehydrogenase (Sigma Chemical Co., 50 U/mg, 10 mg/ml), 0.16 mg of glucose dehydrogenase (Boehringer-Mannheim, 150 U/mg), 4.0 mg of glutamate pyruvate transaminase (aspartate aminotransferase) (Sigma 17 T. t8 T. ,9 T. 2o K.

Kato, S. J. Berger, J. A. Carter, and O. H. Lowry, Anal. Biochem. 53, 83 (1973). M. S. Chang and C. Malouf, Artif. Organs 3, 38 (1979). M. S. Chang, C. Malouf. and E. Resurreccion. Art([] Organs (Suppl.) 3, 284 (1979). F. Gu and T. M. S. Chang, Biomat. Artif. Cells Artif. Organs 15, 289 (1987).

72

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[7]

Chemical Co. 91 U/mg), and 20 mg of NADH (Sigma Chemical Co.). NADH is used to stabilize glucose dehydrogenase during preparation and is dialyzed out of the microcapsules after preparation. The remaining steps for the microencapsulation of hemoglobin-enzyme solution are the same as described under the basic method for cellulose nitrate membrane microcapsules. Ammonia substrate solution is made up of glucose (1 g/dl), KCI (5 mM/liter), MgC12 (5 mM/liter), ammonium acetate (20 mM/liter), a-ketoglutarate (20 mM/liter), pyruvic acid (20 mM/liter), NADH (0.5 mM/ liter), and ADP (0.1 mM/liter). For urea assay, the same solution as above is used except that ammonium acetate is replaced by urea (50 mg/dl). Our results show that this multienzyme system is effective in the sequential conversion of urea into ammonia, then ammonia into glutamic acid, and finally glutamic acid into another amino acid, alanine (Fig. 2). The conversion of urea into ammonia takes place extremely rapidly, nearly all the urea being converted into ammonia within 10 min. However, the rate of conversion of ammonia sequentially into glutamic acid and then into alanine was 0.3 ixM/ml microcapsules per minute. This is not likely due to the delay in the recycling of NADH in the formation of glutamic acid, since when microcapsules were used to convert ammonia into glutamic acid only, the rate of formation of glutamic acid was 2.4 tzM/ ml microcapsule per minute. When the microcapsules were used to convert glutamic acid into alanine only, the rate was only 0.8 tzM/ml microcapsules per minute even with a 20 mM/liter glutamic acid substrate solution. Further analysis showed that by increasing the amount of transaminase the rate of conversion can be increased to 1.80 tzM/ml microcapsules per minute. Thus by optimizing the enzyme concentrations and the other reaction rate, a much more efficient system is possible. The high rate of recycling of NADH was made possible only by the use of a very high specific-activity glucose dehydrogenase (150 U/mg). Microencapsulated multienzyme systems with NAD-dextran can convert ammonia and urea into essential amino acids. 2°

Recycling of NADPH in the Sequential Conversion of Urea and Ammonia into Amino Acid 2j (Fig. 3) Urease (10 rag), glutamate dehydrogenase (8 mg), and glucose-6-phosphate dehydrogenase (2 mg) are dissolved in 1.5 ml of hemoglobin solution. This is then microencapsulated using the procedure of interfacial polymerization described under the basic methods. The rate of formation of 6-phosphogluconate is used to measure the 21 j. C o u s i n e a u and T. M. S. Chang, Biochem. Biophys. Res. Commun. 79, 24 (1977).

[7]

RECYCLINGOF NAD(P) BY MULTIENZYMESYSTEMS / .

urea

' I

/

u~ 3

GLUTAMATE DEHYDROCENASE :

+l

/

NADP ~ - - ~ N A D P

~ \

"--

\

glutamate-I

glucose-6-P

"% ~ U R E A S E ~~H ~3N~

N

+

~ ketoglutarate

t

NADPH ~

GLUCOSE-6-P / DEIIYDROGENASE

73

I

NADPH

/ -~ 6 - p h o s p h o g l u c o n a t e /

/

FIG. 3. High concentrations of three enzymes (urease, glutamate dehydrogenase, glucose-6-phosphate dehydrogenase) for the multistep enzyme reaction of conversion of urea into the amino acid glutamic acid.

cyclic regeneration ofNADP + : NADPH. The microcapsules, 100/~1, containing the muitienzyme system are suspended in 3.0 ml of 0. I M phosphate, pH 7.5, containing 10 mM of urea, 4 mM o f a-ketoglutarate, 0. I mM of ADP, and 5 mM of glucose 6-phosphate, followed I rain later by the addition of the cofactor NADPH at 0.21 mM to initiate the cyclic process. Aliquots of the Suspension are withdrawn at various intervals and heated at 100° for 2 min, and the filtrate analyzed for 6-phosphoglucohate as follows. The filtrate (0.5 ml) is added to 1.0 ml of 0.1 M phosphate, pH 7..5, containing 6-phosphogluconate dehydrogenase (0.15 units) and 2 mM NADP + at 37°. After 30 min of incubation, the observed increase of the absorbance at 340 nm is noted and read on a standard curve. The rate of conversion of urea into glutamate and the recycling of NADPH could be followed by the rate of formation of 6-phosphogluconate, the decrease in urea level, or the formation of glutamic acid. Ammonium acetate could be used in this reaction instead of urea. Glutamate is formed at the same rate irrespective of whether urea or ammonium acetate is used as substrate. This suggests that the rate-limiting step in this sequential reaction is in the conversion of ammonia into glutamic acid. Recycling of NAD + in Stereospecific Steroid Oxidation 22 (Fig. 4) Another application is demonstrated in Fig. 4. Here microcapsules containing 3cz-hydroxysteroid dehydrogenase and the bacteria Leuconostoc mesenteroides with NADH oxidase have been used in stereospecific 22 F. Ergan, D. Thomas, and T. M. S. Chang, Appl. Biochem. Biotechnol. 10, 61 (1984).

74

MULTISTEP ENZYME SYSTEMS AND COENZYMES

/

andros

t erone

\

t3a-HYDROXYSTEROID~ ----~ DEHYDROCENASE

i

,,0++ d o ÷ , water

~

[7]

\ ~

NADH

,!

/ OXIDASE

~androstanedione

•

NADH oxygen

I

j

FIG. 4. Recycling of NAD ÷ in stereospecific steroid oxidation. Here the bacteria Leuconostoc mesenteroides as the source of NADH oxidase is microencapsulated with 3a-

hydroxysteroid dehydrogenase. steroid oxidation. 22 The substrate required to recycle N A D + is oxygen, which is c o n v e r t e d into water. Recycling of D e x t r a n - N A D H Retained within Microcapsules 9A2A3 (Fig. 5) The a b o v e studies show that microencapsulated multienzyme s y s t e m s can carry sequential e n z y m e reactions and can recycle the required cofactors. F o r in vitro studies, cofactors in the external substrate solution can freely equilibrate across the microcapsule m e m b r a n e to be recycled in the microcapsules. H o w e v e r in applications for medicine or in industrial reactors, it will be important to retain the cofactor inside the microcapsules for recycling. This w a y , it will not be necessary to have cofactors supplied f r o m outside the microcapsules. Studies were therefore carried out here to covalently link cofactor to macromolecules, and retain these inside microcapsules for recycling 9a2a3 (Fig. 5).

/ ethanol

\

¢ ALCOHOL " ~ DEHYDROCENASE DEXTRAN

,\ malate

. ~

SAD+

~ ~ acetaldehyde

DEXTRAN

I

SADH )!

/,

MALATE DEHYDROCENASE~"

I oxaloacetate

FIG. 5. Recycling of dextran-NADH retained within semipermeable microcapsules.

[7]

RECYCLING OF N A D ( P ) BY MULTIENZYME SYSTEMS

75

Preparation of Dextran-NAD + NAD+-N6-[N-(6-aminohexyl)acetamide] is coupled to dextran T7023 with slight modifications. 12 NAD + could also be coupled to polyethyleneimine, 16 albumin, 24 or hemoglobin 24 and immobilized in the microcapsules. 16,24

Preparation of Cellulose Nitrate Membrane Microcapsules Containing Dextran-NAD and Multienzyme Systems 9 Dextran-NAD + solution (1.5 ml) containing 3-5 mg of NAD + coupled to 150 mg of dextran TT0 is mixed with 1.5 ml of hemoglobin solution (15 g/dl prepared as described under the procedure of cellulose nitrate membrane microcapsules). The pH is adjusted to 8.5 and 30 mg of yeast alcohol dehydrogenase (300 U/mg) and 100/zl of malate dehydrogenase (1870 U/100/xl) are added to the hemoglobin-dextran-NAD + solution. The pH is again adjusted to 8.5. The microcapsules are then prepared by the method described in the earlier section for cellulose nitrate membrane microcapsules. The microcapsules prepared are stored at 0° in phosphate buffer (0.1 M, pH 8.0).

Preparation of Polyamide Membrane Microcapsules Containing Dextran-NAD ÷ and Multienzyme Systems 13 The microcapsules are prepared as described under basic procedure of preparation using interfacial polymerization. Dextran-NAD + (200 mg), containing 5 mg of immobilized NAD ÷, is dissolved in 1.5 ml of the specially prepared hemoglobin solution. Yeast alcohol dehydrogenase (10 mg of 305 U/mg) and malate dehydrogenase (1.65 mg of 1130 U/mg) are added and dissolved in this solution. The resulting solution is mixed with 1.5 ml of the diamine-polyethyleneimine solution in a 150-ml beaker kept in an ice bath. The rest of the procedure is as described under the section on basic procedure. The resultant microcapsules are suspended in 0.1 M phosphate buffer and stored at 4°.

Measurements of the Recycling of Dextran-NAD + within the Microcapsules This could be carried out in two ways. One way is to carry out stirbatch studies as described for the recycling of free NAD + in the earlier section. Alternatively, recycling activity measurements are made by us23 K. Mosbach, P. O. Larsson, and C. R. Lowe, this series, Vol. 44, p. 859. 24 Ho P. Wahl and T. M. S. Chang, J. Mol. Catal. 39, 147 (1987).

76

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[7]

ing a continuous flow shunt chamber prepared from a microcolumn (0.5 × 3 cm) containing 0.5 ml of microcapsules. 9,13Glass wool is placed at both ends of the column to retain the microcapsules and to allow the passage of fluid. A solution of Tris buffer (0.1 M, pH 8.0) containing the substrates (200 mM ethanol and 5 mM oxaloacetate) is passed through the column at a flow rate of 0.1 ml/min and the effluent is collected in 0.5-ml fractions, which are then boiled at 100° for 5 min to destroy excess oxaloacetate and cooled in an ice bath. The concentration of malate in the column effluent is measured. Results The recycling activity as followed by the production of malate is shown in Table I. The microcapsules contained about 1 mg of immobilized NAD ÷ per 1 ml of microcapsules. Thus, in the continuous flow reactor, no cofactor is required in the substrate solution perfusing the column. In the microcapsules, prepared using purified hemoglobin as described, both the recycling activity and the storage stability are severalfold higher than in the case of the microcapsules prepared using crude hemoglobin. The recycling activity of the shunt is high, 65% of the oxaloacetic acid being converted into malic acid by one passage through the shunt. The dextranNAD ÷ present within the microcapsules is regenerated two times each minute during the reaction. The shunt showed good stability, the reaction rate remaining constant for the first hour and 83% of the original activity TABLE I RECYCLING OF DEXTRAN-NAD + WITHIN CELLULOSE NITRATE MEMBRANE MICROCAPSULES N A D ÷ regenerated

Sample Crude Hb

Purified Hb

Glutaraldehyde-Hb

Storage time (days)

No. of cycles per hour

Quantity in microcapsules (/zmol/hr/ml)

0 1 2 5 0 I 4 7 0 7

7.0 4.2 2.8 1.05 120.0 117.5 84.0 37.3 4.9 4.0

9.7 6.0 4.0 1.5 150.0 147.0 105.0 46.5 6.0 4.8

[7]

RECYCLING OF

NAD(P)

BY MULTIENZYME SYSTEMS

77

being retained even after 3 hr of continuous reaction. The same shunt could be stored and reused several times. The cross-linked microcapsules showed complete stability during 3 hr of continuous reaction although their initial activity was only about 10% of that of the untreated microcapsules. In the case of the polyamide membrane microcapsules, the dextranNAD ÷ within the microcapsules was regenerated twice each minute during the continuous reaction of the shunt. The stability of the recycling activity was very good. After the recycling activity reached its maximal value, it remained stable at this value and no significant decrease in activity was observed during the 3 hr of reaction. After this reaction, each shunt was stored at 4 ° . Sixty-three percent of the original activity was retained after 7 days and 41% was still retained after 12 days. The microcapsules could be further stabilized by cross-linking with glutaraldehyde but this cross-linking caused a sharp decrease in the initial recycling activity within the microcapsules. Both yeast alcohol dehydrogenase and malate dehydrogenase are much more stable in the microencapsulated form when compared with the same enzymes in solutions. Stability of microencapsulated malate dehydrogenase is much higher than that of yeast alcohol dehydrogenase. Furthermore, there is a correlation between the decrease in the microencapsulated dextran-NAD + recycling activity and the decrease in the activity of the microencapsulated yeast alcohol dehydrogenase. The stability of the recycling activity of the microcapsules is therefore related to the stability of yeast alcohol dehydrogenase. Recycling of Free NADH Retained within Lipid-Polymer Membrane Microcapsules j4'25-28 (Fig. 6) Cofactors covalently linked to macromolecules such as dextran or polyethyleneimine can be retained within semipermeable microcapsules to be recycled enzymatically. However, linkage of cofactors to macromolecules increases steric hindrance and reduces their rate of reactions with enzymes. In biological cells such as erythrocytes, free cofactors and multienzymes systems are all retained within the cells in free solution. Thus studies were carded out here to immobilize free cofactors inside microcapsules with membranes impermeable to cofactors but permeable to the initial substrates. In this way, the free cofactor can function without steric 25 y . 26 y . 27 y . 2s E.

T. Yu and T. M. S. Chang, FEBS Lett. 125, 94 (1981). T. Yu and T. M. S. Chang, Trans. Am. Soc. Artif. Intern. Organs 27, 535 (1981). T. Yu and T. M. S. Chang, Enzyme Eng. 6, 163 (1982). Ilan and T. M. S. Chang, Appl. Biochem. Biotechnol. 13, 221 (1986).

78

MULTISTEP ENZYME SYSTEMS AND COENZYMES

urea

[7]

NH 3

Itpld-polymer membrane

alcohol

acetaldehyde

FIG. 6. Recycling of free NADH retained within microcapsules having a lipid-polymer membrane which is impermeable to NADH. The permeability of the lipid-polymer membrane can be adjusted.

hindrance in close proximity to the enzymes. Furthermore, all enzymes and cofactors inside the microcapsules are in free solution. Lipidpolyamide membrane microcapsules have been prepared. 3,4 These are permeable to lipophilic molecules but have negligible permeability to hydrophilic molecules as small as K + and Na+. 29 Preparation ~4

Lipid-polyamide microcapsules of 100 tzm mean diameter containing multienzyme systems, cofactors, and o~-ketoglutarate are prepared as described. 14The first part is similar to the procedure described earlier in this paper under basic procedure for the preparation of polyamide microcapsules. To 2.0 ml of the hemoglobin solution prepared as described earlier in this chapter is added 12.5 mg glutamate dehydrogenase (bovine liver, type III, 40 U per mg, Sigma Co.), 6.25 mg alcohol dehydrogenase (yeast, 330 U per mg, Sigma Co.), 0.5 mg urease (51 U/mg, Millipore Co.), 1.18 mg ADP, and either NAD + (0.52, 105, 2.11, or 21.13 mg) or NADH (21.13 mg) dissolved in 0.25 ml of water. Finally, 0.25 ml of a solution containing ot-ketoglutarate (56.5 rag), MgClz (2.5 mg), KCI (0.93 rag) is added. Two and one-half milliliters of the hemoglobin-enzyme solution so prepared is added to 2.5 ml of the diamine-polyethyleneimine solution prepared as described in the basic method section. The remaining steps are the same as described except that the Tween 20 steps are omitted 29 A. M. Rosenthal and T. M. S. Chang, J. Membr. Sci. 6, 329 (1980).

[7]

RECYCLINGOF NAD(P) av MULTIENZYMESYSTEMS

79

here. Instead, after a wash with the 0.5% Span 85 organic solution as described, the following steps are carried out to apply the lipids to the polyamide membranes. The microcapsules are rinsed twice with 10 ml of a lipid-organic liquid. (This is prepared earlier as follows: 1.4 g lecithin and 0.86 g cholesterol are added to 100 ml tetradecane and stirred for 4 hr at room temperature. If a more permeable lipid membrane is required to allow urea to diffuse across, then the lipid compositions should be 0.43 g cholesterol and 0.7 g lecithin.) Then another 10 ml of the lipid-organic liquid is added and the suspension is slowly rotated for 1 hr at 4° on a multipurpose rotator. After this, the supernatant is decanted, and the lipid-polyamide membrane microcapsules are recovered and left in this form at 4 ° without being suspended in aqueous solution until it is added to the substrate solution just before the reaction. The procedure should be practiced and the microcapsules prepared should be tested for absence of leakage of enzymes or cofactors before being used in experimental studies.

Assay Method The microcapsules so prepared can act on either ammonia or urea. For assaying using ammonia, a substrate solution is made up to contain Tris (100 mM), ammonium acetate (20 mM), and alcohol (200 raM) with a final pH adjusted to 9. At pH 9, about 38% of ammonia is present in the nonionized form (NH3), which can cross the lipid-polyamide membrane. At pH 7.4, most of the ammonia is present in the ionized form (NH4+), which cannot cross the lipid-polyamide membrane. For assaying using urea, urea (50 mg/dl) replaces ammonium acetate in the above substrate solution. In each experiment, 2 g lipid-polyamide microcapsules is added to 4 ml of the substrate solution in a stoppered flask. The reaction is carried out in a rotary shaker at 30° and 160 rpm. At present intervals, 0.4-ml aliquots of the supernatant are taken and 0.04 ml 0.5 N HC1 is added to each. These are then heated in boiling water for I min and the ammonia and/or urea concentrations measured using a Technicon autoanalyzer. Since glutamate formed inside the microcapsules is retained inside, the microcapsules are disrupted and the glutamate measured by HPLC.

Results With cofactor recycling, 0.25 #mol NAD ÷ per 2 g of microcapsules was sufficient to convert 10/~mol of ammonia into glutamate. The number of recyclings was 40 in 3 hr (Table II). By increasing the amount of NAD + in the microcapsules from 0.25 to 0.5, 1.0, and 10/.Lmol, the reaction rates

80

[7]

MULTISTEP ENZYME SYSTEMS AND COENZYMES

TABLE II COFACTOR RECYCLING WITH DIFFERENT AMOUNTS OF NAD÷ RETAINED WITHIN LIPID-POLYAMIDE MEMBRANE MICROCAPSULES

NAD÷ (/~mol) in 2 ml microcapsules 0.25 0.50 1.00 10.00

Number of cycles after 45 min 10 16 8 0.8

-+ 10 -+ 3.2 -+ 1.6 -+ 0.2

90 min 26 -+ 19 -+ 9.6-+ 1.2 ---

11 4 2 0.2

135 min 32 21 12 1.3

-+9 _+ 4 -+ 1.2 -+ 0.2

180 min 40 22 13 1.5

-+9 _+ 4 -+ 1.2 -+ 0.2

were only increased slightly. With recycling of 0.5/xmol NAD ÷ per 2 g of microcapsules, the rate of conversion of ammonia into glutamate is comparable to the use of 10/xmol NADH per 2 g of microcapsules with no recycling. When 10 ~mol of NADH per 2 g microcapsules was used without recycling, the reaction stopped as soon as 10 ~mol ammonia was converted into glutamate. However, when 2 g microcapsules containing 0.25, 0.5, 1.0, or 10/~mol NAD ÷ with recycling were used, the conversion continued even after 10/xmol ammonia was converted into glutamate by each 2 g of microcapsules. By decreasing the proportion of cholesterol in the lecithin-cholesterol lipid solution used for complexing, it was possible to obtain a membrane with good permeability to urea but with no leakage of NAD ÷ and otketoglutarate. When microcapsules containing urease and the multienzyme system with NAD ÷ recycling were used, ammonia formed from urea was converted into glutamate. However, since the rate of conversion of urea into ammonia was faster than the rate of conversion of ammonia into glutamate, the ammonia level increased before it fell as ammonia was converted more slowly into glutamate. The glutamate formed is retained within the microcapsules, which can be disrupted and the glutamate measured by HPLC as shown in Table III. The membrane of the standard lipid-polyamide membrane microcapsules was not permeable to urea, so that no glutamate was detectable inside the microcapsules. When the permeability of the lipid-polyamide membranes was increased by the methods described above, the rate of conversion of urea into ammonia and then into glutamate increased with increasing permeability (Table III). The lipid-polyamide membrane retains enzymes, unmodified cofactors, and a-ketoglutarate in solution within the microcapsules (Fig. 5). Ammonia and urea equilibrating into the microcapsules are converted

[7]

RECYCLINGOF NAD(P) BY MULTIENZYMESYSTEMS

81

TABLE Ill CONVERSION OF UREA INTO GLUTAMIC ACID WITH NADH RECYCLING IN LIPID-POLYAMIDE MEMBRANE MICROCAPSULES

Permeability to urea

Glutamic acid (#tool/liter in 3 hr)

Low Medium High

0 2.42 6.93

into glutamate. Alcohol entering the microcapsules acts as substrate for the conversion and recycling of NAD + to NADH. Thus cofactors can be immobilized without modification. Even some substrates, e.g., t~-ketoglutarate, can also be immobilized. Since none of the cofactors or enzymes is immobilized directly to solid polymers, there is no diffusion restriction or steric hindrance within the "intracellular" confines of the microcapsules. A very small amount of NAD + is required when the cofactor is recycled. A larger amount of NADH can be microencapsulated for the reaction if recycling of cofactor cannot occur, so that the enzymes and cofactors may all be included in the microcapsules. The required substrate, aketoglutarate, can also be retained within the microcapsules. Thus, when used in a microcapsule column for perfusion, cofactors or a-ketoglutarate need not be added to the perfusate. The product, glutamate, can be used as the initial material for another enzymatic reaction. Microcapsules containing more complex multienzyme systems can convert ammonia into glutamate which is then converted into other amino acids (e.g., alanine) by transaminases inside the same microcapsule.19,28 The a-ketoglutarate formed during the transamination reaction can be recycled in the reaction. 19,28 Summary Multistep enzyme systems can be immobilized in solution within semipermeable microcapsules. With the ability to recycle cofactors, a number of potentially useful systems have been made possible. Furthermore NAD + can be retained inside the microcapsules by two approaches. (l) NAD + can be linked to macromolecules such as dextran or polyethyleneime. However, in this form, there are significant increases in steric hindrance and diffusion restrictions. (2) "Artificial cells" consisting of

82

M U L T I S T E P E N Z Y M E SYSTEMS A N D C O E N Z Y M E S

[8]

lipid-polyamide membrane microcapsules containing multienzyme systems, cofactors, and substrates can retain NAD + in the free form. Analogous to the intracellular environments of red blood cells, free NAD + in solution inside the microcapsules is effectively recycled by the multistep enzyme systems which are also in solution. Enzymes in the microcapsules are in high concentrations and in close proximity to one another. Any number and any concentration of different enzyme systems can be microencapsulated all within one artificial cell, within the limit of solubility of the total amount of enzymes. Products of sequential reactions inside the microcapsules are at much higher concentrations than outside. All these factors result in an optimal intracellular environment for multistep enzyme reactions. External substrates in the form of lipophilic or small hydrophilic molecules can equilibrate across the membrane to participate as initial substrates in the multistep reactions in the microcapsules. A number of potential applications are possible using this approach. The lipid-polyamide membrane artificial cell can also be used in basic research as a biochemical cell model for the simpler types of biological cells such as erythrocytes.

[8] Bioluminescent Assays U s i n g C o i m m o b i l i z e d E n z y m e s By G. WIENHAUSEN, L. J. KRICKA, and M. DELUCA The present article will discuss the uses of luciferases coimmobilized with other enzymes for various analytical purposes. The quantitative determination of metabolites and enzymes is coupled to the luciferase reactions which are measured as light production. The assays are sensitive, rapid, and specific. Two different luciferases have been used: the firefly (Photinus pyralis) luciferase (EC 1.13.12.7, Photinus-luciferin 4-monooxygenase) and the bacterial (Vibrio harveyi) luciferase (EC 1.14.14.3, alkanal, reduced-FMN : oxygen oxidoreductase). Firefly luciferase catalyzes the ATP-dependent oxidative decarboxylation of luciferin (LH2) with the emission of light [reaction (1)]. LH2 + ATP + 02

luciferase, L ~ O

+ A M P + PP~ + CO2 + hv

(1)

Any enzyme or substrate which is coupled to ATP production or disappearance can be assayed with firefly luciferase, as shown in reaction (2). X-P + ADP

METHODS IN ENZYMOLOGY, VOL. 136

E

, ATP + X

(2)

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

82

M U L T I S T E P E N Z Y M E SYSTEMS A N D C O E N Z Y M E S

[8]

lipid-polyamide membrane microcapsules containing multienzyme systems, cofactors, and substrates can retain NAD ÷ in the free form. Analogous to the intraceUular environments of red blood cells, free NAD ÷ in solution inside the microcapsules is effectively recycled by the multistep enzyme systems which are also in solution. Enzymes in the microcapsules are in high concentrations and in close proximity to one another. Any number and any concentration of different enzyme systems can be microencapsulated all within one artificial cell, within the limit of solubility of the total amount of enzymes. Products of sequential reactions inside the microcapsules are at much higher concentrations than outside. All these factors result in an optimal intracellular environment for multistep enzyme reactions. External substrates in the form of lipophilic or small hydrophilic molecules can equilibrate across the membrane to participate as initial substrates in the multistep reactions in the microcapsules. A number of potential applications are possible using this approach. The lipid-polyamide membrane artificial cell can also be used in basic research as a biochemical cell model for the simpler types of biological cells such as erythrocytes.

[8] Bioluminescent Assays U s i n g C o i m m o b i l i z e d E n z y m e s By G. WIENHAUSEN, L. J. KRICKA,and M. DELUCA The present article will discuss the uses of luciferases coimmobilized with other enzymes for various analytical purposes. The quantitative determination of metabolites and enzymes is coupled to the luciferase reactions which are measured as light production. The assays are sensitive, rapid, and specific. Two different luciferases have been used: the firefly (Photinus pyralis) luciferase (EC 1.13.12.7, Photinus-luciferin 4-monooxygenase) and the bacterial (Vibrio harveyi) luciferase (EC 1.14.14.3, alkanal, reduced-FMN : oxygen oxidoreductase). Firefly luciferase catalyzes the ATP-dependent oxidative decarboxylation of luciferin (LH2) with the emission of light [reaction (1)]. LH2 + ATP + 02

luciferase

~ L=O

+ A M P + P P t + CO2 + hv

(1)

Any enzyme or substrate which is coupled to ATP production or disappearance can be assayed with firefly luciferase, as shown in reaction (2). X-P + ADP

METHODS IN ENZYMOLOGY, VOL. 136

E

, ATP + X

(2)

Copyright © 1987by Academic Press, Inc. All rights of reproduction in arty form reserved.

[8]

B I O L U M I N E S C E N T ASSAYS

83

Either the concentration of the phosphorylated compound X-P or the specific kinase (E) can be determined. Bacterial luciferase catalyzes the oxidation of FMNH2 and a longchain aldehyde, a reaction which also results in light emission. F M N H z + R C H O + O2

luciferase

> F M N + R C O O H + H20 + hv

(3)

This reaction by itself is of limited analytical usefulness unless it is coupled to the production of FMNH2, a reaction which is catalyzed by an NADH- or NADPH-specific oxidoreductase [Eqs. (4) and (5)]. NADH + H + + FMN NADPH + H + + FMN

oxidoreductase oxidoreductase

~ N A D + + FMNH2

(4)

~ N A D P ÷ + FMNH2

(5)

By coupling the luciferase with the oxidoreductase, any dehydrogenase which produces N A D H or NADPH can be assayed, as shown in reaction (6). XH~ + NAD(P) +

E

, X + NAD(P)H + H +

(6)

Either XH2, the reduced substrate, or the enzyme catalyzing the oxidation of XH2 can be measured. What are the advantages of using coimmobilized enzymes for these assays? In general the enzymes are more stable than the soluble forms. The sensitivity of the assays is greatly increased due to the fact that the immobilized luciferase is present in a microenvironment with locally high concentrations of ATP or FMNH2. For example, coimmobilized N A D H : FMN oxidoreductase and bacterial luciferase produce about 100 times more light per picomole of NADH than comparable amounts of soluble enzymes. Another important aspect is that immobilized enzymes can be incorporated into flow cells where they can be used for multiple assays. Source

of the Enzymes

The experiments reported here have all been performed with enzymes purified in our laboratory. Firefly luciferase is purified as described in Vol. LVII of this series.~ Bacterial luciferase is purified from Vibrio harveyi according to Hastings et al. z The oxidoreductases are purified also from V. harveyiJ There is a diaphorase from microorganisms available 1 M. D e L u c a and W. D. McElroy, this series, Vol. 42, p. 3. 2 j. W. Hastings, T. O. Baldwin, and M. Z. Nicoli, this series, Vol. 42, p. 135. 3 E. Jablonski and M. D e L u c a , Biochemistry 16, 2932 (1977).

84

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[8]

from Boehringer-Mannheim which can be used instead of the oxidoreductases. This enzyme will use either NADH or NADPH as a substrate. We do not have any experience in immobilizing commercially available enzymes; however, there is one report in which the commercial bacterial oxidoreductase and luciferase have been coimmobilized on nylon tubing. 4 Other experiments have indicated that the commercial enzymes can be successfully immobilized onto Sepharose using the same procedures described here. 5 A combined luciferase/oxidoreductase preparation partially purified from Photobacterium fischeri can be purchased from the following companies: Boehringer-Mannheim, Lumac/3M, Calbiochem-Behring/ Hoechst, LKB-Wallac, Analytical Luminescent Laboratory, and Packard Instruments. Sigma Chemical Co. Ltd. sells the enzymes either from P. fischeri or as a preparation from V. harveyi. Firefly luciferase and luciferin are commercially available from the following companies: Boehringer-Mannheim, Calbiochem-Behring/ Hoechst, Sigma Chemical Co. Ltd., LKB-WalIac, Lumac/3M, Turner Design, Packard Instruments, and Analytical Luminescent Laboratory. Instrumentation

The instruments used for measuring light emission can be a very simple design with a photomultiplier tube, power supply, and a light-tight reaction chamber. It is also possible to buy more sophisticated automated luminometers with microprocessor-controlled automatic injection systems. The various commercially available instruments were reviewed by Picciolo et al. in 1978. 6 A more recent discussion of the then currently available instruments can be found in an article by Stanley. 7 A different approach to measuring the light emission from luminescent assays is the use of a high-speed (ASA 20000) instant photographic film, as described by Green et al. ~ Assays

Firefly Luciferase. This enzyme is assayed by injecting 100 ~1 of 0.02 M ATP into 500/.d of 0.025 M glycylglycine buffer, pH 7.8, containing 1 × 10 -4 M D-luciferin, 8 x 10 -3 M MgC12, and an appropriate amount of 4 A Roda, S. Girotti, S. Ghini, B. Gricolo, G. Carrea, and R. Bovara, Clin. Chem. 30, 206 (1984). 5 K. Green, L. J. Kricka, G. H. G. Thorpe, and T. P. Whitehead, Talanta 31, 173 (1984). 6 G. L. Picciolo, J. W. Deming, D. A. Nibley, and E. Chappelle, this series, Vol. 42, p. 550. 7 p. E. Stanley, in "Clinical and Biochemical Luminescence" (L. J. Kricka and T. J. N. Carter, eds.), p. 219. Dekker, New York, 1982.

[8]

BIOLUMINESCENT ASSAYS

85

either soluble or immobilized luciferase. The amount of enzyme to be used is dependent on the sensitivity of the luminometer. The peak light intensity is proportional to the concentration of enzyme present and is expressed in relative light units. If the amount of enzyme is kept constant and limiting amounts of ATP are injected, then light intensity is proportional to the ATP concentration. Bacterial Luciferase. Either the soluble or immobilized enzyme is assayed by the injection of 100 Izl 1.5 x 10 -4 M FMNH2 into 500 tzl of 0.1 M potassium phosphate buffer, pH 7.0, containing 0.0005% (v/v) decanal and luciferase. The FMNH2 is prepared by photoreducing FMN in the presence of 5 × 10-3 M EDTA. 8 The buffer-decanal mixture is made from a 0.05% decanal-water suspension which should be prepared freshly every 4 hr. Oxidoreductases. These enzymes are assayed spectrophotometrically by measuring the initial rate of oxidation of NADH or NADPH as the decrease in absorbance at 340 nm. The reaction is initiated by adding I00 /A of 2 mM N A D H or NADPH to 0.9 ml of potassium phosphate buffer, pH 7.0, containing the enzyme and 0.13 mM FMN. The immobilized oxidoreductases are assayed in the same way except the reaction mixture in the cuvette is stirred continuously to avoid settling of the Sepharose. Diaphorase. The conditions for the determination of the activity of the diaphorase are the same as those used in the oxidoreductase assays. The oxidation of N A D H or NADPH in the absence of added FMN is subtracted from that in the presence of FMN.

Coupled Reaction of Coimmobilized Oxidoreductase and Luciferase. Maximum initial light intensity is measured upon the addition of 10-20/~1 of the Sepharose-bound enzyme suspension into 0.5 ml of 100 mM potassium phosphate buffer, pH 7.0, containing 0.04 mM NADH or NADPH, respectively, 0.003 mM FMN, and 0.0005% decanal. Other Enzymes. All of the other enzymes which are coimmobilized with the oxidoreductase and luciferase are assayed spectrophotometrically according to published procedures. 9 Immobilization Procedure The mixture of enzymes to be immobilized contain 5 mg of bacterial luciferase and 2-3 IU of either the NADH or NADPH : FMN oxidoreduc8 j. Lee, Biochemistry 11, 3350 (1972). 9 H. U. Bergmeyer, "Methods of Enzymatic Analysis." Verlag Chemie Weinheim/Academic Press, New York, 1971.

86

MULTISTEP

c.,

P

r~

ENZYME

c,l

t"q

SYSTEMS

AND

COENZYMES

e,l

[8]

i-,,I

~

t'q

?

?

?

e~ i

×

E

?

X

T T T x

~

b

T

x

? ~-

x

T

x

T

x

~

x

-×××~-×~-×-m M d ~

md

md----

,t"q e,~

0 .<

t',,~ t-.2 t'q

~"-.I ~"~

e'q

[."~ ,r'~ eq

i.

Z

e

Z

e Z 0 ¢,.) >. ,<

0) "~

< ~Z.~

~ _

~Z~

,~ Z.~ ~ Ez~~' ~

~ Z . ~~ .. ~~

~ ~Z

~Zm E

~2

2

z

f~

[8]

BIOLUMINESCENT

O,

~

t~ ~

~

87

ASSAYS

t~ ~

t~

o.

,

~.

.~

-

"~

X

X

~

X

--~'~x-

-

X

~

X

X

z

-~.~ _~.~ "" ~

0

~-

~

~

.~

.~ ~,~.~ =-~ ~

'-

~

~

-~

~

z~

,t~ e~ ~

t~

~n

n~

-~

~.x× ~x××

r....~ X

t~

×

~-~ e,,.h vh

z~z~

X

X

X

~,~ X

x X

X

~

=~

~

.~ ~.~,~'~ . . . ~ o _ ~

~'~'~

~

X

.-

n~

<

X

X

~

88

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[8]

tase or diaphorase. Additional enzymes are added usually in the amount of 2-8 IU. If the final protein concentration of the mixture is less than 8 mg/ml, BSA is added to reach this concentration. The volume of the coupling mixture is 2.5-3.0 ml. The enzymes are dialyzed overnight in the cold against 4 liters of 0.1 M sodium pyrophosphate, pH 8.0. One gram of Sepharose was activated according to the method of March et al.~° and this is added to the enzyme mixture. The suspension is stirred gently at 4 ° for 16 or 2 hr at room temperature. The Sepharose is then washed extensively with 0.1 M potassium phosphate buffer, pH 7.0, and 1 M NaC1 in 0.1 M potassium phosphate, pH 7.0.~1 The Sepharose-enzymes are suspended in 6 ml of 0.1 M potassium phosphate, pH 7.0, containing 0.02% sodium azide and 2 mM dithiothreitol (DTT). Additional DTT, 20/zl of 0.1 M, is added each week. If diaphorase or the NADH : FMN oxidoreductase is used instead of the NADH-dependent oxidoreductases, the storage buffer contains 0.5 mM glutathione rather than DTT. This is essential because DTT is a substrate for these enzymes. The efficiency of the immobilization is determined by measuring the enzyme activities recovered on the Sepharose. The activities recovered on the Sepharose are expressed as a percentage of the initial activities added to the Sepharose. Storage and Stability of Immobilized Enzymes The immobilized enzymes can be stored at 0-4 ° in buffer containing either DTT or glutathione and azide for several months. If more DTT is added weekly, as discussed above, the enzymes remain fully active. For longer storage glycerol is added to the enzyme-Sepharose suspension to give a final concentration of 15%. The immobilized enzymes are then rapidly frozen in liquid nitrogen and kept at - 2 0 ° indefinitely with no loss of activity. Repeated slow freezing and thawing result in a considerable loss of activity. Metabolites Assayed In general, the immobilized enzymes are added to the buffer mixture containing everything except the compound to be measured and the background light is determined. Then the specific compound is added, the contents of the tube are mixed, and either the peak light or the rate of increase of light is measured. When diaphorase is used instead of the ~0 S. C. March, I. Parikh, and P. Cuatrecasas, Anal. Biochem. 60, 149 (1974L tl j. Ford and M. DeLuca, Anal. Biochem. 110, 43 (1981).

[8]

BIOLUMINESCENT ASSAYS

89

TABLE II RECOVERY OF ENZYME ACTIVITIES IMMOBILIZED ON SEPHAROSE Activity recovered on Sepharose Enzyme coupled

(%)

NADH : FMN oxidoreductase Creatine kinase Firefly luciferase Hexokinase Malate dehydrogenase Lactate dehydrogenase 7a-Hydroxysteroid dehydrogenase Bacterial luciferase

90 84 80 60 45 25 20 15

oxidoreductase, the most reproducible results are obtained when the reaction is initiated by the injection of NAD. Table I contains the data 1~-15for the various metabolites which have been measured using coimmobilized enzymes. The final volume for all of the assays listed is 0.5 ml. The buffers and substrate concentrations are different for the various assays. While it is very possible that other assay conditions could be substituted we have found these concentrations to give the best results. In general the various compounds can be assayed in the range of 1010,000 pmol. The lower limit of detection will depend upon the amount of active enzymes that are immobilized as well as on the equilibrium constants for the various enzymes. In addition, the sensitivity of the assay is limited by the amount of background light, which is determined by the purity of the substrates and enzymes used. Many of the commercially available dehydrogenases contain an aldehyde dehydrogenase that gives light in the presence of decanal. These enzymes must be further purified before use in the luminescent assays as described previously. ~1It is necessary to determine the range of linearity of light vs concentration of substrate for each new preparation of immobilized enzymes. Standards should be included every day with unknown samples. Table II shows some representative enzyme recoveries obtained after 12 G. Wienhausen and M. DeLuca, Anal. Biochem. 127, 380 (1982). 13 j. Schoelmerich, G. P. van Berge Henegouwen, A. F. Hofmann, and M. DeLuca, Clin. Chim. Acta 137, 21 (1984). ~4j. Schoelmerich, J. E. Hinkley, I. A. Macdonald, A. F. Hofmann, and M. DeLuca, Anal. Biochem. 133, 244 (1983). l~ A. Roda, L. J. Kricka, M. DeLuca, and A. F. Hofmann, J. Lipid Res. 23, 1354 (1982).

90

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[8]

a

_3,----'-"-'-"

./

03 I-Z I.--r"

t.u

"~

m ee"

/

TIME (min) FIG. 1. Time course o f light output in the assay of D-glucose: (a) background, (b) 2, (c) 5, and (d) 10 pmol of glucose.

immobilization on Sepharose. These recoveries range from almost 100% to a low of 15%. Kinetics The kinetics of light emission are also variable with the different combinations of enzymes. ~2 For example, in the assay for glucose with the four-enzyme system--hexokinase, glucose-6-phosphate dehydrogenase,

bf ,

0'3

r..D

._,.~_ L.I.I

c,,s

0

1

0

t

i

1

2

TIME (min) FIG. 2. Time course of light output in the assay of NADP+: (a) background, (b) 2 pmol of N A D P ÷.

~3.1, ~I31]dLLq/1/N O.l,OHd

t~

=

~.IF

oqoi

°

I I I I

I

L

92

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[8]

100

0 0

10

20

30

40

50 60 70 TIME (rain)

80

90

lOO 110

FIG. 4. Typical analytical traces obtained for repeated analysis of 10 pmol/ml of NADH; 2-min samples, 4-min wash.

N A D H : FMN oxidoreductase, and bacterial luciferase--the slope of increasing light is measured as shown in Fig. 1. With other assays, such as shown for NADP +, the peak light intensity is measured (Fig. 2).12 Each combination of enzymes must be evaluated independently and the best method for measuring the compound of interest is then used. Automation One of the advantages of the coimmobilized enzymes is the possibility of using them in an automated system. The Sepharose-enzymes can be packed into small flow cells and the cells placed in front of the phototube in a standard luminometer. Such a system was developed by Kricka et al. 16The flow cells could be reused in some cases for 500-700 consecutive assays. The reproducibility was good and the lower limit of detection of various metabolites was approximately 1 pmol. Recently the automated system has been modified. ~7A schematic of the modified flow manifold is shown in Fig. 3A, and Fig. 3B shows the design of the immobilized flow 16 L. J. Kricka, G. K. Wienhausen, J. E. Hinkley, and M. DeLuca, Anal. Biochem. 129, 392 (1983). 17 D. Vellom, M. DeLuca, J. Hinkley, A. Loucks, and H. Egghart, in "Analytical Applications of Bioluminescence and Chemiluminescence" (L. J. Kricka, P. E. Stanley, G. H. P. Thorpe, and T. P. Whitehead, eds.). Academic Press, New York, 1984.

[9]

E N Z Y M E C H A N N E L I N G I M M U N O M E T R I C ASSAYS

93

cell and its orientation in the luminometer. By reducing the total volume in the system and positioning the immobilized enzymes very close to the phototube the sensitivity of the assays has been increased significantly. It is possible to measure as little as 6 fmol of NADH in a 60-/zl sample. A typical analytical trace for repeated analysis of NADH is shown in Fig. 4. The reproducibility is good and 10 samples per hour can be measured. We believe that with some further modifications we will be able to increase the sensitivity and the speed of these automated assays.

[9] N o n s e p a r a t i o n E n z y m e C h a n n e l i n g Immunometric Assays

By IAN GIBBONS, RICHARD ARMENTA, ROBERT K. DINELLO, and EDWIN

F.

ULLMAN

Enzymes are often used as labels in immunoassays because of their stability and ease of detection. In many enzyme immunoassays, the enzyme-labeled reagent binds to a surface and the immobilized enzyme is subsequently measured. For example, in the enzyme-linked immunosorbent assay (ELISA), l antigen binds to antibody which has previously been adsorbed to a solid surface. Usually, unbound antigen is removed. Enzyme-labeled antibody is then added and binds to antigen on the solid surface. After carefully washing away unbound reagent, measurement of the bound enzyme gives a direct indication of the quantity of antigen originally present. "Homogeneous" enzyme immunoassays 2 were devised as a way to eliminate the tedious separation steps required by conventional techniques. In homogeneous methods, the binding of antigen to antibody modulates the activity of the bound label. The original homogeneous methods were adapted to measure both drugs 3 and proteins 4 and rely on competition between enzyme-labeled antigen and sample antigen for a limited quantity of antibody. When measuring small amounts of antigen, low concentrations of antibody must be used and the binding reactions J A. H. W. M. Schuurs and B. K. Van Weeman, Clin. Chim. Acta 81, 1 (1977). 2 K. E. Rubenstein, R. S. Schneider, and E. F. Ullman, Biochem. Biophys. Res. Commun. 47, 846 (1972). 3 G. L. Rowley, K. E. Rubenstein, J. Huisjen, and E. F. Ullman, J. Biol. Chem. 250, 3759 (1975). 4 I. Gibbons, C. Skold, G. L. Rowley, and E. F. Ullman, Anal. Biochem. 102, 167 0980).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[9]

E N Z Y M E C H A N N E L I N G I M M U N O M E T R I C ASSAYS

93

cell and its orientation in the luminometer. By reducing the total volume in the system and positioning the immobilized enzymes very close to the phototube the sensitivity of the assays has been increased significantly. It is possible to measure as little as 6 fmol of NADH in a 60-/zl sample. A typical analytical trace for repeated analysis of NADH is shown in Fig. 4. The reproducibility is good and 10 samples per hour can be measured. We believe that with some further modifications we will be able to increase the sensitivity and the speed of these automated assays.

[9] N o n s e p a r a t i o n E n z y m e C h a n n e l i n g Immunometric Assays

By IAN GIBBONS, RICHARD ARMENTA, ROBERT K. DINELLO, and EDWIN

F.

ULLMAN

Enzymes are often used as labels in immunoassays because of their stability and ease of detection. In many enzyme immunoassays, the enzyme-labeled reagent binds to a surface and the immobilized enzyme is subsequently measured. For example, in the enzyme-linked immunosorbent assay (ELISA), l antigen binds to antibody which has previously been adsorbed to a solid surface. Usually, unbound antigen is removed. Enzyme-labeled antibody is then added and binds to antigen on the solid surface. After carefully washing away unbound reagent, measurement of the bound enzyme gives a direct indication of the quantity of antigen originally present. "Homogeneous" enzyme immunoassays 2 were devised as a way to eliminate the tedious separation steps required by conventional techniques. In homogeneous methods, the binding of antigen to antibody modulates the activity of the bound label. The original homogeneous methods were adapted to measure both drugs 3 and proteins 4 and rely on competition between enzyme-labeled antigen and sample antigen for a limited quantity of antibody. When measuring small amounts of antigen, low concentrations of antibody must be used and the binding reactions J A. H. W. M. Schuurs and B. K. Van Weeman, Clin. Chim. Acta 81, 1 (1977). 2 K. E. Rubenstein, R. S. Schneider, and E. F. Ullman, Biochem. Biophys. Res. Commun. 47, 846 (1972). 3 G. L. Rowley, K. E. Rubenstein, J. Huisjen, and E. F. Ullman, J. Biol. Chem. 250, 3759 (1975). 4 I. Gibbons, C. Skold, G. L. Rowley, and E. F. Ullman, Anal. Biochem. 102, 167 0980).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

94

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[9l

become inconveniently slow. This problem can be overcome by adopting a noncompetitive method and driving the binding reactions with excess reagents. However, the use of excess labeled reagents frequently leads to high unmodulatable background signals which limit sensitivity. The present article describes a set of nonseparation immunoassays in which the reagents are enzyme-labeled antibodies. These methods were designed to measure low concentrations of macromolecular antigens with excess reagents while limiting the background due to the unbound reagents. The assays provide sensitivity comparable to that of ELISA while greatly simplifying the assay protocol. The principle on which these assays are based is called enzyme channeling immunoassay. 5 Channeling occurs when two enzymes catalyzing sequential reactions [reaction (1)] are brought together at a surface or other microenvironment where diffusion-controlled exchange with bulk solution is limited. Substrates

enzyme 1 enzyme2 ~ Product 1 ~ Product 2

(1)

Under these conditions, the reduced rate of diffusion of Product 1 away from the microenvironment permits the local concentration of the product to exceed that of the bulk phase. Product 2 is therefore produced more rapidly than when the enzymes are free in solution or when only one enzyme is on the surface. Several immunoassay methods have been demonstrated in which antigen-antibody binding reactions bring together enzyme-labeled reagents at a surface? '6 For the methods reported here, glucose oxidase (GO) was selected as the first enzyme and horseradish peroxidase (HRP) as the second. Together they catalyze the reaction shown in reaction (2). Glucose + O~ -

H202 OO~ + Gluconolactone )

Colored (oxidized) product

HRP , f ~. Chromogen

+

(2)

H 20

These enzymes were chosen for their high turnover numbers and stability. A variety of chromagens are substrates for HRP. 2,2'-Azinodi-(3ethylbenzothiazoline-6-sulfonic acid) (ABTS), which produces an easily 5 D. J. Litman, T. M. Hanlon, and E. F. Ullman, Anal. Biochem. 106, 223 (1980). 6 E. F. Ullman, I. Gibbons, L. Weng, R. DiNello, S. N. Stiso, and D. J. Litman, in "Diagnostic Immunology: Technology Assessment and Quality Assurance" (R. M. Nakamura and J. H. Rippey, eds.), pp. 31-46. College of American Pathologists, Skokie, Illinois, 1983.

[9]

ENZYME CHANNELING IMMUNOMETRIC ASSAYS

95

measured green product, is useful for this purpose. In order to reduce background signal and provide linear enzyme rates, a scavenging agent is employed to destroy hydrogen peroxide in the bulk solution and suppress color due to unbound reagent. Catalase, which converts hydrogen peroxide to water and oxygen, was found to be very effective. The theory of these assays has been described elsewhere 6 and will not be further discussed here. The capsular antigen of Haemophilus influenzae known as polyribose phosphate (PRP) was used in our work as typical of many antigens of clinical interest. PRP is a repeating heterogeneous polymer with many similar antigenic sites. 7 Three enzyme-channeling assays for PRP will be described. The advantages and limitations of each will be discussed. The materials and reagents employed for the assays are largely the same and are specified below. Materials Used and Reagent Preparation

Enzymes. HRP (type VI), glucose oxidase (type V from Aspergillus niger), and catalase (bovine liver) were the products of the Sigma Chemical Co., St. Louis, MO 63128. Antigen. A pure preparation of PRP, the capsular antigen of Haemophilus influenzae, type b, was the kind gift of Dr. Porter Anderson. The antigen is a polyribosylribitol phosphate 7 which had an average molecular weight of 500,000. Antibodies. Rabbit anti-PRP was obtained from the New York State Department of Health. The immunoglobulin fraction was purified from the antiserum by precipitation with ammonium sulfate. The stock solution had a concentration of 80 mg/ml. Anti-GO was raised in sheep by repeated immunization with 100 /zg of the amine derivative of GO (described later) in incomplete Freund's adjuvant. The immunoglobulin fraction of the antiserum was prepared by precipitation with 50% ammonium sulfate. Chemicals. ABTS, Tween 20, ovalbumin, and glucose: Sigma Chemical Co., St. Louis, MO 63128. Sodium metaperiodate: Fisher Chemical Co., Pittsburgh, PA 15219. Sodium borohydride: J. T. Baker Co., Phillipsburg, NJ 08865. 1-Fluoro-2,4-dinitrobenzene: Eastman Kodak, Rochester, NY 14650. Equipment. Microtiter plates and covers: "Serocluster" Costar, Cambridge, MA 02139. Absorbances are measured either with a Gilford Stasar 7 R. M. Crisel, R. S. Baker, and D. E. Dorman, J. Biol. Chem. 250, 4926 (1975).

96

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[9]

III spectrophotometer, Gilford Instrument Labs, Oberlin, OH 44074, or with a Model 210 vertical beam reader, Artek Systems Corp., Farmingdale, NY 11735. Assay incubations are generally done with shaking using a Bellco (Vineland, NJ) miniorbital shaker set at "4."

Measurement of Protein Concentrations Proteins are quantitated spectrophotometrically. HRP is measured at 403 nm assuming a molar extinction of 1.12 x 105 cm -~ M -~ and a molecular weight of 40,000. Native GO is measured at 450 nm assuming a molar extinction of 2.95 x 104 cm -~ M -~ and a molecular weight of 160,000. After derivatization with 1-fluoro-2,4-dinitrobenzene, the extinction of GO at 450 nm increases to 6.76 x 10 4 c m -1 M -1. IgG is assumed to have a molar extinction of 2.4 x 105 cm -l M -t at 280 nm and a molecular weight of 160,000. The composition of conjugates of GO and HRP is deduced from the appropriate spectral ratios correcting the absorbance at 280 nm for the contribution of the enzymes as follows: A4o3/A28o = 2.0 (HRP) and A45o/A28o = 0 . 1 7 (fluorodinitrobenzene-modified GO).

HRP-Labeled Antibody Conjugation of HRP to anti-PRP is according to Wilson and Nakane. s HRP is dissolved in distilled water at 4.3 mg/ml. Four milliliters of freshly prepared sodium periodate (0.1 M) is added to 20 ml of the HRP solution to oxidize the enzyme carbohydrate. After the reaction mixture is stirred for 20 min at room temperature, 2.4 ml of a 1 M glycerol solution is added to terminate the reaction. Stirring is continued for 30 min, followed by overnight dialysis against three changes of 500 ml 2 mM sodium acetate, pH 4.5, to remove the reagents. Anti-PRP diluted to 25 mg/ml in water is dialyzed against 10 mM sodium carbonate, pH 9.6. Coupling to HRP is achieved by mixing 70 mg oxidized HRP (22 ml) with 110 mg antibody (5 ml). The pH of the mixture is adjusted to 9.6 with NaOH (0.1 M) and the reaction allowed to proceed for 2 hr at room temperature. After the mixture is cooled to 0°, the Schiff base complex is reduced by addition of 1.5 ml of a freshly prepared solution of NaBH4 (4 mg/ml of water). This mixture is incubated at 0° for 2 hr, dialyzed overnight against PBS to remove reagents (PBS, 10 mM sodium phosphate, pH 7.2, with 150 mM NaCI), and then concentrated to 5 ml by vacuum ultrafiltration through a collodion bag. The conjugates are s M. B. Wilson and P. Nakane, in "Immunofluorescence and Related Staining Techniques Proc. Int. Conf., 6th" (W. Knapp, K. Holubar, and G. Wick, eds.), p. 215. Elsevier/North Holland, Amsterdam, (1978).

[9]

ENZYME CHANNELING IMMUNOMETRIC ASSAYS

97

separated from unconjugated enzyme and antibody by chromatography on a column (2.5 × 90 cm) of BioGel A-5m eluted with PBS. Fractions of 5 ml are collected. For us, fractions 39 to 57 represented the bulk of the conjugates. These fractions were pooled and concentrated 10-fold by ultrafiltration. A conjugate molar ratio of 1.4 HRP/IgG was determined spectroscopically. The conjugates were heterogeneous in molecular weight ranging from about 1 to 3 × 106 as judged by their elution position. Conjugated enzyme retained 80% of its activity, and complete activity was retained by conjugated antibody, as shown by the ability to compete with native antibody for PRP. Glucose Oxidase-Labeled Antibody

The method of Johnson et al. 9 is used. Glucose oxidase is first allowed to react with 1-fluoro-2,4-dinitrobenzene to block amino groups. GO (200 mg) is dissolved at 5 mg/ml in 0.1 M NaHCO3, pH 8.2. 1-Fluoro-2,4-dinitrobenzene (3.5 ml of a 1%, w/v, ethanol solution) is then added and the reaction allowed to proceed for 1 hr at room temperature. The reagent is removed by dialysis against 0.3 M NaHCO3, and the product concentrated by ultrafiltration to 20 ml. Oxidation of the enzyme carbohydrate is achieved by addition of 2 ml NalO4 (0.4 M). After reaction for 20 min at room temperature, 1 ml glycerol (1 M) is added. One hour later the product is dialyzed against 10 mM sodium carbonate, pH 9.6. A solution of 90 mg of anti-PRP in 3 ml of 10 mM NaHCO3 is mixed with the solution of oxidized GO (200 mg) giving a final volume of 25 ml. After reaction for 14 hr at room temperature, the solution is cooled to 0° and the Schiffbase complexes are reduced by addition of 1.6 ml NaBH4 (4 mg/ml). The mixture is allowed to incubate for 2 hr at 0° before dialysis against PBS. The product is then concentrated to 5 ml and chromatographed as described for the HRP conjugate. In our studies, fractions 5156 were pooled. The conjugate pool had a molar ratio of 0.8 GO per IgG. Conjugates ranged in molecular weight from about 1 to 2 × 106. The GO activity was 30% compared with starting material. Derivatization o f Glucose Oxidase with Amino Groups

A solution of GO (40 mg/ml) in 2 mM sodium acetate (pH 4.5) is combined with an equal volume of freshly prepared NaIO4 (0.2 M in the same solvent). Oxidation is allowed to proceed for 20 rain at room temperature and is followed by purification on a Sephadex G-50 column by elution with 2 mM sodium acetate buffer, pH 4.5. Ethylenediamine (3 M 9 R. B. Johnson, R. M. Libby, and R. M. Nakamura, J. lmmunoassay 1, 27 (1980).

98

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[9]

in 0.5 M Na2CO3, pH 9.5) is added to a final concentration of 0.2 M. After incubation for 1.5 hr at room temperature, the solution is cooled to 0° and the Schiff base complex is reduced by addition of 0.1 volume freshly prepared NaBH4 (4 mg/ml). Following a 2-hr incubation at 0°, aminated GO is recovered by gel filtration over Sephadex G-50.

Microtiter Plates Coated with Glucose Oxidase and Anti-PRP Aliquots (200/A) of a mixture of the amino derivative of GO (40/zg/ml) and anti-PRP (40/zg/ml) dissolved in 60 mM Na2CO3 (pH 9.6) are placed in the wells of microtiter plates. After incubation at 37° for 4 hr unbound material is removed by washing with PBS containing 0.05% Tween 20. Glucose oxidase activity equivalent to about 50 ng of free GO is bound to each well. Coated plates are stored at 4° with PBS-Tween in the wells.

Assay Reagents The HRP and GO conjugates of anti-PRP are stable when stored at 4° in PBS containing 0.01% NaN3 as a preservative. Working dilutions are made up immediately before use. NaN3 is an inhibitor of HRP, but the concentration in the assays is so low as to be negligible. ABTS solutions when freshly made are strongly colored due to traces of oxidized material. Since the color would interfere with measurements of HRP activity, the ABTS solutions are bleached by stirring at room temperature overnight. Stock concentrated solutions of catalase are made by diluting the enzyme suspension to 4 mg/ml in PBS and dialysing against the same solvent to remove the thymol preservative. Concentrations are measured spectrophotometrically assuming a molar extinction coefficient at 450 nm of 548,000 cm -I M -1 (MW = 250,000). The solutions can be stored at 4° and used for up to 30 days.

Enzyme Channeling Immunoassays

Coated-Surface Enzyme Channeling lmmunoassay This assay resembles ELISA except that no separation steps are required. Usually, microtiter plates are used as the solid support. GO is coimmobilized on the solid phase together with antibody. An aminated derivative of GO is used because a much greater activity can be adsorbed to the microtiter plate surface than is possible with native GO. The assay principle is illustrated in reaction (3).

[9]

H20 BULK SOLUTION

99

E N Z Y M E C H A N N E L I N G I M M U N O M E T R I C ASSAYS

G1

I ~TS

~TS

~

02 ~

Cat

ii

/I /

H202

Colored P duct

GIu

Ag

UNSTIRRED LAYER GO

Ab

/11

[Catalase]>[HRP] in Bulk Solution

GO

i

AboAg-(Ab-HRP)

f

T~ I ) / I I

(3)

[ I I

[HRP]>[Catalase] at Surface

Reagents are prepared using PBS containing 0.02% ovalbumin as buffer. Reagent A contains 4.5 ~g HRP conjugated to anti-PRP per milliliter. Reagent B contains 25 mM ABTS, 625 mM glucose, and 500 /zg catalase/ml. Microtiter plates to which GO and anti-PRP have been adsorbed are used as the coated surface. Samples (I00/xl) of PRP dissolved in PBS are added to the wells of the microtiter plate. The plate is covered and incubated for 3 hr at 37°. Then reagent A (50/xl) is added and incubation continued for a further 2 hr at 37°. Finally, reagent B (100/xl) is added to generate color. A405is read using an Artek vertical beam reader after incubation for 30 min at 37 °. The assay response obtained (Fig. 1) resembles that of an ELISA assay using similar reagents. Absorbance increased with added antigen over the range 0.1-3 ng. In the absence of antigen, the enzyme channeling assay gave a higher absorbance and higher background rate than did the corresponding ELISA. This was due to absorbance by catalase and to a background rate of color production not suppressed by catalase. At high antigen concentrations (>3 ng per assay), the response diminished because the reagents were no longer in excess over antigen. This is typical of immunometric assays with no separation. This assay has a more convenient protocol than ELISA in that the need for the usual separation and washing steps has been eliminated. However, like ELISA it is slow because of the time required for macromolecules to bind to the solid surface, and the solid-phase reagent is inconvenient to store and to prepare reproducibly.

Soluble Reagent Enzyme Channeling Immunoassay: Microtiter Plate Protocol To solve some of the problems inherent in the coated-surface assay, an assay was constructed that only required reagents in solution. Since

100

[9]

M U L T I S T E P E N Z Y M E SYSTEMS AND COENZYMES

2.2

1.8

g

,~

1.4

1.o~0.6~

II II

//

t

t

0.01

t

0.1 PRP (ng)

t

1

10

FIG. 1. Coated-surface enzyme channeling immunoassay. Experimental details are given

in the text.

enzyme channeling requires an immobilized phase, an appropriate phase was caused to form during the assay by precipitating one of the reagents [see reaction (4)]. Ab-GO

+

Ab-HRP

I o~-GO

Ag

>

(Ab-GO).Ag. (Ab-HRP)

l

a-GO Ag,(Ab-HRP)

Ab Ib /~b Abt Abl

OyOvyyO

Ab Ab Ib Ab Ab

Precipitate forms

Precipitate forms with channeling

without channeling

(4)

do

YYYY

In this assay, GO-labeled antibody replaces GO and antibody adsorbed to the microtiter plate. Since the antigen is polyepitopic, it can simultaneously bind both GO-labeled antibody and HRP-labeled antibody in solution. The use of labeled antibodies at high concentrations ensures that the reaction is almost instantaneous. Anti-GO is then added to pre-

[9]

ENZYME CHANNELING IMMUNOMETRIC ASSAYS

101

cipitate all the GO-labeled antibody in a process that can be accelerated by shaking. The precipitate forms as small particles. Coprecipitation of the HRP-antibody conjugate with the GO immune complex serves to signal the presence of antigen. Reagents are made up as follows in PBS containing 0.05% Tween 20 and ovalbumin at 1% (reagents A and B) or 0.1% (reagent C). Reagent A contains 120/.Lg HRP conjugated to anti-PRP and 200/.Lg GO conjugated to anti-PRP per milliliter. Reagent B contains anti-GO diluted to 0.44 mg/ ml. Reagent C contains 12 mM ABTS, 284 mM glucose, and 450 /zg catalase/ml. In the assay protocol, 10-~1 samples containing PRP are mixed with 10 /zl of reagent A and 10/xl of reagent B in the wells of a microtiter plate. The plate is covered and gently shaken for 90 min at 37°. Reagent C (220 /zl) is then added and the initial absorbance of the wells at 415 nm is recorded using an Artek vertical beam reader. The increase in absorbance over 30 min at 37° is measured, the plate being shaken between readings. Results obtained with this method are given in Fig. 2. A linear response from 0 to 500 pg per tube (microtiter plate well) was seen. As little as I0 pg of PRP was detected. This represents a considerable improvement over the coated-surface assay. Two factors contribute to the improvement: (1) All the antigen is bound to reagent in the soluble reagent

0.4

.__q 0.3 0

E

t" LF)

"~ o.2

<

0.1

0

I

I

I

I

I

100

200

300

400

500

Polyribose Phosphate (PRP) per Tube (pg) FIG. 2. Soluble reagent enzyme channeling immunoassay: microtiter plate protocol. Experimental details are given in the text. Reproduced by permission from Ullman e t al. 6

102

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[9]

method, whereas only a small fraction of antigen becomes bound to the surface in the coated-surface method. (2) There is a lower background rate in the soluble reagent method. Although much more sensitive, the soluble reagent method is still quite slow because of the time required to form the GO precipitate and the separate incubation needed to generate color.

Soluble Reagent Enzyme Channeling lmmunoassay: Rapid Protocol A much more convenient method resulted from two changes. First, polyethylene glycol is added to speed up particle formation. Second, the assay response is measured over only I rain in a temperature-regulated flow cell. Not only is this protocol faster, but it provides somewhat better sensitivity than the microtiter plate protocol. Reagents are made up as follows in PBS containing 0.1% ovalbumin. Reagent A contains 55 /zg HRP conjugated to anti-PRP and 20 Izg GO 0.08

0,06 0 O3 0 if3 ,T, ,< <1

0.04 .m

(~

E UJ

0.02

o

I 50

I 1oo

1,50

PRP (PO)

FIG. 3. Solublereagent enzyme channeling immunoassay: rapid protocol. Experimental details are given in the text.

[10]

ENZYME COMPLEXES COMPOSED OF A D H AND L D H

103

conjugated to anti-PRP per ml. Reagent B contains anti-GO diluted to 0.44 mg/ml. Reagent C contains 10% polyethylene glycol 6000, plus 0.2% Tween 20. Reagent D contains 10 mM ABTS, 250 mM glucose, and 500 ~g catalase/ml. The PRP samples are dissolved in human serum. In the assay protocol, a 10-/A serum sample is mixed with 10/A of reagent A, 10 ~1 of reagent B, and 10/~1 of reagent C. After incubation at room temperature with orbital shaking for 15 min, 960/zl of reagent D is added. The mixture is subjected to brief vortex mixing prior to aspiration into the flow cell of a Stasar spectrophotometer thermostated to 37°. The change in A415is measured between 10 and 60 sec after the preset temperature is reached. As seen in Fig. 3, the assay has detected as little as 5 pg of PRP. This assay is suitable for automation in a flow analyzer since the time required to read the signal is short (<1 min). The overall time of the assay (16 min) is much less than a typical ELISA and comparable sensitivity is achieved. The protocol can be further simplified by combining reagents A and C. Conclusions The homogeneous methods discussed here demonstrate the potential of nonseparation techniques to match ELISA in performance. The assay protocols represent a significant advance over ELISA in convenience and flexibility.

[10] I m m o b i l i z e d a n d S o l u b l e S i t e - t o - S i t e D i r e c t e d E n z y m e Complexes Composed of Alcohol Dehydrogenase and Lactate Dehydrogenase

By NILS SIEGBAHN, MATS-OLLE M,~NSSON, and KLAUS MOSBACH Coimmobilization of enzymes on solid supports leads to heterogeneous preparations in which the active sites of the enzymes are found "at random" in relation to one another. A bifunctional NAD analog, bisNAD, has been reported j as a useful reagent for affecting affinity precipitation of enzymes. The bis-NAD can also be used to obtain both an immobilized and a soluble two-enzyme system in which the two different P. O. Larsson and K. Mosbach, FEBS Left. 98, 333 (1979).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[10]

ENZYME COMPLEXES COMPOSED OF A D H AND L D H

103

conjugated to anti-PRP per ml. Reagent B contains anti-GO diluted to 0.44 mg/ml. Reagent C contains 10% polyethylene glycol 6000, plus 0.2% Tween 20. Reagent D contains 10 mM ABTS, 250 mM glucose, and 500 ~g catalase/ml. The PRP samples are dissolved in human serum. In the assay protocol, a 10-/A serum sample is mixed with 10/A of reagent A, 10 ~1 of reagent B, and 10/~1 of reagent C. After incubation at room temperature with orbital shaking for 15 min, 960/zl of reagent D is added. The mixture is subjected to brief vortex mixing prior to aspiration into the flow cell of a Stasar spectrophotometer thermostated to 37°. The change in A415is measured between 10 and 60 sec after the preset temperature is reached. As seen in Fig. 3, the assay has detected as little as 5 pg of PRP. This assay is suitable for automation in a flow analyzer since the time required to read the signal is short (<1 min). The overall time of the assay (16 min) is much less than a typical ELISA and comparable sensitivity is achieved. The protocol can be further simplified by combining reagents A and C. Conclusions The homogeneous methods discussed here demonstrate the potential of nonseparation techniques to match ELISA in performance. The assay protocols represent a significant advance over ELISA in convenience and flexibility.

[10] I m m o b i l i z e d a n d S o l u b l e S i t e - t o - S i t e D i r e c t e d E n z y m e Complexes Composed of Alcohol Dehydrogenase and Lactate Dehydrogenase

By NILS SIEGBAHN, MATS-OLLE M,~NSSON, and KLAUS MOSBACH Coimmobilization of enzymes on solid supports leads to heterogeneous preparations in which the active sites of the enzymes are found "at random" in relation to one another. A bifunctional NAD analog, bisNAD, has been reported j as a useful reagent for affecting affinity precipitation of enzymes. The bis-NAD can also be used to obtain both an immobilized and a soluble two-enzyme system in which the two different P. O. Larsson and K. Mosbach, FEBS Left. 98, 333 (1979).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

104

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[10l

active sites of two different enzymes are facing one another. 2,3 The bisNAD is used as a directing aid for the orientation of the active sites. After cross-linking of the two enzymes with glutaraldehyde, the bis-NAD template is removed, leaving the active sites still positioned against one another. By such an active site arrangement of the two enzymes it can be expected that the diffusion of the product of the first enzyme to the juxtaposed active site of the second enzyme would be facilitated when compared with "at random" immobilized enzyme systems. Such site-tosite enzyme systems might also serve as models for enzyme complexes of consecutively operating enzymes, which are believed to be of importance in the channeling of labile intermediates and in the regulation of metabolism. 4 Site-to-Site Orientation

General Procedure Buffer A: 0.1 M sodium phosphate, pH 7.5, 0.1 M NaC1. Buffer B: 0.1 M sodium phosphate, pH 7.5, 0.5 M NaC1. Buffer C: 0.1 M sodium phosphate, 0.1 M NaCI, 0.05 M oxalate, and 10 mM pyrazole. All washing steps of the gels are performed on a glass filter funnel. Incubation of the gels is normally carried out in 25 ml Erlenmeyer flasks under gentle stirring.

Immobilized System Moist Sepharose 4B, 2 ml, is activated with 25 /zl tresyl chloride (Fluka Buchs, Switzerland) as described elsewhere in this series) The activated gel is subsequently washed with several volumes of buffer A. Horse liver alcohol dehydrogenase (ADH, 2.7 U/mg) (Boehringer-Mannheim, FRG), 12 mg, is dissolved in about 2 ml buffer A and added to the activated gel. The coupling is allowed to proceed for 2 hr at room temperature and the immobilized enzyme preparation is then washed with several volumes of buffer A. The remaining active tresyl groups on the 2 M.-O. M~nsson, N. Siegbahn, and K. Mosbach, Proc. Natl. Acad. Sci. U.S.A. 80, 1487 (1983). 3 N. Siegbahn, M.-O. MAnsson, and K. Mosbach, Appl. Biochem. Biotechnol. 12, 91 (1986). 4 G. R. Welch (ed.), "Organized Multienzyme Systems." Academic Press, New York, 1985. 5 K. Nilsson and K. Mosbach, this series, Vol. 104, p. 56, and Vol. 135 [3].

[10]

ENZYME COMPLEXES COMPOSED OF A D H AND L D H

105

Sepharose are quenched with 0.25 M Tris-HCl, pH 8.0, for 2 hr at room temperature. The gel is then washed with several volumes of buffer B and finally with buffer C. About 200 nmol bis-NAD (Sigma Chem. Co., St. Louis, MO) in 2 ml buffer C is added and during 10 min at 0° allowed to form a ternary complex with the active site of ADH in the presence of pyrazole (Fig. 1, step 1). In order to remove excess of bis-NAD and to maintain the formed ternary complex the gel is washed 5 times with 5 volumes of buffer C. The amount of bis-NAD that remains affinity bound is calculated by subtracting the bis-NAD removed during washing from the amount initially added. Lactate dehydrogenase (LDH) (from beef heart, 490 U/mg, Sigma Chem. Co., St. Louis, MO), two times the molar amount of affinity-bound bis-NAD, dissolved in a minimal volume of buffer C, is added to the gel, for affinity coupling of LDH to the NAD entity of the affinity-bound bisNAD pointing out from the immobilized ADH (Fig. 1, step 2). After 5 min of incubation the gel is washed 5 times with 2 ml of cold buffer C. The amount of LDH that remains affinity bound to the gel is determined in the same way as for the determination of remaining bis-NAD described above. The gel is subsequently suspended in 2 ml cold buffer C and glutaraldehyde is added to a final concentration of 0.06% (Fig. 1, step 3). The cross-linking is allowed to proceed for 2.5 hr at room temperature. After the cross-linking step the gel is first washed and then suspended in 0.2 M Tris-HCl, pH 8.0, in order to quench the remaining unreacted aldehyde groups and kept at 4° overnight. After quenching, the gel is washed with several volumes each of buffer A, then buffer B, and finally buffer A containing 0.1 M isobutyramide, 0.05 M oxamate, and 1 mM

ADH

LDH

FIG. 1. Schematic presentation of preparation of immobilized site-to-site directed twoenzyme complex. See text for further explanation. The oligomeric nature of the enzymes has not been taken into account. ('~) Pyrazole; (O) Oxalate; ([] []) bis-NAD.

106

MULTISTEP ENZYME SYSTEMS AND COENZYMES

ADH

([]

~

[10]

ADH

FIG. 2. Preparation of reversibly immobilized alcohol dehydrogenase. ( 4 ) Pyrazole; []) bis-NAD.

NADH. This final treatment is carried out because it is expected that NADH, together with isobutyramide and oxamate, will form new ternary complexes with ADH and LDH, respectively. The NADH will then compete with remaining enzyme-bound bis-NAD and thereby displacing it from the active sites of the two enzymes (Fig. 1, step 4). The gel is finally washed with several volumes of buffer B and buffer A in order to remove the NADH.

Soluble System A soluble site-to-site oriented enzyme complex is obtained by immobilizing ADH on solid phase with a cleavable disulfide bond (Fig. 2, steps 14) prior to the active site orientation of LDH. After the cross-linking of the two enzymes the enzyme complex is split off from the solid phase by cleaving the disulfide bond between the ADH part of the complex and the matrix. An amino-group-containing gel is obtained by coupling diaminoethane to cyanogen bromide-activated Sepharose 4B.6 In order to introduce thiol groups onto the gel, 2 ml 20 mM N-succinimidyl-3-(2-pyridyldithio)propionate (SPDP) (Pharmacia Fine Chem. Co., Uppsala, Sweden) in ethanol is added to 5 ml amino-gel suspended in 5 ml buffer A. After coupling for 2 hr at 25° the SPDP-activated gel is washed with several volumes of buffer A. The 2-pyridyl disulfide-Sepharose obtained is then easily converted into the free thiol form by treating 5 ml of moist gel with 5 ml 100 mM 1,4dithioerythritol (DTE) (Boehringer-Mannheim, FRG), dissolved in buffer A, for 30 rain at room temperature (Fig. 2, step 2). The gel is subsequently washed with several volumes of buffers B, A, and C. The amount of thiol groups on the gel is determined by cleavage of the disulfide bond and then spectrophotometric measurement of the amount 6 S. C. March, I. Parikh, and P. Cuatrecasas, Anal. Biochern. 60, 149 (1974).

[10]

ENZYME COMPLEXES COMPOSED OF A D H AND L D H

107

of released pyridine-2-thione (extinction coefficient at 343 nm of 8080 cm-~ M-~ 7). Since the free thiol groups can take part in unwanted side reactions, the thiol-Sepharose is used immediately after the cleavage. SPDP-activated ADH is prepared as follows 7 (Fig. 2, step 3). ADH (312 nmol), 25 mg, is dissolved in 6 ml of buffer C and in order to protect the active sites of ADH during the SDDP activation step, about 1/zmol of bis-NAD, dissolved in 100/xl distilled water, is added. A ternary complex between the active sites of ADH, pyrazole, and bis-NAD is thereby formed. A five molar excess of SPDP (20 mM in ethanol) is dropwise added to the solution of protected ADH and the reaction is allowed to proceed 30 min at 25°. The unreacted SPDP is removed by dialysis of the ADH solution against three changes of 1 liter of buffer C. The amount of SPDP coupled to ADH can then be determined as described above for the solid-phase approach. The SPDP-activated and dialyzed ADH is mixed with the thiolSepharose immediately after the above-mentioned reduction of the 2pyridyl disulfide-Sepharose, and the coupling is allowed to proceed 16 hr at 4°. The immobilized ADH is then carefully washed with cold buffers A, B, and C. The affinity immobilization of LDH and the subsequent cross-linking with glutaraldehyde are performed as described above (see Fig. 1). Quenching of unreacted aldehyde groups is accomplished by incubation for 24 hr at 4 ° with 20 ml 0.2 M Tris-HC1, pH 8.0, containing 33 mM NaCNBH3 and 10 mM AMP. This gentle treatment with NaCNBH3 reduces the Schiff bases that are formed by the reaction of aldehyde and Tris. 8 AMP is added to the quenching solution to remove possible remaining affinity-bound bis-NAD by competition. AMP may also protect the active sites of ADH and LDH during the quenching procedure. After quenching, the gel is washed with buffers B and A to remove all AMP. Finally 5 ml of buffer A, containing 100 mM DTE, is added to the gel in order to cleave the disulfide bonds between ADH and the matrix. After 25 min of incubation at 25° the gel is washed 4 times with 5 ml of 50 mM DTE in buffer A. After this treatment about 40% of immobilized ADH activity is recovered in the effluent. The effluent is made 25 mM with respect to oxalate in order to carry out affinity purification of the A D H LDH complex. The oxalate-containing solution (50 ml) is subsequently applied to a 10-ml column of weakly substituted NAD-Sepharose. The NAD-Sepharose is obtained by coupling 0.55 /.tmol N6-[N-(6-amino 7 T. Stuchburg, M. Shipton, R. Norris, J. P. G. Malthouse, K. Brocklehurst, J. A. L. Herbert, and H. Susckitzky, Biochem. J. 151, 417 (1975). 8 R. Borch, M. Bernstein, and D. Durst, J. Am. Chem. Soc. 93, 2897 (1971).

108

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[10]

hexyl)carbamoylmethyl]-NAD, 9 available from Sigma Chem. Co., to 10 ml Sepharose 4B activated with CNBr (50 mg CNBr/ml gel). 6 The NAD-Sepharose is packed in a column and equilibrated with 25 mM oxalate in 50 mM sodium phosphate, pH 7.5, prior to the affinity purification of the A D H - L D H complex. The complex is, in the presence of oxalate, affinity bound to the NAD column since a ternary complex is formed with NAD, oxalate, and the active sites of LDH of the complex. After application of the sample, the NAD-gel is washed with 25 mM oxalate in 50 mM sodium phosphate, pH 7.5, until all non-affinity-bound enzymes have been removed. The L D H - A D H complex is then eluted from the affinity column by changing the elution buffer to buffer A, thereby excluding the oxalate in the elution buffer. The protein fractions that contain the A D H - L D H complex are collected and used for further experiments. Enzyme Assays All assays with immobilized enzymes (5-100 mg of moist gel) are carried out in a stirred 25-ml vessel, containing the immobilized enzyme suspended in 15 ml of assay solution which is continuously pumped with a peristaltic pump through a filter, allowed to pass through a flow cuvette in a spectrophotometer (Hitachi 181), and recycled back to the reaction vessel. All assays with soluble enzymes are carried out in I- or 3-ml cuvettes. Lactate Dehydrogenase (LDH) Activity. LDH activity is determined (in buffer A) by measuring the decrease in absorbance at 340 nm with 0.3 mM and 5 mM pyruvate as substrates. Lipoamide Dehydrogenase (LiDH) Activity. LiDH activity is determined in buffer A by the method described by Massey. 10The reduction of 0.7 mM ferricyanide (e420 = 1040 cm -1 M -1) is followed with 0.2 mM N A D H as second substrate. Alcohol Dehydrogenase (ADH) Activity. ADH activity is measured in buffer A as the increase in absorbance at 340 nm with 1 mM NAD and 50 mM ethanol as substrates, unless a coupled substrate assay is used. In the coupled substrate assay the ADH activity is measured in buffer A containing 50 mM oxalate using 5 mM benzyl alcohol and 5 mM acetaldehyde as substrates (Fig. 3). The coenzyme NAD is regenerated in the active site of ADH after addition of the substrates. The required coenzyme is added to the immobilized enzyme system in the following order. First the immo9 M. Lindberg, P. O. Larsson, and K. Mosbach, Eur. J. Biochem. 40, 187 (1973). 10 V. Massey, this series, Vol. 9, p. 272.

[10]

ENZYME COMPLEXESCOMPOSEDOF ADH AND LDH

ADH ~

109

LDH

ethanol benzylaleohol FIG. 3. Site-to-site directed complex with readded Dis-NADas coenzyme in a coupled substrate assay. The oligomeric nature of the enzymes has not been taken into account. (Q) Oxalate; ( t ~ [ ] ) bis-NAD.

bilized A D H - L D H complex is washed with several volumes of buffer A containing 50 mM oxalate. Bis-NAD, 50 nmol, is dissolved in 1 ml of the same buffer and added to the gel, allowing its affinity binding to the active sites of LDH. After 30 min at room temperature excess of bis-NAD is removed by washing the gel with a few volumes (2-3) of buffer A containing 50 mM oxalate. One of the two NAD entities of the affinity-bound bisNAD is accordingly pointing outward from the active site of L D H (Fig. 3). This NAD entity is then able to reach the active site of ADH and can there act as a coenzyme for the coupled reaction. A molar extinction value at 279 nm for the product, benzaldehyde, of 1400 cm -1 M -I ~1 is used. In order to determine the maximum activity the solution is subsequently made 50/xM with respect to NAD. In the corresponding coupled substrate assay for the soluble system, bis-NAD is added to the enzyme complex to a final concentration of 50 nM in the cuvette and incubated for 10 min at room temperature in order to saturate all active sites of L D H with bis-NAD. Benzaldehyde formation is measured as before. In order to obtain maximum coupled substrate activity, 5/xl 10 mM NAD is added to the assay solution.

Scavenger Enzyme Assay In order to investigate the effects of the orientation and general proximity of the active sites of L D H and ADH, a third enzyme, lipomide dehydrogenase (LiDH), was introduced as a scavenger enzyme (Fig. 4). For the immobilized enzyme system, ADH (2 mg) and LiDH (3 mg, 136 U/mg) are coimmobilized to tresyl chloride-activated Sepharose 4B (2.5 g of moist gel). L D H is subsequently site-to-site immobilized to the immobilized ADH as described above. In the control experiment all three en11C. W. Fuller, J. R. Rubin, and H. J. Bright, Eur. J. Biochern. 103, 421 (1980).

I I0

[10]

MULTISTEP ENZYME SYSTEMS AND COENZYMES

A

B

/

NA

\

\

NAD+ ~ /

NAD+ NADH,/

NAD+/

NAD+

FIG. 4. Scavenger assay with lipoamide dehydrogenase (LiDH) as an enzyme competing with LDH for NADH formed by ADH.

zymes (alcohol, lactate, and lipoamide dehydrogenases) are immobilized randomly : 1.2, 0.6, and 2.0 mg, respectively, are used per 2.5 g of tresyl chloride-activated moist gel. The scavenger enzyme assay is performed as follows, The gel is suspended in 13.5 ml buffer A containing 1 mM NAD and 0.75 mM ferricyanide. Ethanol (final concentration 50 mM) is added and the reduction of ferricyanide is measured as the decrease in absorbance at 420 nm. Pyruvate is subsequently added to a final concentration of 5 mM in order to measure the competition between LiDH and LDH for the NADH formed by ADH. The amount of NADH oxidized by LDH is thus measured indirectly as the decrease in LiDH activity. The soluble site-to-site enzyme complex was also assayed with the scavenger enzyme assay using the same system as for the immobilized enzymes. Characterization of Site-to-Site Enzyme Complexes

ADH Activity In order to characterize both the immobilized and soluble enzyme complexes the ADH activity was determined by a coupled substrate assay with bis-NAD affinity bound to LDH acting as coenzyme. As is shown in Table I, as much as 33% o f the ADH activity obtained in the presence of excess of NAD (50/~M) was obtained upon readdition of 50 nM bis-NAD to the immobilized A D H - L D H complex. In a reference experiment, ADH and LDH were randomly coimmobilized and for this preparation much less activity (4%) was observed upon readdition of bis-NAD. The soluble enzyme complex was characterized in a similar way. For the first measurement no coenzyme was added to the assay solution, However, some ADH activity was observed. The same effect was also

[10]

ENZYME COMPLEXES COMPOSED OF A D H AND L D H

111

TABLE I RELATIVE A D H ACTIVITY OF SOLUBLE AND IMMOBILIZED COMPLEXES OF A D H AND L D H AT DIFFERENT N A D CONCENTRATIONS

Relative ADH activitya

System Immobilized Site-to-site oriented enzymes Randomly coupled Soluble Site-to-site oriented enzymes Separate enzymes

0

50 nM bis-NADb

50 ~tM NAD

11 0

33 4

100 100

20 0

55 <1

100 100

a ADH activity was determined with a coupled substrate assay using three different coenzyme concentrations. Relative ADH activity = (ADH activity at different coenzyme concentration/activity with 50/zM NAD) × 100. b The bis-NAD affinity bound to LDH in the immobilized system corresponds to a nominal concentration of approximately 50 nM.

noted for the immobilized site-to-site directed enzyme complex as shown in Table I. This was probably due to some remaining bis-NAD that had not been removed by washing. As the washing procedure and the purification of both the immobilized and the soluble site-to-site directed enzyme complexes were rather extensive, it indicates that the bis-NAD template was physically entrapped between the active sites of the two enzymes. This was probably due to the cross-linking treatment since adding more glutaraldehyde during the cross-linking procedure caused an even higher coupled substrate activity prior to adding any coenzyme to the assay solution. The coupled substrate activity for the soluble enzyme complex increased when the bis-NAD concentration increased to 50 nM in the presence of oxalate. A 50 nM concentration of bis-NAD corresponds to the nominal concentration of bis-NAD in the corresponding immobilized system. The soluble enzyme complex reached, at 50 nM bis-NAD, about 55% of the ADH activity obtained with an excess of NAD (50/zM) NAD. The corresponding system of separate soluble enzymes showed, in the presence of oxalate, no measurable activity (< 1%) at 50 nM bis-NAD compared to the maximum activity obtained with 50/zM NAD. From these experiments it becomes apparent that one moiety of the bis-NAD can interact with ADH while the other half of the molecule is still affinity bound to LDH. This indicates that the overall geometry of the two enzymes relative to one another, obtained upon cross-linking, is re-

112

MULTISTEP ENZYME SYSTEMS AND COENZYMES

[10]

tained also after the majority of the bis-NAD templates had been washed away. The results from the two reference experiments, i.e., randomly immobilized and soluble separate enzymes (Table I), show that the bisNAD analog added does not interact to any large extent with ADH under the conditions used. This is most likely due to the fact that the two enzymes are too far apart or are not properly oriented toward one another.

Scavenger Enzyme Assay To further investigate the effects of the orientation of the active sites of ADH and LDH, a third enzyme, lipoamide dehydrogenase (LiDH), was introduced in an assay as a scavenger enzyme (Fig. 4). When NAD and ethanol were added to such an enzyme system, NADH and acetaldehyde were formed. The N A D H formed was subsequently oxidized by LiDH when ferricyanide was added. When pyruvate then was added, LDH competed with LiDH for the NADH formed by ADH. This competing reaction was measured as the decrease in LiDH activity. Only ADH operated under substrate saturation conditions and, since LDH and LiDH were in excess, it was ensured that all NADH molecules produced by ADH were rapidly oxidized. The amount of N A D H oxidized by LDH and LiDH, respectively, would be expected to be determined by the relative total numbers of TABLE II RELATIVE NADH-OxIDIZING ACTIVITIES OF THE THREE-ENZYME SYSTEM COMPRISING ADH, LDH, Ar~r~LiDH NADH oxidized in "scavenger" assay (%p By LiDH System Immobilized Site-to-site oriented enzymes Random coupling Soluble Site-to-site oriented enzymes Separate enzymes

Theoretical

By LDH

Found

Theoretical

19

Found

81

50

95.5

99.5

37

17

63

83

37

39

63

61

4.5

50 0.5

The theoretical percentage of NADH oxidized by either LDH or LiDH was calculated from the obtained separate enzyme activities as the ratio of found Vma~activities of LDH and LiDH, LDH/(LHD + LiDH), and LiDH/(LDH + LiDH), respectively.

[10]

ENZYME COMPLEXES COMPOSED OF A D H AND L D H

113

enzyme units of the two enzymes since the Kn~values for NADH of LiDH and LDH are approximately the same under the conditions used. In a randomly coupled three-enzyme system the amount of NADH oxidized by LiDH was that to be expected on these grounds. As is shown in Table II the values were 99.5% found and 95.5% expected. For the system with LDH juxtaposed to ADH, it was predicted that 81% of formed NADH would be oxidized by LiDH. However, it was found that only 50% of the formed NADH was oxidized by LiDH, showing the channeling effect. The soluble site-to-site enzyme complex was also studied with the scavenger enzyme assay and in principle the same kind of results were obtained as for the immobilized system (Table II). The results obtained for both the immobilized system and the soluble system indicate that, given two alternative routes, the one to LDH is preferred over the one to LiDH when LDH and ADH are site-to-site oriented.

[12]

PRODUCTION

OF OPTICALLY

ACTIVE COMPOUNDS

117

[11] O v e r v i e w

By L. ANDERSSON,M.-O. M/~NSSON,and K. MOSBACH In this section we have collected a relatively large number of contributions reflecting the growing interest in the use of immobilized enzymes/ cells for organic syntheses. Striking is the fact that many reactions are carried out in organic solvents. We would like to draw the reader's attention to a recent publication and references cited therein related to this topic describing enzyme modification with polyethylene glycol resulting in enzymes formidably stable in organic solvents [K. Takahashi, Y. Kodera, T. Yoshimoto, A. Ajima, A. Matsushima, and Y. Inada Biochem. Biophys. Res. Commun. 131, 532 (1985)]. Also in an article by Nilsson and Mosbach in Vol. 135 [3] of this series an example of enzyme application in organic solvents, more specifically on peptide synthesis, is given. Other aspects not treated here but of potential importance include the increased use of thermostable enzymes in connection with coenzyme recycling [see, for instance, Enzyme Eng. 8, in press (1987)]. Surely it can be anticipated that in the future we will see increasing use of immobilized systems containing genetically engineered enzymes, such as those obtained by site-directed mutagenesis.

[12] E n z y m a t i c P r o d u c t i o n o f O p t i c a l l y A c t i v e C o m p o u n d s in B i p h a s i c A q u e o u s - O r g a n i c S y s t e m s

By

ALEXANDER M. KLIBANOV a n d BERNARD CAMBOU

Production of optically active compounds is one of the most challenging problems in modern organic chemistry. 1.2 In addition to its fundamental interest, this area is also of great practical significance. As new generations of specialty chemicals (pharmaceuticals, agricultural chemicals, flavors, food additives, etc.) become increasingly complex, a growing number of them contain chiral carbon atoms. Since the ultimate function of most of the aforementioned substances is to interact with biological i E. L. Eliel and S. Otsuka (eds.), "Asymmetric Reactions and Processes in Chemistry." American Chemical Society, Washington, D.C., 1982. 2 j. D. Morrison (ed.), "Asymmetric Synthesis." Academic Press, New York, 1983.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc, All rights of reproduction in any form reserved,

[12]

PRODUCTION

OF OPTICALLY

ACTIVE COMPOUNDS

117

[11] O v e r v i e w

By L. ANDERSSON,M.-O. M/~NSSON,and K. MOSBACH In this section we have collected a relatively large number of contributions reflecting the growing interest in the use of immobilized enzymes/ cells for organic syntheses. Striking is the fact that many reactions are carried out in organic solvents. We would like to draw the reader's attention to a recent publication and references cited therein related to this topic describing enzyme modification with polyethylene glycol resulting in enzymes formidably stable in organic solvents [K. Takahashi, Y. Kodera, T. Yoshimoto, A. Ajima, A. Matsushima, and Y. Inada Biochem. Biophys. Res. Commun. 131, 532 (1985)]. Also in an article by Nilsson and Mosbach in Vol. 135 [3] of this series an example of enzyme application in organic solvents, more specifically on peptide synthesis, is given. Other aspects not treated here but of potential importance include the increased use of thermostable enzymes in connection with coenzyme recycling [see, for instance, Enzyme Eng. 8, in press (1987)]. Surely it can be anticipated that in the future we will see increasing use of immobilized systems containing genetically engineered enzymes, such as those obtained by site-directed mutagenesis.

[12] E n z y m a t i c P r o d u c t i o n o f O p t i c a l l y A c t i v e C o m p o u n d s in B i p h a s i c A q u e o u s - O r g a n i c S y s t e m s

By

ALEXANDER M. KLIBANOV a n d BERNARD CAMBOU

Production of optically active compounds is one of the most challenging problems in modern organic chemistry. 1.2 In addition to its fundamental interest, this area is also of great practical significance. As new generations of specialty chemicals (pharmaceuticals, agricultural chemicals, flavors, food additives, etc.) become increasingly complex, a growing number of them contain chiral carbon atoms. Since the ultimate function of most of the aforementioned substances is to interact with biological i E. L. Eliel and S. Otsuka (eds.), "Asymmetric Reactions and Processes in Chemistry." American Chemical Society, Washington, D.C., 1982. 2 j. D. Morrison (ed.), "Asymmetric Synthesis." Academic Press, New York, 1983.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc, All rights of reproduction in any form reserved,

118

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

receptors (consisting of asymmetric building blocks), it turns out that very frequently only one out of two or more isomers of chiral molecules is biologically active. This means that one has to be able either to readily resolve racemic mixtures of enantiomers or to asymmetrically synthesize optical isomers. Conventional chemical methods for the former, although often used, are laborious, time-consuming, and inefficient.3 Chiral chemical syntheses, on the other hand, despite some spectacular recent developments, 4,5 are still not generally applicable to large-scale preparations. Enzymes are uniquely suitable for asymmetric transformations. ~ This is because they are built of only L-amino acids and hence their active centers constitute asymmetric environments which are likely to react differently with different enantiomers. Despite this and other virtues of enzymes as catalysts in organic syntheses, 6-9 there are also some major obstacles. 9 One of the most serious of them stems from the fact that enzymes are supposed to function in water and therefore have traditionally been used in aqueous solutions. That represents a severe limitation with respect to organic syntheses, because (1) a majority of compounds on interest to organic chemists are poorly soluble in water, and (2) in a number of instances the position of equilibrium for a desired reaction is unfavorable in aqueous solutions or water participates in a side reaction which competes with the desired one. In 1977, Klibanov e t al. 1° proposed an approach to solve both of the above problems. The gist of the method was to employ biphasic aqueousorganic mixtures. The enzyme is located in the aqueous phase, while the substrates are dissolved in a water-immiscible organic solvent constituting the organic phase. Substrates and products can freely diffuse between the two phases and yet the solubility problem does not come up. This approach was illustrated using chymotrypsin-catalyzed esterification reaction between N-acetyl-L-tryptophan and ethanol: whereas in water the yield of N-acetyl-L-tryptophan ethyl ester was only about 0.01%, in a biphasic system consisting of approximately 99% chloroform and 1% wa-

3 j. Jacques, A. Collet, and S. H. Wilen, "Enantiomers, Racemates and Resolutions." Wiley, New York, 1981. 4 C. H. Behrens and K. B. Sharpless, Aldrichimica Acta 16, 67 (1983). H. S. Mosher and J. D. Morrison, Science 221, 1013 (1983). 6 j. B. Jones and J. F. Beck, in "Applications of Bochemical Systems in Organic Chemistry" (J. B. Jones, C. J. Sih, and D. Perlman, eds.), Part I, p. 107. Wiley, New York, 1976. 7 C.-H. Wong and G. M. Whitesides, Aldrichimica Acta 16, 27 (1983). 8 M. A. Findeis and G. M. Whitesides, Annu. Rep. Med. Chem. 19, 263 (1984). 9 A. M. Klibanov, Science 219, 722 (1983). 10 A. M. Klibanov, G. P. Samokhin, K. Martinek, and I. V. Berezin, Biotechnol. Bioeng. 19, 1351 (1977).

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

119

ter it was nearly 100%. 10 Since the first paper, ~° the theory of the enzymatic esterifications in biphasic aqueous-organic mixtures has been refined and applied to the production of other esters. 11-~s Although the approach described above has not been used for asymmetric transformations (either already optically pure N-acetyl-L-amino acids or nonchiral acids served as substrates), it is clear that the methodology, in principle, may have a potential in that regard. However, detailed analysis TMof enzymatic esterifications in biphasic systems shows that this approach is not generally applicable for preparative conversions. The latter require high (preferably molar/liter) concentrations of substrates to boost productivity. Most acids (one of the substrates) will partition from organic to the aqueous phase where they will dissociate. Hence a high concentration of the acid in the organic phase will result in its high concentration in the aqueous phase. This, in turn, will drastically reduce the pH in the aqueous phase (a simple calculation shows that there is no way even a concentrated buffer could prevent that) and, consequently, cessation of the enzymatic reaction (unless the enzyme is an acid hydrolase, and those are rare). It occurred to us ~5that in order to avoid the above complications, one could use enzymatic transesterifications [Eq. (1)] instead of esterifications. Since all reactants are esters and alcohols, there is no pH change during the reaction. In the present paper, we describe the use of this technology for the preparative production of a number of optically active compounds. Rationale Most hydro-lyases, 16 in particular carboxylesterases, 17 can catalyze not only their "natural" reaction of hydrolysis, but also that of transesterification: RICOOR2 + RsOH --~ RtCOOR3 + R2OH

(1)

In the absence of water or in a biphasic system with a very low overall H K. Martinek, A. N. Semenov, and I. V. Berezin, Biochim. Biophys. Acta 658, 76 (1981). 12 K. Martinek and A. N. Semenov, Biochim. Biophys. Acta 658, 90 (1981). ~3j. L. Vidaluc, M. Baboulene, V. Speziale, A. Lattes, and P. Monsan, Tetrahedron 39, 269 (1983). ~4 B. Cambou and A. M. Klibanov, Biotechnol. Bioeng. 27, 1449 (1985). 15 B. Cambou and A. M. Klibanov, J. Am. Chem. Soc. 106, 2687 (1984). ~6 L. B. Spector, "Covalent Catalysis by Enzymes," Chap. 4. Springer-Verlag, New York, 1982. 17 p. Greenzaid and W. P. Jencks, Biochemistry 10, 1210 (1971).

120

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

fraction of water, the reaction of hydrolysis may be almost completely suppressed. Suppose that either Rl or R2 or R3 contains a chiral carbon. Furthermore, let us assume that the carboxylesterase catalyzing reaction (1) is stereospecific, i.e., it accepts only one enantiomer (of either the ester or the alcohol) as a substrate. Then enzymatic transesterification shown in Eq. (1) can be used for preparative resolution of racemic ester R1COOR2 or racemic alcohol R3OH provided that the second, nonchiral substrate is present in a molar excess. The reactive enantiomer will be completely enzymatically converted to a new chemical entity, thereby making it easy to separate it from the unreacted one. A solution of a stereospecific carboxylesterase in an appropriate aqueous buffer serves as the aqueous phase. A mixture of the substrates (RICOOR2 and R3OH) constitutes the organic phase; that is, to enhance the productivity, no additional solvent for the substrate is used. Using pig liver carboxylesterase and lipase from Candida cylindracea as stereospecific catalysts, we have implemented transesterifications in biphasic systems to preparatively resolve both racemic alcohols (R3 contains a chiral carbon) and racemic esters (R1 contains a chiral carbon). ~4,~5 Experimental Design Conceptually, the system described above represents an emulsion of an aqueous solution of an enzyme in a mixture of substrates. However, from the experimental standpoint it is advantageous instead to use porous supports whose pores are filled with an aqueous solution of an enzyme. First, such beads are more mechanically robust than water droplets; second, they can be readily separated from products at the end of the enzymatic reaction; third, they can be used repeatedly. The organic (substrate) phase should be presaturated with an aqueous buffer (of the pH optimal for the enzyme action) to avoid loss of water from the beads due to its partitioning. Three different porous supports are used for entrapment of stereospecific esterases: Sepharose, Chromosorb, and titania. A racemic alcohol or ester is dissolved in a molar excess of the second substrate, the mixture is saturated with an aqueous buffer solution, and then the enzyme confined to the pores of a solid support is added. The system is vigorously shaken at room temperature. Periodically, the organic phase is analyzed by gas chromatography. After the reaction is completed, the enzyme-containing beads are recovered by filtration, and the organic phase obtained is separated by aqueous extractions and/or fractional distillation.

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

CH3CH2COCH3 + HOCH2CH2R

121

-CH3CH2COCH2CH2R÷ CH30H

R = CH(0CH3)CH3 (3-meth0xy-l-buton0l) R = CH(CH3)CH2CH3 (3-methyl-l-pentonol)

R = CH(CH3)CH2CH2CH2CH(CH3) 2 (3,7-dlmethYl-l-octanol) R = CH(CH3)CH2CH2CH=C(CH3) 2

(cltronello])

FIG. 1. Hog liver carboxylesterase-catalyzed resolution of racemic primary alcohols (1-4) via transestefification reactions.

Preparative Resolution of Racemic Primary Alcohols Catalyzed by Pig Liver Carboxylesterase

Materials Pig liver carboxylesterase (EC 3.1.1.1) was purchased from Sigma Chemical Co. in the form of a suspension in 3.2 M ammonium sulfate, and had a specific activity of 160 units/mg protein. Porous supports used for "immobilization" of the enzyme, Sepharose 4B and Chromosorb 101, were also obtained from Sigma. The former was an aqueous suspension of 40- to 190-~m beads (agarose concentration 4%). Chromosorb 101 (commonly used as a carrier in gas-liquid chromatography) consisted of 125- to 150-/xm particles with pore diameter of 0.30.4/xm. All primary alcohols used in this work, both in the racemic and in the optically pure form (when available), were obtained commercially. Propionic esters of 1-4 (see Fig. 1) were synthesized from propionyl chloride and the corresponding alcohol following the general procedure of Morris and Green. 18The boiling points of the propionic esters thus prepared were as follows (760 mm Hg): of 1, 181°; of 2, 174°; of 3, 342°; and of 4, 250°. Methyl propionate employed in this study as a "matrix" ester was obtained from Aldrich and was used without further purification.

Methods Assays. All alcohols and esters in this work are determined gas chromatographically using a 6-foot glass column packed with Analab's Super Pak 20 M (N2 cartier gas, 10 ml/min; detector and injector port temperatures 250°). (1) In the case of 1, the temperature of the column is increased 18 j. F. Morris and E. H. Green, J. Am. Chem. Soc. 26, 293 (1901).

122

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

from 51 ° to 240 ° at 16°/min. The retention times observed are 1.6 min for methyl propionate and methanol, 6.4 min for 1, 7.1 min for 3-methoxy-1butyl propionate, and 5.2 min for propionic acid. (2) In the case of 2, the temperature of the column is 90 °. The retention times observed are 1.1 min for methyl propionate and methanol, 2.0 min for propionic acid, 2.8 min for 2, and 3.6 min for 3-methyl-l-pentyl propionate. (3) In the case of 3 and 4, the temperature of the column is 135°. The retention times observed are 1.0 min for methyl propionate and methanol, 1.5 min for propionic acid, 3.0 min for 3, 4.1 min for 3,7-dimethyl-l-octyl propionate, 3.5 min for 4, and 4.5 min for citronellyl propionate. All optical rotations are measured at 589 nm (sodium line) and 25 ° using a Perkin-Elmer 243B polarimeter. Preparation of the Enzymatic Catalyst. Solutions of hog liver carboxylesterase are entrapped either in Sepharose or in Chromosorb as described below. (1) Swollen Sepharose 4B, 2ml, is washed with water and with 0.1 M phosphate buffer (pH 8.0). The resin is thoroughly dried on a glass filter, and then the resultant gel is cut into small (approximately 2 x 2 mm) pieces with a razor blade. The beads produced are added to 4.6 ml of an enzyme suspension, 10.9 mg/ml, and allowed to absorb it (there was no liquid left afterward). (2) Chromosorb 101 (100-120 mesh), 1 g, is washed with water and with 0.1 M phosphate buffer, and then dried on a glass filter. After that, the beads are added to 1.1 ml of an enzyme suspension, 10.9 mg/ml, and allowed to completely absorb it. We have found that (for reasons not presently clear) Sepharose is a better support in the transesterification with 1 as a nucleophile, while Chromosorb is superior with 2, 3, and 4 (in terms of the half-life of the reaction at a given amount of the enzyme present in the system). Both enzymatic preparations are stored at 4 ° for several days with no appreciable loss of the catalytic activity. Alkaline Hydrolysis of Esters. Enzymatically prepared optically active esters of propionic acid can be converted to the corresponding optically active alcohols by alkaline hydrolysis. Propionic esters of 1, 2, 3, and 4 are hydrolyzed at pH 12 and 37 ° in a pH-stat's cuvette. Saturated solutions of the esters in 0.1 M aqueous solution of KCI with 10% of acetone are placed in a 100 ml water-jacketed cuvette and are hydrolyzed until completion (several hours). Then the solutions are partly evaporated under vacuum and the alcohols are extracted with ether, followed by its evaporation.

Results Hogs liver carboxylesterase entrapped in Sepharose or Chromosorb was used to asymmetrically catalyze transesterifications shown in Fig. 1:

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

123

methyl propionate (an inexpensive and readily available ester) was employed as a matrix in which the methyl moiety was stereoselectively replaced with alcohols 1-4. (-)-3-Methoxy-l-butanol (1) was dissolved in a molar excess of methyl propionate and the mixture was saturated with phosphate buffer, followed by addition of carboxylesterase entrappedsin Sepharose beads. The system was shaken at room temperature; periodically, the organic phase was analyzed gas chromatographically. Figure 2 shows the time course of the enzymatic transesterification. One can see that the concentration of 1 gradually decreases, while the concentration of the transesterification product, 3-methoxy-l-butyl propionate, concomitantly increases (no propionic acid is formed, indicating that hydrolysis was completely suppressed); after 20 hr no further reaction takes place. The enzyme-containing beads were then recovered by filtration. The organic phase obtained was separated by aqueous extractions and evaporation under vacuum. As a result, optically active (+)-1 and (-)-3-methoxy-1butyl propionate were collected. We have found that the former was

c o n c e nt r a t ion, M 2

I

fI

14 hours

I

Ii

28

FIG. 2. The time course of the reaction between methyl propionate and (-)-1 catalyzed by hog liver carboxylesterase entrapped in Sepharose beads. (O) Disappearance of methyl propionate; (C]) accumulation of (-)-3-methoxy-1-butyl propionate. The organic phase consisted of methyl propionate containing 2 M (-+)-1 (and was saturated with 0.1 M phosphate buffer, pH 8.0). The aqueous phase contained 9600 units of the enzyme entrapped in 6 g of Sepharose beads swollen with 0.1 M phosphate buffer (pH 8.0). The mixture was shaken at room temperature.

124

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

totally unreactive in the esterase-catalyzed reaction with methyl propionate. On the other hand, (-)-1 produced from the (-)-ester by an alkaline hydrolysis was nearly fully acylated in the enzymatic transesterification. These results suggest that hog liver carboxylesterase has essentially absolute stereoselectivity in the reaction of 1 with methyl propionate, and that the transesterification stops at a 50% conversion (Fig. 2) because the reactive (-)-isomer is exhausted. The enzymatic catalyst recovered after the preparative synthesis was rinsed with ether presaturated with phosphate buffer (to remove accumulated methanol), and repeatedly employed for the preparative transesterification. After each use, no more than a quarter of the enzymatic activity was lost (as judged on the basis of the half-life of the enzymatic reaction). Sepharose is by no means a unique support for the enzyme entrapment; e.g., we have successfully used porcine liver carboxylesterase confined to Chromosorb beads for preparative production of three other couples of optically active esters and alcohols. Using the same approach as described above for 1, we have found that only S- (but not R-) isomers of 2 and 3 can act as nucleophiles in the transesterification reaction shown in Fig. 1 catalyzed by the carboxylesterase. As a result, gram quantities of R(+)-2 and -3 [and their S(-)-propionic esters] were prepared with the enantiomeric excess of 94 and 97%, respectively. Both ( - ) - and (+)-fl-citronellols (4) are natural products which are widely used in perfumery. 19,20Although methods for the synthesis of the optical isomers of 4 have been developed, they are rather inefficient and usually afford low optical purity, zl-24 (---)-4 can be easily made by reduction of synthetic citronellal or geranio119,2° but its resolution, as well as that of most racemic alcohols, 3 is very cumbersome. We have attempted to employ our enzymatic approach for the resolution of (+-)-4. Starting with 20 g of racemic 4, 6.9 g of R(+)-4 [ee (enantiomeric excess) = 96%] and 4.8 g of S(-)-4 (ee -- 92%) were obtained. As in the case of the other alcohols tested, only S(-)-4 was reactive toward the enzyme. Hence hog liver carboxylesterase exhibits absolute S-stereoselectivity in the transesterification reaction. Remarkably, the enzyme recognizes configuration of 19 A. B. Booth, in "Kirck-Othmer Encyclopedia of Chemical Technology," Vol. 19, 2nd Ed., p. 803. Wiley (Interscience), New York, 1969. 2o "The Merck Index," 9th Ed., p. 310. Merck, Rahway, New Jersey, 1976. 21 R. von Rienacker and G. Ohroff, Agnew. Chem. 73, 240 (1961). 22 R. von Rienacker, Chimia 27, 97 (1973). 23 T. Shono, Y. Matsmara, K. Hibino, and S. Miyawaki, Tetrahedron Lett. 15, 1295 (1974). 24 M. Hidai, H. Ishiwatari, H. Yagi, E. Tanaka, K. Onozawa, and Y. Uchida, Chem. Commun. 170 (1975).

[12]

P R O D U C T I O N O F O P T I C A L L Y ACTIVE C O M P O U N D S

125

the chiral carbon which is two methylene groups away from the alcohol's reactive hydroxyl. It is significant that there is a close parallelism between the stereospecificity of the enzyme in the transesterification reaction (Fig. 1) and that in the reaction of hydrolysis. We have found that in water (pH 8.0) hog liver carboxylesterase catalyzes the hydrolysis of only the (-)-isomer of propionic esters of 1, 2, 3, and 4 (the reactions were carried out to completion in a pH-stat's cuvette; then the products were extracted with ether and characterized). An ideal enzyme to be used as a catalyst in organic chemistry has to meet the following two requirements: (1) it should have a very broad substrate specificity, and (2) it should display a marked selectivity (e.g., stereo) in the conversion of its substrates. Porcine liver carboxylesterase clearly satisfies the latter requirement: it is stereospecific with respect to the nucleophile in reaction (1). However, it does not meet the former requirement. We found that the carboxylesterase has a rather narrow specificity with regard to the alcohols it accepts as nucleophiles in the transesterification reaction; only primary alcohols having no substituents in the first two methylene groups adjacent to the hydroxyl react at appreciable rates. Since a number of chiral secondary alcohols are of practical significance, it was necessary to extend our methodology to them, i.e., to identify an estei-ase that will be both stereospecific and able to accept secondary alcohols as nucleophiles in the transesterification reaction. This work is described in the next section.

Procedure Enzymatic Production of (-)- and (+)-3-Methoxy-l-butanol. A mixture containing 92 ml of methyl propionate and 28 ml of racemic 1 is saturated with 0.1 M aqueous sodium phosphate (pH 8.0). Following removal of the aqueous phase on a separation funnel, 6 g of Sepharose beads containing hog liver carboxylesterase is added. The mixture is transferred to a 200-ml screw-cap bottle and shaken on an orbit shaker at 250 rpm and room temperature for 20 hrs. Then the enzymatic catalyst is recovered by filtration, and the liquid phase obtained is washed with 3 volumes of distilled water. The washings are combined and evaporated at 15 mm Hg; at 50 ° 11.7 g of (+)-1 (93% yield, 97% purity as judged by gas chromatography) with [a]D = 15.1° (neat) was collected. The organic phase remaining after aqueous extraction is evaporated at 15 mm Hg; at 40 ° 16.7 g of (-)-3-methoxy-l-butyl propionate (80% yield, 98% purity by GC) with [a]o = -16.3 ° (neat) was obtained. The ester is hydrolyzed as described above, resulting in 6.1 g of (-)-1 (48% overall yield, 95% purity

126

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

by GC) with [ a ] D = - - 14.3 ° (c = 9, methanol). To our knowledge, optically active 1 has not been synthesized before. Therefore, we could not determine the enantiomeric excess of the enzymatically prepared (+)- and ( - ) 1 (or the (-)-ester).

Enzymatic Production of R(+)-2 and S(-)-Methyl-l-pentyl Propionate. A mixture containing 22.3 ml of methyl propionate and 7.7 ml of racemic 2 is saturated with the phosphate buffer as described above and then 2 g of Chromosorb beads with hog liver carboxylesterase is added. The mixture is placed in a 50-ml screw-cap bottle and shaken for 53 hr. Following r e c o v e r y of the enzymatic catalyst, methanol and methyl propionate from the organic phase are removed by rotary evaporation and the remainder is distilled at 10 mm Hg. At 54 ° 2.6 g of R(+)-2 (75% yield, 96% purity by GC) with [a]D = 9.8 ° (C = 10, methanol), which corresponds to ee = 94% [lit. 25 [a]D = 10.4 ° (C = 10, methanol)], was obtained. At 71 ° 2.5 g of S ( - ) - 3 - m e t h y l - l - p e n t y l propionate (44% yield, 94% purity by gas chromatography) with [odd = - 8 . 6 ° (c = 10, methanol) was collected.

Enzymatic Production of R( +)-3 and S(-)-3, 7-Dimethyl-l-octyl Propionate. The procedure is the same as for 2 except the reaction time is 96 hrs. Upon distillation at 10 mm Hg, at 105 ° 3.3 g of R(+)-3 (61% yield, 96% purity by GC) with [~]D = 4.9 ° (C = 5, ether), which corresponds to ee = 97% [lit. 26 [C~]D = 5.1 ° (C = 5.03, ether)], was obtained. At 121 ° 3.0 g of S(-)-3,7-dimethyl-l-octyl propionate (40% yield, 95% purity by GC) with [aiD = --5.1 ° (C = 5, ether) was collected. Enzymatic Production of R(+)- and S(-)-4. A mixture containing 39.7 ml of methyl propionate and 24.8 ml of (-+)-4 is saturated with the buffer as described above and 4.9 g of Chromosorb beads with hog liver carboxylesterase is added. The mixture is placed in a 100-ml screw-cap bottle and shaken for 96 hr. Then the beads are filtered out, and the organic phase is distilled at 10 mm Hg. At 108 ° 6.9 g of R(+)-4 (69 ° yield, 96% purity by GC) with [a]D = 5.0 ° (neat), which corresponds to ee = 96% [lit. 26 [a]D = 5.2 ° (neat)], was obtained. At 123° 8.0 g of S(-)-citronellyl propionate (58% yield, 97% purity by GC) with [a]D = --3.8 ° (neat) was collected. The ester is hydrolyzed as described in Methods, resulting in 4.9 g of S ( - ) - 4 (48% overall yield, 94% purity by GC) with [a]D = --4.4 ° (C = 20, methanol), which corresponds to ee -- 92% [as compared to [a]D = --4.8 ° (C = 20, methanol) for the authentic samples of S ( - ) - 4 purchased from Pfaltz and Bauer]. 25 K. Mori, T. Suguro, and M. Uchida, Tetrahedron 34, 3119 (1978). 26 N. Cohen, C. Scott, C. Neukom, R. J. Lopresti, G. Weber, and G. Saucy, Helv. Chim. Acta 64, 1158 (1981).

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

127

Preparative Resolution of Racemic Secondary Alcohols Catalyzed by Yeast Lipase

Materials Lipase from Candida cylindracea (EC 3.1.1.3, triacylglycerol lipase) was purchased from Sigma Chemical Co.; it was a powder with a specific activity of 2415 units/mg protein. Chromosorb I01, described in the preceding section was used as a support for immobilization. All secondary alcohols used in this work, both in the racemic and in the optically pure form (when available), were obtained commercially. Butyric esters of 5-10 (see Fig. 3) were synthesized from butyryl chloride and the corresponding alcohol following the general procedure of Morris and Green. ~8The boiling points of the butyric esters thus prepared were as follows (760 mm Hg): of 5, 151°; of 6,210°; of 7, 250°; of 8, 223°; of 9, 170°; and of 10, 220 °. Tributyrin, employed as a matrix ester, was obtained from Sigma and was used without further purification.

Methods Assays. All alcohols and esters in this work are determined gas chromatographically using a 6-foot glass column packed with Analab's "Super Pak 20 M " (N2 carrier gas, 10 ml/min; detector and injector port temperatures 250°). The temperature of the column is increased from 51 ° to 240° at 16°/min. The retention times observed are 11-13 min for tributyrin and

0 TRIBUTYRIN + H O C H R I R 2 ~ CH3CH2CH2~0CHRi R2 + DIBUTYRIN

Z

Ri = CH3

R2 = CH2CH3

RI = CH3

R2 = (CH2)5CH3

(2-BUTANOL)

R! = CH3

R2 = C6H5

(2-OCTANOL)

(SEC....N-PHENETHYLALCOHOL)

RI = CH3 R2 = CH2CH2CH=C(CH3)2 RI = CH3

R2 = CH2Ct

RI = H

R2 = CH(CL)CH20

I_I RI = H

R2 = CH(0H)CH2CH3

10

(6-METHYL-S-HEPTEN-2-OL)

(I-CHLORO-2-PROPANOL) (2,3-DICHLOROPROPANOL) (I,2-BUTANEDIOL)

FIG. 3. Yeast lipase-catalyzed resolution of racemic alcohols (5-11) via transesterification reactions.

128

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

dibutyrin (a broad shapeless peak), 6.0 min for butyric acid, 1.8 min for 5, 3.2 min for 2-butyl butyrate, 5.0 min for 6, 6.6 min for 2-octyl butyrate, 6.6 min for 7, 7.8 min for sec-phenethyl butyrate, 5.0 rain for 8, 6.7 min for 6methyl-5-hepten-2-yl butyrate, 3.0 min for 9, 4.9 min for 1-chloro-2-propyl butyrate, 6. I min for 10, 7.4 min for 2,3-dichloroprop-l-yl butyrate, 6.1 min for 11, and 7.1 min for 2-hydroxybut-l-yl butyrate. All optical rotations are measured at 589 nm (sodium line) and 25 ° using a Perkin-Elmer 243 B polarimeter. Preparation of the Enzymatic Catalyst. Lipase from C. cylindracea is entrapped in Chromosorb beads. One gram of the beads is impregnated with 0.1 M phosphate buffer (pH 8.0), and they are dried on a filter paper. Then 67 mg of the enzyme powder is added to the beads and mixed thoroughly, and the mixture is allowed to equilibrate at room temperature for 5 hr. Alkaline Hydrolysis of Esters. Enzymatically prepared optically active butyric esters can be converted to the corresponding optically active alcohols by alkaline hydrolysis. (1) Butyric esters of 5 and 11 are hydrolyzed at pH 12 and 37° in a pH-stat's cuvette. Saturated solutions of the esters in 0.1 M aqueous solution of KCl with 10% of acetone are placed in a 100-ml water-jacketed cuvette and hydrolyzed until completion (several hours). Then the solutions are partly evaporated under vacuum and the alcohols are extracted with ether, followed by its evaporation. (2) Butyric esters of 6, 7, and 8 (about 10 g) are dissolved in 50 ml of dry methanol saturated with KOH. The esters are completely solvolyzed within 1 hr (as judged by gas chromatography). Then methanol is evaporated under vacuum and the precipitated KOH is removed by filtration; the resultant solution is washed with ether and dried with anhydrous MgSO4. (3) In the case of butyric esters of 9 and 10, the alkaline treatment results both in hydrolysis and in epoxidation 27 of the alcohols produced. Butyric ester of 9, 11.9 g, and butyric ester of 10, 10.5 g, are dissolved in I 1 ml of dry nbutanol and 12 ml of dry methanol, respectively. Then the resultant solutions are added dropwise to 50 ml of dry n-butanol or methanol, respectively, saturated with KOH, at 4 °. The mixtures are stirred at room temperature for 1 hr during which time the epoxidation is completed as judged by gas chromatography. (Under condition 1 of the first Assays section, the retention times of both authentic, commercially obtained and enzymatically prepared propylene oxide and epichlorohydrin were 1.3 and 2.5 min, respectively.) Then the mixtures are stored at 4 ° overnight to precipitate KCI, and filtered; the solvents are evaporated under vacuum; the solutions are filtered again, washed with ether, and dried with anhydrous MgSO4. 27 N. F. Blau, J. W. Johnson, and C. G. Stuckwisch, J. A m . Chem. Soc. 76, 5106 (1954).

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

129

Epoxidation of 9 and 10. The procedure used is a modification of the method of Blau et al. 27 and is identical to that described in (3) of the preceding section except that 6.3 g of 9 and 8.2 g of 10 are used instead of their butyric esters.

Results

As mentioned earlier, the major drawback of hog liver carboxylesterase as a practical catalyst is its narrow nucleophile specificity. To overcome this obstacle, we have endeavored to search for another commercially available esterase which will be both stereoselective and nonspecific in the transesterification process. Such an esterase was found among lipases. Since the natural function of lipases is to hydrolyze triglycerides, 28-3° one of them, tributyrin (glyceryl tributyrate), was employed by us as a matrix ester. This ester is inexpensive and readily available. (Another likely candidate, triacetin, is too soluble in water and also is less reactive with most lipases. 29 Triglycerides of fatty acids are inferior because of their high molecular weight and cost per mole.) We have discovered that lipase from the yeast C. cylindracea, when placed in a biphasic system, where the aqueous phase is a solution of the enzyme in water confined to the pores of Chromosorb and the organic phase is a solution of an alcohol in tributyrin, (1) accepts a wide variety of alcohols (both secondary and primary) in the transesterification reaction (Fig. 3), and (2) if the alcohol is chiral, displays a marked stereoselectivity in accepting only one isomer as a nucleophile. In the case of all chiral alcohols tested, the time course of lipase-catalyzed transesterifications resembled that shown in Fig. 2. In all cases the process was carried out to apparent completion; no significant amount of butyric acid was detected (i.e., enzymatic transesterification nearly totally suppressed hydrolysis). Table I shows some examples of preparative resolution of secondary alcohols catalyzed by yeast lipase. The general scheme of this process was conceptually the same as in hog liver carboxylesterase-catalyzed resolutions and is illustrated in Fig. 4. To test the reusability of lipase entrapped in the Chromosorb beads, the enzymatic catalyst was used for four consecutive transesterifications 28 p. Desnuelle, in "The Enzymes" (P. D. Boyer, ed.), Vol. 7, 3rd Ed., p. 575. Academic Press, New York, 1972. 29 H. Brockerhoff and R. G. Jensen, "Lipolytic Enzymes." Academic Press, New York, 1974. 30 K. M. Shahani, in "Enzymes in Food Processing" (G. Reed, ed.), 2nd Ed., Chap. 8. Academic Press, New York, 1975.

130

[12]

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

TABLE I PREPARATIVE RESOLUTION OF RACEMIC SECONDARY ALCOHOLS VIA TRANSESTERIFICATION IN BIPHASIC SYSTEMS CATALYZED BY YEAST LIPASEa Production of Unreactive isomer

Alcohol

Configuration of the reactive isomer

Reaction time (days)

2-Butanol 2-Octanol sec-Phenethyl alcohol 6-Methyl-5-hepten-2-olc

R R R S

1.6 5.8 6.2 5.3

Yield (g) 4.9 7.7 7.9 8.3

(70%) (77%) (79%) (83%)

ee (%) 89 95 88 91

Reactive isomer b Yield (g) 4.6 6.0 6.1 5.8

(65%) (60%) (61%) (58%)

ee (%) 93 92 85

In all cases the organic phase consisted of 2 M solution of the racemic alcohol in tributyrin; the aqueous phase consisted of a solution of lipase from Candida cylindracea in 0.1 M phosphate (pH 8.0). The mixtures were shaken at room temperature, followed by removal of the catalyst by filtration and separation of the products by distillation. b Obtained from the enzymatically produced butyric esters by alkaline hydrolysis, followed by ether extraction. ' Resolution of optical isomers of this alcohol (also called sulcatol) is particularly interesting because their nonequimolar mixture serves as an aggregation pheromone of a scolytid beetle [K. J. Byrne, A. A. Swigar, R. M. Silverstein, J. H. Borden, and E. Stokkink, J. Insect. Physiol. 20, 1895 (1974)].

with 5 as a nucleophile. After each resolution, about 75% of the initial enzymatic activity was retained. Enzymatically prepared optically active alcohols and esters can then be chemically converted to other optically active compounds. We have illustrated this in the lipase-catalyzed transesterification (Fig. 3) using 9 and 10 as nucleophiles. In each case only the R-isomer was reactive. Both the remaining S-alcohols and butyric esters of the R-alcohols were converted to epoxides using KOH treatment, z2 All four optically active epoxides produced are useful as chiral building blocks, 31 for example: R(+)propylene oxide in the synthesis of R-reciferolide 3z or chiral a_methylene-y-lactones33; S(-)-propylene oxide for the production of Ssulcatop4; both R ( - ) - and S(+)-epichlorohydrin for the synthesis of a variety of pharmaceuticals. 31'35 3~ A. S. Rao, S. K. Paknikar, and J. G. Kirtane, Tetrahedron 39, 2323 (1983). 32 K. Utimoto, K. Uchida, M. Yamaga, and H. Nozaki, Tetrahedron Lett. 18, 3641 (1977). 33 L. D. Martin and J. K. Stille, J. Org. Chem. 47, 3630 (1982). 34 B. D. Johnston and K. N. Slesser, Can. J. Chem. 57, 233 (1979). 35 j. j. Baldwin, A. W. Raab, K. Mensler, B. H. Alison, and D. E. McClure, J. Org. Chem. 43, 4876 (1978).

[12] matrix ester

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

esterase

racemic

+

alcohol

>

filtration

• o o• •

new, optically active ester

washing_

+

131

remaining optically active alcohol

reuse

I organic phase I aq ueousd~Xtrr oc t ion n under vacuum

optically active alcohol (Ist enantiomer)

new, optically active ester I ~= alkaline hydrolysis optically active

alcohol (2nd enantiomer) FIG. 4. Schematic illustration of the resolution of racemic alcohols using esterase-catalyzed transesterifications.

All the alcohols enzymatically resolved thus far were monohydric. It was of obvious interest to explore whether the lipase-catalyzed transesterifications can be used for the preparation of optically active dials; 1,2butanediol was chosen as a model. To avoid formation of both 1- and 2esters, the enzyme, in addition to stereoselectivity, must also exhibit positional specificity. To examine that we studied the lipase-catalyzed transesterification with 2 M 1-butanol (mimicking the primary hydroxyl in 11) and 2 M 5 (mimicking the secondary hydroxyl in 11) as nucleophiles. It was determined that the half-life of the former reaction exceeds that of the latter by the factor of 12; that is, the enzyme is indeed regiospecific. This led to the facile enzymatic resolution according to the scheme in Fig. 4: from 20 g to racemic 11, 7.2 g o f R ( + ) - l l (ee = 88%) and 7.0 g of S ( - ) - l l (ee = 90%) (the unreactive and reactive isomers, respectively) were prepared.

Procedure Enzymatic Production of S(+)- and R(-)-5. A mixture containing 77.7 ml of tributyrin and 17.3 ml of racemic 5 is equilibrated with the phosphate buffer, and 3.8 g of Chromosorb beads with yeast lipase is added. The mixture is placed in a 200-ml screw-cap bottle and shaken at 250 rpm and room temperature for 40 hr. Then the enzymatic catalyst is recovered

132

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

by filtration, and the organic phase is distilled at 18 mm Hg. At 28 ° 4.8 g of S(+)-5 (69% yield, 97% purity by GC) with [a]D = 12.0 ° (neat), which corresponds to ee = 89% [lit. 36 [a]D = 13.5 ° (neat)], was obtained. At 50 ° 10.2 g o f R ( - ) - 2 - b u t y l butyrate (75% yield, 98% purity by GC) with laID = --11.3 ° (neat) was collected. The ester was hydrolyzed as described above, resulting in 4.6 g of R ( - ) - 5 (65% overall yield, 95% purity by GC) with [a]D = --12.8 ° (C = 5, ether), which corresponds to ee = 93% [as compared to [a]D = --13.8 ° (C = 5, ether) for the authentic sample of R ( - ) - 5 purchased from Fluka]. Enzymatic Production of S(+)- and R(-)-6. The procedure employed is the same as for 5 except the mixture contains 52 ml of tributyrin, 25 ml of racemic 6, and 3.1 g of the catalytic beads, and it is shaken for 140 hr; distillation is done at 4 mm Hg. At 33 ° 7.7 g of S(+)-6 (77% yield, 95% purity by GC) with [a]D = 9.4 ° (neat), which corresponds to ee = 95% [lit. 37 [OdD = 9.9 ° (neat)], was obtained. At 50 ° 12.6 g of R ( - ) - 2 - o c t y l butyrate (82% yield 97% purity by GC) with [~]D = --11.8 ° (neat) was collected. Alkaline hydrolysis of the ester resulted in 6.0 g of R ( - ) - 6 (60% overall yield, 96% purity by GC) with [OdD = - 9 . 1 ° (neat), which corresponds to ee = 92% [lit. 37 [a]D = --9.9 ° (neat)]. Enzymatic Production of S(-)- and R( +)-7. The procedure employed is the same as for 6 except the mixture consists of 63.3 ml of tributyrin, 19.8 ml of racemic 7, and 3.3 g of the catalytic beads, and is shaken for 150 hr. Upon vacuum distillation, at 44 ° 7.9 g of S ( - ) - 7 (79% yield, 94% purity by GC) with [a]D = --36.4 ° (neat), which corresponds to ee = 88% [lit. 38 [et]D = --41.3 ° (neat)], was obtained. At 62 ° 11.8 g of R(+)-sec-phenethyl butyrate (82% yield, 93% purity by GC) with [a]D = 42.3 ° (neat) was collected. Alkaline hydrolysis of the ester resulted in 6.1 g of R(+)-7 (61% overall yield, 97% purity by GC) with [a]D = 33.6 ° (neat), which corresponds to ee = 85% [lit. 38 [a]D = 39.5 ° (neat)]. Enzymatic Production of R(-)- and S(+)-8. The procedure employed is the same as for 6 except the mixture consists of 53.5 ml of tributyrin and 23.8 ml of racemic 8, and it is shaken for 128 hr. Upon vacuum distillation, at 30 ° 8.3 g of R ( - ) - 8 (93% yield, 97% purity by GC) with [a]D = --16.9 ° (neat), which corresponds to ee = 91% [lit. 34 [C~]D = --18.5 ° (neat)], was obtained. At 44 ° 12.2 g of S(+)-6-methyl-5-hepten-2-yl butyrate (79% yield, 96% purity by GC) with [a]D ---- 14.8 ° (neat) was collected. Alkaline hydrolysis of the ester resulted in 5.8 g of S(+)-8 (58% overall yield, 95% 36 "The Merck Index," 9th Ed., p. 197. Merck, Rahway, New Jersey, 1976. 37 y . y . Lin, D. N. Palmer, and J. B. Jones, Can. J. Chem. 52, 469 (1973). 38 "Aldrich Catalog/Handbook of Fine Chemicals," p. 918. Aldrich, Milwaukee, Wisconsin, 1982-1983.

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

133

purity by GC) with [O/]D = 16.5 ° (neat), which corresponds to ee = 95% [lit. 34 [Ct]D = 18.6 ° (neat)]. Enzymatic Production of S(-)- and R( + )-Propylene Oxide. A mixture consisting of 79.1 ml of tributyrin and 16.2 ml of racemic 9 is saturated with the buffer, and 3.8 g of the lipase-containing Chromosorb beads is added. The mixture is placed in a 200-ml screw-cap bottle and shaken for 52 hr. Following separation of the catalyst, the organic phase is distilled at 4 mm Hg. At 18° 6.3 g of S(-)-9 (70% yield, 95% purity by GC) with [a]D = -- 5.7 ° (neat) were obtained; this alcohol was converted to 2.4 g of S ( - ) propylene oxide (43% overall yield, 93% purity by GC) with [a]D = --4.9 ° (C = 1, chloroform) which corresponds to ee = 68% [lit. 39 laiD = --7.2 ° (C ---- 1, chloroform)]. At 28 ° 11.9 g of R(+)-l-chloro-2-propyl butyrate (76% yield, 94% purity by GC) with [a]D = 6.3 ° (neat) was collected; the ester was converted (see above) to 2.7 g of R(+)-propylene oxide (50% overall yield, 94% purity by GC) with [a]D = 4.8 ° (C = 1, chloroform) which corresponds to ee = 67% [lit. 39 [ a ] D = 7.2 ° (c = 1, chloroform)]. Enzyma tic Production of S (+)- a nd R (-) -Epichlorohydrin. The procedure employed is the same as for propylene oxide except the mixture consists of 62 ml of tributyrin, 15 ml of racemic 10, and 3.1 g of the enzymatic catalyst, and is shaken for 50 hr. At 28 ° 8.2 g of S(+)-10 (82% yield, 94% purity by GC) with [aid = 5.6 ° (neat) was obtained; this alcohol was converted to 3.6 g of S(+)-epichlorohydrin (51% overall yield, 95% purity by GC) with [a]D = 23.5 ° (C = 1.2, methanol) which corresponds to ee = 71% [lit. 35 [aid -- 33.0 ° (C = 1.2, methanol)]. At 52 ° 10.5 g of R(-)-2,3-dichloro-l-propyl butyrate (70% yield, 95% purity by GC) with [aid = --8.9 ° (neat) was collected; the ester was converted (see above) to 3.0 g of R(-)-epichlorohydrin (43% overall yield, 95% purity by GC) with [a]D = --23.0 ° (C = 1.5, methanol) which corresponds to ee -67% [lit. 35 [a]D = --34.3 ° (c = 1,5, methanol)]. It is worth noting that the most likely reason for a relatively low enantiomeric excess obtained is temperature-induced racemization of optically active 9 and 10 (and their butyric esters). In agreement with this hypothesis, when the enzymatically produced 9 was distilled at a lower vacuum (20 mm Hg) and higher temperature (38 °) only 10% ee was obtained for the resultant propylene oxide. Enzymatic Production of R(+)- and S(-)-ll. A mixture consisting of 92 ml of tributyrin and 20 ml of racemic 11 is saturated with the buffer, and 4.4 g of the lipase-containing Chromosorb beads is added, followed by shaking for 36 hr. The beads are removed by filtration and the resultant 39 "Aldrich Catalog/Handbook of Fine Chemicals," p. 997. Aldrich, Milwaukee, Wisconsin, 1982-1983.

134

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

liquid is washed with 3 volumes of water. The washings are combined and evaporated in a rotary evaporator; the remainder is washed with ether and dried with anhydrous MgSO4. As a result, 7.2 g of R ( + ) - l l (72% yield, 96% purity by GC) with [a]D = 11.0 ° (C = 16, ethanol), which corresponds to ee = 89% [lit. 4° [aiD = 12.4 ° (C = 16, ethanol)], was obtained. The organic phase left after aqueous extraction is distilled at 5 mm Hg. At 50 ° 13.9 g of (-)-1,2-butyl butyrate (79% yield, 95% purity by GC) with [a]D = --14.6 ° (neat) was collected; this ester was converted to 7.0 g of S ( - ) - l l (70% overall yield, 94% purity by GC) with [a]D = --11.6 ° (C = 2.5, ethanol) which corresponds to ee = 90% [lit. 41 [a]D = --12.9 ° (C = 2.5, ethanol)]. Preparative Resolution of Racemic Esters Catalyzed by Yeast Lipase

Materials Lipase from Candida cylindracea (EC 3.1.1.3) was purchased from Sigma Chemical Co.; it was a powder with a specific activity of 2415 units/ mg protein. The enzyme was confined to the pores of titania. Titania ST6A was a generous gift of Dr. N. R. Clark of Sterling Organics and was in the form of 400- to 700-/zm particles with pore diameter of 2000-10,000 ,~. Racemic 2-(p-chlorophenoxy)propionic acid and n-butanol were obtained from Aldrich. Methyl 2-(p-chlorophenoxy)propionate and butyl 2(p-chlorophenoxy)propionate were synthesized by us using Fischer esterification. 42

Methods Assays. All esters and alcohols are determined by gas chromatography as described in the previous section. The retention times observed are 1.5 min for methanol, 1.6 min for butanol, 9.1 min for 12, and 11.0 min for butyl 2-(p-chlorophenoxy)propionate. Optical rotations are measured at 589 nm (sodium line) and 25 ° using a Perkin-Elmer 243B polarimeter. Preparation of the Enzymatic Catalyst. Candida cylindracea lipase is entrapped in the pores of titania beads. One gram of the beads is impregnated with 0.1 M phosphate buffer (pH 8.0), followed by drying on a filter 4o K. Mori, M. Sasaki, S. Tamada, T. Suguro, and S. Masuda, Heterocycles 10, 111 (1978). 4~ K. Mori, M. Sasaki, S. Tamada, T. Suguro, and S. Masuda, Tetrahedron 35, 1601 (1979). 4: B. A. E. MacClement, R. G. Carrier, D. J. Phelps, and P. R. Carey, Biochemistry 20, 3438 (1981).

[12]

135

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

C,~/~O~HCOOCH3 CH3

+ HOCH2CH2CH2CH3~C,~/-~O~HCOOCH2CH2CH2CH3 ~ CH3

+ HOCH3

~2 FIG. 5. Yeast lipase-catalyzed resolution of racemic ester 12 via a transesterification reaction.

paper. Then 300 mg of the enzyme powder is added to the beads, mixed thoroughly, and the mixture allowed to equilibrate for 5 hr at room temperature. Alkaline Hydrolysis of Esters. Compound 12 and the corresponding butyl ester are hydrolyzed at pH 12.0 in 0.1 M aqueous solution of KCI containing 20% (v/v) of acetone at 37°. The reaction (followed potentiometrically) is complete after 4 hr. Then the pH is adjusted to pH 1.0 with concentrated HC1, followed by extraction of 1 with ether. The etheral solution is dried with anhydrous MgSO4 and evaporated under vacuum.

Results Thus far, we have used esterase-catalyzed asymmetric transesterifications in biphasic systems for preparative resolution of racemic alcohols (Fig. 4). The technology we have developed, 15 however, is not limited to that particular class of compounds. To illustrate this point, we have applied our method for the resolution of a racemic ester. Derivatives of 2-(p-chlorophenoxy)propionic acid are widely used as herbicides 43 with only one enantiomer being biologically active. 44 We have attempted to resolve methyl ester of racemic 2-(p-chlorophenoxy)propionic acid (12) using the reaction shown in Fig. 5. In this case nbutanol (the smallest water-immiscible alcohol) was used both as the "matrix" nucleophile and the solvent for 12. (---)-12 is dissolved in a molar excess of n-butanol, and the mixture is saturated with phosphate buffer, followed by addition of lipase entrapped in titania beads. The mixture is shaken at room temperature and the extent of the enzyme-catalyzed transesterification is followed gas chromatographically. After 30 hr the degree of conversion reaches 50% and the transesterification stops. Subsequent analysis of the products (see below) has shown that this is due to absolute stereospecificity of yeast lipase: only the R-isomer of the ester is reactive toward the enzyme and could be transesterified. This affords a facile preparative resolution of 43 R. Cremlin, "Pesticides: Preparation and Mode of Action," p. 143. Wiley, Chichester, England, 1978. 44 K. H. Buchel (ed)., "Chemistry of Pesticides," p. 357. Wiley, New York, 1983.

136

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[12]

racemic 12 via enzymatic transesterification in the biphasic system described above. Procedure

Twenty milliliters of 2 M solution of 12 in n-butanol is saturated with 0.1 M phosphate buffer (pH 8.0). Then 3 g of the titania beads containing 0.7 g of lipase is added. The mixture is placed in a 100-ml screw-cap bottle and shaken on an orbit shaker at 250 rpm and 25 °. Periodically, 1-/xl aliquots are withdrawn and analyzed gas chromatographically as described above. When the reaction stops, the enzyme-containing beads are separated from the liquid phase by filtration, followed by its vacuum (4-5 mm Hg) distillation. At 85 ° 3.5 g of (S)-12 (81% yield, 98% purity by GC) which has a specific optical rotation [a]D = --41.1 ° (C ---- 50, ethanol) is obtained; the latter corresponds to an enantiomeric excess of 96%. 45 At 95 ° 4.1 g of (R)-butyl 2-(p-chlorophenoxy)propionate (80% yield, 97% purity by GC) with [a]D = +43.5 ° (C = 50, ethanol) (which corresponds to ee of 97% 45) is obtained. In both cases the ee is determined on the basis of the optical purity of the acid produced from the ester by alkaline hydrolysis. The enzymatically produced (S)-12 can be racemized and subjected to another enzyme-catalyzed transesterification, thereby leading to a continuous process for the preparation of the biologically active R-isomer 44from the racemic mixture. We have found that the racemization can be conveniently based on the Claisen condensation: incubation of a methanol solution of the S-ester at 60° in the presence of catalytic amounts of sodium methoxide (prepared by dissolution of metallic sodium in methanol in the 1 : 10, w/w, ratio) for 16 hr affords complete racemization of the ester. Concluding Remarks As reported in this paper, our approach, enzymatic stereoselective transesterification in biphasic systems, has been successfully tested with two different esterases, hog liver carboxylesterase and yeast lipase. As a result, a number of optically active alcohols and esters were produced with high yield and enantiomeric purity. It appears that our work represents a significant advance in the use of enzymes for preparative production of chiral compounds 6-8 as the methodology developed (1) is very simple, efficient, and readily amenable to a scaleup; (2) is applicable to a wide variety of water-insoluble substances; (3) employs commercially available enzymes; (4) does not require expensive cofactors; and (5) is 45 D. T. Witiak, T. C.-L. Ho, R. E. Hakney, and W. E. Connor, J. Med. Chem. 11, 1086 (1968).

[12]

PRODUCTION OF OPTICALLY ACTIVE COMPOUNDS

137

carried out in the presence of only about 1% of water, which greatly simplifies product recovery. A natural extension of the biphasic system with a low water content would be a monophasic, purely organic medium. We have recently found 46 that at least one esterase, porcine pancreatic lipase, can vigorously function in nearly anhydrous organic solvents. Under such conditions the enzyme acquires remarkable new properties; for example, it becomes extremely thermostable and more specific. 46 We believe that the use of enzymes in aqueous-organic or organic media represents one of the most promising directions in enzyme technology. Addendum In addressing the basic question of how much water in the reaction medium enzymes actually require, we discovered that that amount in many cases can be close to z e r o . 47 This phenomenon has led to a new area of enzyme technology, biocatalysis in organic s o l v e n t s . 48'49 When placed in nonaqueous solvents instead of water, enzymes become extremely thermostable, 46 exhibit a new substrate specificity, 5° and can catalyze reactions impossible in aqueous solutions. 47 The last feature has been utilized by us to carry out novel efficient bioconversions including lipasecatalyzed regioselective acylation of glycols 5~ and sugars 52 and stereoselective conversions 53 in dry organic solvents, polyphenol oxidase-catalyzed regiospecific oxidations of phenols in organic media, 54 alcohol dehydrogenase-catalyzed asymmetric oxidoreductions in nonaqueous milieu, 55 and peroxidase-catalyzed depolymerization of lignin in organic media (but not in water)) 6 Biocatalysis in organic solvents indeed promises to have a profound impact on both basic and applied enzymology.

46 A Zaks and A. M. Klibanov, Science 224, 1249 (1984). 47 A. Zaks and A. M. Klibanov, Proc. Natl. Acad. Sci. U.S.A. 82, 3192 (1985). 48 A. M. Klibanov, CHEMTECH 16, 354 (1986). 49 M. Waks, Proteins 1, 4 (1986). 50 A. Zaks and A. M. Klibanov, J. Am. Chem. Soc. 108, 2767 (1986). 5t p. Cesti, A. Zaks, and A. M. Klibanov, Appl. Biochem. Biotechnol. 11, 401 (1985). 52 M. Therisod and A. M. Klibanov, J. Am. Chem. Soc. 108, 5638 (1986). 53 G. Kirchner, M. P. Scollar, and A. M. Klibanov, J. Am. Chem. Soc. 107, 7072 (1985). 54 R. Z. Kazandjian and A. M. Klibanov, J. Am. Chem. Soc. 107, 5448 (1985). 55 j. Grunwald, B. Wirz, M. P. Scollar, and A. M. Klibanov, J. Am. Chem. Soc. 108, 6732 (1986). 56 j. S. Dordick, M. A. Marietta, and A. M. Klibanov, Proc. Natl. Acad. Sci. U.S.A. 83, 6255 (1986).

138

I M M O B I L I Z E D E N Z Y M E S / C E L L S IN O R G A N I C S Y N T H E S I S

[13]

[13] T w o - L i q u i d P h a s e B i o c a t a l y t i c R e a c t o r s

By M. D. LILLY, S. HARBRON,and T. J. NARENDRANATHAN Two liquid phase biocatalytic reactors have been developed over the last I0 years to meet the growing need for biological catalysis of selective and stereospecific conversions of a range of water-immiscible and waterinsoluble organic compounds of interest to the chemical and biological industries. In these systems the organic phase consists of either the reactant alone or the reactant dissolved in a water-immiscible solvent. Advantages and disadvantages of two-liquid phase systems have been described elsewhere.l,2 One particular advantage of this approach is the possibility of operating the reactor with a large proportion of the volume occupied by organic phase which can be readily separated from the aqueous phase at the end of the reaction. This overcomes the usual need to work in dilute aqueous solutions with all the associated problems of product recovery. The development of biocatalytic reactors for these conversions involving a water-immiscible organic phase requires an understanding of both the kinetic properties of the biocatalyst and the diffusion and mass transfer characteristics of the reactants and products. A knowledge of the influence that the organic phase has on the biocatalyst and the factors likely to contribute to the long-term stability of the biocatalyst is also needed. Types of Reactions It is worth distinguishing in a simple manner between the possible configurations of two-liquid phase biocatalytic reactions so that a rational approach can be made to a better understanding of their kinetic behavior. Such a scheme is given in Table I in which the reaction types are classified according to four criteria. It is important to distinguish first between those reactions which take place in the aqueous phase and those which occur at the aqueous-organic interface. Representative examples of systems where the reaction site is in the aqueous phase or at the aqueous-organic interface are given in Table II and Table III, respectively. The second major distinction depends on how much aqueous component is added: if a discrete aqueous phase is present then it may be either the continuous (type 1) or the discontinuous phase (type 2). For this i M . D . L i l l y , J. Chem. Technol. Biotechnol. 32, 162 (1982). 2 M . D . L i l l y , Philos. Trans. R. Soc. London, Ser. B 300, 391 (1983).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[13]

TWO-LIQUID PHASE BIOCATALYTIC REACTORS

139

TABLE I CLASSIFICATIONOF Two-LIQUID PHASE BIOCATALYTICREACTIONS Criterion Reaction site Aqueous phase

Biocatalyst Organic phase

Possible configuration Aqueous phase Aqueous-organic interface Discrete aqueous continuous phase (type 1) Discrete aqueous discontinuous phase (type 2) No discrete aqueous phase (type 3) Free Immobilized No added solvent Added solvent

purpose "discrete" indicates the presence of a definite aqueous phase external to the biocatalyst. This is illustrated in Fig. la, which relates to reactor types 1 and 2, and shows the substrate concentration profile arising from mass transfer across the aqueous-organic interface to the surface of the biocatalyst (immobilized enzyme, free or immobilized cell). For reactions involving enzymes dissolved in the aqueous phase, the corresponding concentration profiles are shown in Fig. lb. For simplicity TABLE II TYPES OF Two-LIQUID PHASE BIOCATALYTICREACTIONS:AQUEOUSPHASE REACTION Catalyst Type 1. Discrete aqueous continuous phase Free biocatalyst No added solvent Bacillus subtilis Nocardia corallina Pseudomonas oleovorans Pseudomonas putida Various microorganisms Added solvent Steroid 20/3-dehydrogenase

Nocardia rhodochrous Pseudomonas oleovorans

Reaction

Substrate (phase ratio)

Ester hydrolysis Epoxidation Epoxidation Epoxidation Ester hydrolysis

Menthyl acetate (0.05-1.0) a Tetradecene (0.05) b Octene (0.02-1.0F 1,7-Octadiene (0.02-1.00) a Menthyl acetate (0.04) e

20/3-Reduction

Prednisolone (1.0Y Cortisol (1.0) f Cortisone (1.0)Y Deoxycorticosterone (1.0y Prednisone (1.0) r Cholesterol (0.8)g 1,7-Octadiene (0.02-1.0) h

3r-Dehydrogenation Epoxidation

(continued)

140

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[13]

TABLE II (continued) Catalyst Immobilized biocatalyst No added solvent Pseudomonas putida Added solvent Steroid 3a-dehydrogenase Steroid fl-dehydrogenase Steroid 20fl-dehydrogenase Type 2. Discrete aqueous discontinuous phase Free biocatalyst No added solvent Bacillus subtilis Pseudomonas oleovorans Added solvent Chymotrypsin Nocardia erythropolis Nocardia rhodochrous Pseudomonas oleovorans Rhodotorula minuta Immobilized biocatalyst Added solvent Chymotrypsin Nocardia erythropolis Type 3. No discrete aqueous phase Free biocatalyst No added solvent No examples Added solvent Nocardia rhodochrous

Nocardia rhodochrous Nocardia rhodochrous Rhodotorula minuta Immobilized biocatalyst No added solvent Bacillus subtilis

Reaction

Substrate (phase ratio)

Epoxidation

1,7-Octadiene (0.02-1.0) h

3a-Dehydrogenation fl-Dehydrogenation 20/3-Reduction

Androsterone (0.5) ~ Testosterone (0.5) ~ Cortisone (0.5) /

Ester hydrolysis Epoxidation

Menthyl acetate (1.0-2.3) ° Octene (1.0-4.0y

Ester formation

N-Acetyltryptophan and leucineamide (5-50) / Cholesterol (1.5) k Cholesterol (1.1-2.9F 1,7-Octadiene (1.0-1.63) h Menthyl succinate (3.33) I

3/3-Dehydrogenation 3/3-Dehydrogenation Epoxidation Ester hydrolysis

Ester formation 3fl-Dehydrogenation

N-benzoylphenylalanine and ethanol (15)" Cholesterol (0.53-0.67) k

Ester hydrolysis

Cholesterol (3.7-10F," Dehydroepiandrosterone (10).,o ¢i-Sitosterol (10)" Stigmasterol (10)" Pregnenalone (10)° Testosterone (10)P Androstenedione (15-20)q Testosterone (10)P Menthyl succinate (10)~

Ester hydrolysis

Menthyl acetate (7)~

3/3-Dehydrogenation

17/3-Dehydrogenation A~-Dehydrogenation

[13]

TWO-LIQUID PHASE BIOCATALYTIC REACTORS

141

TABLE II (continued) Catalyst Added solvent Chymotrypsin

Reaction

Ester hydrolysis Ester formation

Lipoxygenase Nocardia erythropolis Nocardia rhodochrous

Fat oxidation 3/3-Dehydrogenation 3/3-Dehydrogenation

Nocardia rhodochrous Nocardia rhodochrous

AI-Dehydrogenation

Rhodotorula minuta

Ester hydrolysis

17/3-Dehydrogenation

Substrate (phase ratio)

N-Acetyltyrosine ethyl ester (100) r N-Acetyltryptophan and ethanol (100)r Linoleic acid (4)s Cholesterol (0.78-1.13) k Cholesterol (2.5-10) n,° Dehydroepiandrosterone (2.5-10)~, o Pregnenalone (2.5-10) ° fl-Sitosterol (2.5-10y Stigmasterol (2.5-10) n Testosterone (2.5-10)P Androstenedione (15-20)q Testosterone (2.5-10)p Menthyl succinate (2-2.5) t

"I. K. Brookes, Ph.D. Thesis, University of London, 1984. b K. Furuhashi, A. Taoka, S. Uchida, I. Karube, and S. Suzuki, Eur. J. Appl. Microbiol. Biotechnol. 12, 39 (1981). ¢ M. J. de Smet, H. Wynberg, and B. Witholt, Appl. Environ. Microbiol. 42, 811 (1981). d S. Harbron, T. J. Narendranathan, and M. D. Lilly, Proc. 3rd Eur. Congr. Biotechnol. 4, 369 (1985). e y. Yamaguchi, T. Oritami, N. Tajima, A. Komatsu, and T. Moroe, Nippon Nogei Kagaku Kaishi 50, 475 (1976). s p. Cremonesi, G, Carrea, F. L. Ferrara, and E. Antonini, Biotechnol. Bioeng, 17, 1101 (1975). g B. C. Buckland, P. Dunnill, and M. D. Lilly, Biotechnol. Bioeng. 17, 815 (1975). h R. D. Schwartz and C. J. McCoy, Appl. Environ. Microbiol. 34, 47 (1977). i G. Carrea, F. Colombi, G. Mazzola, P. Cremonesi, and E. Antonini, Biotechnol. Bioeng. 21, 39 (1979). J A. N. Semenov, I. V. Berezin, and K. Martinek, Biotechnol. Bioeng. 23, 355 (1981). k p. Atrat, E. Huller, and C. Horhold, Z. Allg. Mikrobiol. 20, 7 (1980). l T. Omata, N. Iwamoto, T. Kimura, A. Tanaka, and S. Fukui, Eur. J. Appl. Microbiol. Biotechnol. U , 199 (1981). " K. Martinek, A. N. Semenov, and I. V. Berezin, Biotechnol. Bioeng. 23, 1115 (1981). n T. Omata, T. Iida, A. Tanaka, and S. Fukui, Eur. J. Appl. Microbiol, Biotechnol. 8, 143 (1979). o T. Omata, A. Tanaka, and S. Fukui, J. Ferment. Technol. 58, 339 (1980). P S. Fukui, S. A. Ahmed, T. Omata, and A. Tanaka, Eur. J. Appl. Microbiol. Biotechnol. 10, 289 (1980). q T. Yamane, H. Nakatani, E. Sada, T. Omata, A. Tanaka, and S. Fukui, Biotechnol. Bioeng. 21, 2133 (1979). • A. M. Klibanov, G. P. Samokhin, K. Martinek, and I. V. Berezin, Biotechnol. Bioeng. 19, 1351 (1977). s T. Yamane, Enzyme Eng. 6, 141 (1982).

142

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[13]

T A B L ~ III TYPES OF Two-LIQUID PHASE BIOCATALYTIC REACTIONS: REACTION AT INTERFACE

Catalyst

Reaction

Substrate (phase ratio)

Free biocatalyst

Rhizopus arrhizus Immobilized biocatalyst Lipase

Triglyceride hydrolysis

Olive oil (50) a

Interesterification

Olive oil and stearic

Interesterification

Myristic acid and palm midfraction (35-85) c

acid (22.5) b

Lipase

G. Bell, J. R. Todd, J. A. Blain, J. D. E. Patterson, and C. E. Shaw, Biotechnol. Bioeng. 23, 1703 (1981). b K. Yokezeki, S. Yamanaki, K. Takinami, Y. Hirose, A. Tanaka, K. Sonomoto, and S. Fukui, Eur. J. Appl. Microbiol. Biotechnol. 14, 1 (1982). c R. Wisdom, P. Dunnill, M. D. Lilly, and A. Macrae, Enzyme Microbiol. Technol. 7, 567 (1985). a

it has been assumed that the aqueous phase is well mixed (i.e., C 4 = C [ ) . With microbial cells, as the biocatalyst concentration is increased, mixing in the aqueous phase is reduced, and eventually a cell paste is obtained with no discrete aqueous phase outside the biocatalyst as shown in Fig. lc (reactor type 3). A similar situation can occur when immobilized cells or enzymes are used, for instance, in a packed bed. The substrate concentra(a) BIOCATALYST

AOUEOUS PHASE

OROANICPHASE

:film lfitm ',

I i C~

( b} BIOCATALYST

ORGANIC PHASE

(d,~ved in a~leous phase)

(c) BIOCATALYST

ORGANIC PHASE

:f+tmIf+t~_

FIG. 1. Concentration profiles of reactant at steady state in a two-liquid phase systems. Mass transfer from the organic phase to the biocatalyst. (a) Discrete aqueous phase, separate biocatalyst; (b) discrete aqueous phase, biocatalyst dissolved in aqueous phase; (c) increased biocatalyst concentration such that there is no aqueous phase.

[13]

143

TWO-LIQUID PHASE BIOCATALYTIC REACTORS

tion profile within the biocatalyst (Fig. la, b, c) will depend on the nature, form, and intrinsic activity of the biocatalyst. For an immobilized biocatalyst reactor with a discrete aqueous phase, it has been assumed in Tables II and III that the immobilized particles remain associated with the discrete aqueous phase, and reactor types 1 and 2 are possible. However, if the particles are found predominantly in the organic phase, then a type 3 configuration will result. Further subdivisions in Table I relate to whether the biocatalyst is free or immobilized and whether the organic phase contains added solvent. In Tables II and III the phase ratio is defined as the volume of organic phase per volume of aqueous phase including biocatalyst. In our own experiments with either octadiene or menthyl acetate in the absence of biocatalyst, phase inversion occurred between 1.2 and 1.6. 3 In many of the twoliquid phase systems described in the literature, no information is given on the phase ratio at which inversion occurs. We have therefore assumed that at phase ratios below one, the aqueous phase is continuous and at phase ratios above unity the organic phase is continuous. Kinetic Considerations Once a system has been classified in this manner it is possible to develop equations to describe its performance. For a reaction catalyzed by cells or immobilized biocatalyst with discrete aqueous phase (Fig. la), the following relationships 4 for steady-state conditions can be obtained: ds/dt = v, the aqueous phase volumetric reaction rate = k2A2(C1

--

---- k 3 A 3 ( C 3

--

C2) C4)

(1) (2) (3)

(4)

= ksAs(C~ - C5) C3 = K32C2 C5 -- K56C6

(5) (6)

where k, A, C, and K are mass transfer coefficients (liter • T-l), areas (liter 2 • liter-3), substrate concentrations (mol • liter-3), and partition coefficients, respectively. Equations (1) to (6) give C6 \K32 ~

= 1

K32C, \k2Az + ~

+~

3 T. J. Narendranathan, S. Harbron, and M. D. Lilly, unpublished results. 4 S. Harbron, T. J. Narendranathan, and M. D. Lilly, P r o c . 3 r d Eur. C o n g r . 369 (1985).

(7)

B i o t e c h n o l . 4,

144

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[13]

3= zoo-

100

0

0.0

a0z

oh6 o.b8 o.io

o.i2

0i4

g DCW/ml aqueousphaseor g dry foam

FIG. 2. The conversion of 1,7-octadiene to 7,8-epoxy-l-octene by free (filled symbols) and entrapped (open symbols) Pseudomonas putida in small tubes agitated on a reciprocating arm shaker at 1300 strokes per minute. Phase ratios: (©) 0.11; (A) 0.43; ([~) 1.00.

The right-hand side of Eq. (7) is a measure of the effectiveness of the overall mass transfer, and ranges between 0 (high mass transfer resistance) and 1 (low mass transfer resistance). The relationship between v and C6 will depend on the nature and form of the biocatalyst, and if the biocatalyst is immobilized, will need to take into account the diffusion of the substrate into and product out of the support. From a knowledge of the kinetic parameters involved it is therefore possible to predict the behavior of the reactor. An illustration of the influence of phase ratio and biocatalyst concentration on the performance of two-liquid phase reactors is shown in Fig. 2. 5 As the concentration of free cells in the aqueous phase is increased, the system becomes progressively limited by the rate of transfer of the substrate, resulting in the observed fall in specific activity. The effect is less pronounced at higher phase ratios, since an increase in the phase ratio toward the point of inversion gives an increased interracial area per unit volume of aqueous phase, 3 and the mass transfer limitation is thereby reduced. The ability to reduce mass transfer limitations in this way is offset by the increased volume of the organic phase. When this is taken into account the overall volumetric rate for the reactor is reduced at higher phase ratios. With immobilized biocatalyst, diffusion into the catalyst may be the rate-limiting step, in which case changes in phase ratio will have little or no effect. This diffusion limitation will become more pronounced as the s S. Harbron and M. D. Lilly, unpublished results.

[13]

TWO-LIQUID PHASE BIOCATALYTIC REACTORS

145

cell loading in the biocatalyst is increased, resulting in an observed fall in specific activity. Such a situation with immobilized Pseudomonas putida is illustrated in Fig. 2. Agitation is the other major factor likely to influence the interfacial area. For example, in the conversion of cholesterol to cholestenone by Nocardia rhodochrous 6 in a stirred tank reactor, an increase in agitation from 9 to 15 rev/sec gave a corresponding increase in reaction rate. This increase was attributed to an increased interfacial area between the organic phase and the aqueous cell suspension. For menthyl ester hydrolysis by Bacillus subtilis the specific reaction rate in a 50-ml working volume reactor was independent of agitation above 13 rev/sec. Under these conditions the specific reaction rate decreased above and below an organic phase ratio of 0.3 to 0.35. At a lower speed of 6.5 rev/sec, the optimum phase ratio shifted to about 1.0. 7 Effect of Organic Solvent

The organic solvent can have two general effects on two-liquid phase biocatalytic reactions. First, it can affect the partitioning of reactants and products between the aqueous and organic phase. This can be used to advantage if the product has an inhibitory effect on the biocatalyst, as for example with the epoxidation activity ofPseudomonas oleovorans 8 which is inhibited by the accumulation of epoxide product in the aqueous phase. Addition of relatively large volumes of cyclohexane resulted in the favorable partitioning of the epoxide into the organic phase and over 90% conversion of the reactant, octadiene, was achieved as a result. In another case N. rhodochrous entrapped in a hydrophilic resin had activity that was closely related to the partition coefficient of the reactant cholesterol, between the resin and different organic solvents used. 9 A similar relationship was obtained when the same solvent was used, but cells were entrapped in a range of different resins. 9,1° Second, the solvent may have a deleterious effect on the structure of the enzyme and the integrity of the cell. Several workers have observed large differences in catalytic activities with different solvents. Various organic extraction reagents have been examined for their toxic and inhibi-

6 B. C. Buckland, P. Dunnill, and M. D. Lilly, Biotechnol. Bioeng. 17, 815 (1975). 7 I. K. Brookes, Ph.D. Thesis, University of London, England, 1984. 8 R. D. Schwartz and C. J. McCoy, Appl. Environ. Microbiol. 34, 47 (1977). 9 T. Omata, A. Tanaka, and S. Fukui, J. Ferment. Technol. 58, 339 (1980). i0 T. Omata, T. Iida, A. Tanaka, and S. Fukui, Eur. J. Appl. Microbiol. Biotechnol. 8, 143 (1979).

146

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[13]

tory properties on anaerobic bacteria. 11 At concentrations above those required to saturate the aqueous phase, alcohols (C5-C12), ketones (C5C8), benzene derivatives, isoamyl acetate, and diisopropyl ether were toxic, whereas alkanes (C6-Clz) and diisoamyl ether were not. For the conversion of cholesterol to cholestenone by N. rhodochrous, properties of the solvent which appear to minimize any damage have been reported to be low polarity, 6 low solubility in water, lz and a low value for the dielectric constant. 9 Similar observations have been made with 20fl-hydroxysteroid dehydrogenase. 13 Immobilization of the cell or enzyme can reduce the magnitude of any deleterious effects of organic solvents, fl-Hydroxysteroid dehydrogenase immobilized on Sepharose gave a preparation that retained about 60% of its activity after 2 months of continuous use in the presence of ethyl acetate, whereas the free cell preparation under similar conditions lost 80-90% of its activity after 4-5 days.14 Long-term stability of Rhodotorula minuta, which catalyzes the conversion of menthyl succinate to menthol, could similarly be achieved by immobilization. 15Free cells had a half-life of about 50 hr whereas some immobilized cell preparations had a half-life of 1520 hr. Immobilization of N. rhodochrous in polyacrylamide was found to protect the cholesterol degradation pathway from the organic solvent in which the cholesterol was dissolved. 2 Choice of Reactors For a reaction which is not limited by diffusion or mass transfer, the degree of conversion achieved in a packed bed reactor will be higher than in a continuous stirred tank reactor, except at very high substrate concentrations and low conversion rates. If the reaction is diffusion limited then a packed bed reactor becomes kinetically even more favorable than the well-mixed continuous reactor. The kinetic behavior of a batch reactor is similar to that of a packed bed reactor but the overall productivity is less because of reactor downtime. For the above and other reasons 16 packed beds of immobilized enzymes and cells have been widely used for reactions involving water-soluble reactants and products. H M. J. Playne and B. R. Smith, Biotechnol. Bioeng. 25, 1251 (1983). 12j. M. C. Duarte and M. D. Lilly, Enzyme Eng. 5, 363 (1980). ~3 p. Cremonesi, G. Carrea, L. Ferrara, and E. Antonini, Biotechnol. Bioeng. 27, 1101 (1975). 14 G. Carrea, F. Colombi, G. Mazzola, P. Cremonesi, and E. Antonini, Biotechnol. Bioeng. l l , 39 (1979). 15 T. Omata, N. Iwamoto, T. Kimura, A. Tanaka, and S. Fukui, Eur. J. Appl. Microbiol. Biotechnol. 11, 199 (1981).

[13]

TWO-LIQUID PHASE BIOCATALYTIC REACTORS

147

For two-liquid phase reactions it is essential to create sufficient interfacial area to allow adequate rates of transfer of the reactant and product between the phases. Much of the published experimental work has been done in shaken tubes and there are only a few reports on the use of reactors, either stirred tank or packed bed. Lipase-catalyzed interesterification reactions have been carried out in both batch stirred tank reactors and continuous packed bed reactors.17 Since dried catalyst particles have virtually no activity, up to 10% (v/v) water was added to hydrate the catalyst particles even though this resulted in some hydrolysis of the substrate. These catalyst particles were used for 600 hr in the packed bed and for 10 successive batches each of about 16 hr duration in the stirred tank reactor without appreciable loss of activity. A packed bed reactor gave better yields than a stirred tank reactor for the continuous hydrolysis of triglyceride in organic solvent systems using immobilized lipase from Rhizopus arrhizus. 18 The conversion of cholesterol to cholestenone by Nocardia 6 was studied in a 500-ml stirred tank reactor and the reaction rate found to be independent of agitation above a stirrer speed of 15 rev/sec. The cells retained 52% of their activity after seven batches lasting a total of 69 hr. A large stirred tank is also being used in the United States for a two-liquid phase steroid conversion. Experimental Methods Immobilization Techniques Only a limited number of immobilization techniques have been used so far in two-liquid phase systems. Although in some cases support materials already used for completely aqueous reaction systems have been used (porous glass, 19Sepharose,14 polyacrylamide, and alginate12), several new techniques have been developed which are particularly suitable for use in two-liquid phase systems. Several examples of methods are given below. Using hydrophobic photo-cross-linkable resin prepolymers. 2° One gram thawed cells suspended in a 2 ml of water-saturated solvent under ~6M. D. Lilly and P. Dunnill, this series, Vol. 44, p. 717. 17 A. R. Macrae, J. Am. Oil Chem. Soc. 60, 243A (1983). 18 G. Bell, J. R. Todd, J. A. Blain, J. D. E. Patterson, and C. E. Shaw, Biotechnol. Bioeng. 23, 1703 (1981). 19 A. M. Klibanov, G. P. Samokhin, K. Martinek, and I. V. Berezin, Biotechnol. Bioeng. 19, 1351 (1977). 2o S. Fukui, S. A. Ahmed, T. Omata, and A. Tanaka, Eur. J. Appl. Microbiol. Biotechnol. 10, 289 (1980).

148

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[13]

mild sonication is mixed with 1 ml of same solvent containing 1 g hydrophobic photo-cross-linkable resin polymer (ENTP-2000, 2~ ENTB-1000, or mixtures of ENTP-2000 and ENT-4000) and 10 mg photosensitizer (benzoin ethyl ether). The mixture is placed on a sheet of transparent polyester and illuminated with near-UV light for 3 min. The resin gel formed is cut into 1.5-mm-square pieces for use. Using hydrophilic photo-cross-linkable resin prepolymers. 2° The procedure for the immobilization of hydrophilic photo-cross-linkable resin prepolymers (ENT 4000 or mixtures of ENT-4000 and ENTP-2000) is the same as above, except potassium phosphate buffer (pH 7.0, 20 mM) is substituted for the organic solvent. Using urethane prepolymers. 4,z° One gram thawed cells is suspended under mild sonication in 2 ml of potassium phosphate buffer (pH 7.0, 20 mM) and mixed with previously melted 1 g urethane prepolymer (PU3, PU6, or mixtures of PU3 and PU6) at 40 °. The gel formed is cut into 3-mm cubes for use. Polyurethane beads 4 are prepared by pouring the above mixture into 50 ml liquid paraffin being agitated at 1250 rpm. After 20 min the liquid paraffin is decanted, and the beads formed are washed with buffer, dried between filter paper, and stored on ice till used. Immobilization on Celite. 22 Two parts of the lipase from Rhizopus delemar (7200 U/g) are dissolved in 20 parts of distilled water and 5 parts Celite added with stirring at 0°. Then 30 parts acetone are added over 5 min and stirring is continued for 30 min more. The solid product formed is filtered off and dried at 20° under reduced pressure. Reactor Systems Stirred tank reactors. 6 A glass stirred tank (0.5 liter working volume, 75-mm diameter) with four stainless-steel baffles (7 mm width) is used for the oxidation of cholesterol to cholestenone by N. rhodochorous. Agitation is carried out with a four-blade fiat plate turbine impeller (22 mm diameter) located 22 mm above the reactor bottom and driven by a variable speed (0-6000 rpm) motor. Either air is supplied by a small pump, or oxygen is fed directly from a pressurized cylinder to the gas inlet which is a capillary glass tube ending directly beneath the impeller. A reflux condenser is fitted to the air outlet of the reactor. 2t Abbreviations: ENTP-2000, a derivative of polypropylene glycol 2000; ENTB-1000, a derivative of polybutadiene; ENT-4000, a derivative of polyethylene glycol 4000; PU3, a polyester diol of urethane containing 57% polyethylene glycol; and PU6, a polyester diol of urethane containing 91% polyethylene glycol. 22 M. H. Coleman and A. R. Macrae, United Kingdom Patent 1,577,933 (1980).

[13]

TWO-LIQUID PHASE BIOCATALYTIC REACTORS

149

One hundred grams of thawed Nocardia cells suspended in 200 ml of carbon tetrachloride containing 16% (w/v) cholesterol (at 20°C, oxygen flow rate 100 ml/min and stirrer speed 2200 rpm) converted all the cholesterol to cholestenone in 5.5 hr. The reaction rate was constant at 35 g/liter-hr for the first 4 hr. Shaken f l a s k reactor. 2° One gram of thawed Nocardia cells were washed with water-saturated solvent and mixed with a given amount of crystalline electron acceptor (if required) by mild sonication. This mixture is then added to a solution of steroid (e.g., testosterone) in 10 ml solvent in a 30-ml glass-stoppered conical flask and incubated in a thermostated bath at 30° with shaking at 120 strokes/min. A reaction mixture containing 1 g thawed cells and 173/zmol of testosterone and 16/zmol of phenazine methosulfate (electron acceptor) in 10 ml water-saturated benzene-n-heptane solvent (4 : 1 v/v) mixture at 30° converted 167/zmol of testosterone in 2 hr. P a c k e d bed reactor. 22 A mixture of palm midfraction (1 part) and myristic acid (0.4 parts) dissolved in 100-120C petroleum ether (3.2 parts) is saturated with water and then continuously pumped at a flow rate of 22 ml/hr through a bed of hydrated catalyst (5.0 g) prepared from Rhizopus niveus lipase and kieselguhr. The reactor temperature is 40 °. Almost complete (greater than 95%) interesterification was obtained during a 400-hr reactor operation. Future Prospects The examples listed in Tables II and III indicate the increasing interest in conversions involving water-insoluble or water-immiscible reactants and products. Because of their complex nature much more work needs to be done before our understanding of the kinetic behavior of these reactions reaches the same level as that for reactions involving water-soluble reactants and products. Nevertheless the operational stabilities that are being reported for some systems are encouraging, and it seems likely that several commercial processes will be introduced in the next few years.

150

IMMOBILIZED

ENZYMES/CELLS

IN ORGANIC SYNTHESIS

[14]

[14] E n z y m e - C a t a l y z e d S t e r o i d T r a n s f o r m a t i o n s in Water-Organic Solvent Two-Phase Systems

By

GIACOMO CARREA a n d PIERO CREMONESI

The potential of enzyme-catalyzed reactions in organic chemistry is now remarkable. This is due partly to the increased number of available enzymes, to their use in the immobilized form, and to the existence of efficient systems of coenzyme regeneration. However, substrates poorly soluble in water such as steroids present special difficulties in this area since the enzyme-catalyzed transformations must be carried out using large reaction volumes and consequently large amounts of biological catalysts and cofactors as required. Also, the use of large reaction volumes makes the recovery of products more difficult. Attempts to increase steroid solubility by adding water-miscible organic solvents to the reaction medium have given unsatisfactory results since high solvent concentrations progressively give rise to inhibition, decreased specificity, and denaturation of the enzymes.l In contrast, when organic solvents which are practically immiscible or poorly miscible with water are used, the situation becomes substantially improved. In this case, a two-phase system consisting of water and an organic solvent is established. The water phase contains enzymes and hydrophilic cofactors, and the organic phase high concentrations of hydrophobic substrates. On shaking, substrates diffuse into the aqueous phase where they undergo the enzyme-catalyzed transformation and then the formed products return to the organic phase. In this system, the concentration of solvents in water is low and not dependent on the ratio of the two phases, even if the partial volume of the organic phase is much greater than that of the aqueous phase, and the inhibitory and denaturing effects are much lower than those induced by comparable concentrations of miscible solvents. 2-8 L. G. Butler, Enzyme Microb. Technol. 1, 253 (1979). 2 p. Cremonesi, G. Carrea, G. Sportoletti, and E. Antonini, Arch. Biochem. Biophys. 159, 7 (1973). 3 G. Lugaro, G. Carrea, P. Cremonesi, M. M. Casellato, and E. Antonini, Arch. Biochem. Biophys. 159, I (1973). 4 p. Cremonesi, G. Carrea, L. Ferrara, and E. Antonini, Eur. J. Biochem. 44, 401 (1974). s G. Carrea, P. Cremonesi, M. M. Casellato, and E. Antonini, Steroids Lipids Res. 5, 162 (1974). 6 p. Cremonesi, G. Carrea, L. Ferrara, and E. Antonini, Biotechnol. Bioeng. 17, 1101 (1975).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[14]

ENZYMATIC CATALYSIS IN TWO-PHASE SYSTEMS

151

In the present article, the specific oxidation-reduction of hydroxylketo groups of steroids, catalyzed by NAD(P)÷-dependent hydroxysteroid dehydrogenases, is described and discussed. The reactions are carried out in two-phase systems and are coupled to the enzymatic regeneration of the coenzymes. Materials

20fl-Hydroxysteroid dehydrogenase (17,20/3,21-trihydroxysteroid: NAD ÷ oxidoreductase, EC 1.1.1.53) (specific activity, l0 U/mg protein), yeast alcohol dehydrogenase (alcohol:NAD ÷ oxidoreductase, EC 1.1.1.1) (specific activity, 200 U/mg protein), formate dehydrogenase (formate : NAD ÷ oxidoreductase, EC 1.2.1.2) (specific activity, 3 U/mg protein), glutamate dehydrogenase (L-glutamate : NAD(P) ÷ oxidoreductase, EC 1.4.1.3) (specific activity, 90 U/mg protein), L-lactate dehydrogenase (L-lactate : NAD ÷ oxidoreductase, EC 1.1.1.27) (specific activity, 220 U/ mg protein), NAD ÷, NADP ÷, NADH, and NADPH are obtained from Boehringer-Mannheim. 3 (or 17)fl-Hydroxysteroid dehydrogenase (fl-hydroxysteroid : NAD(P) ÷ oxidoreductase, EC 1.1.1.51) (specific activity, 18 U/mg protein) and steroids are purchased from Sigma. 12a-Hydroxysteroid dehydrogenase (specific activity 2 U/mg protein) is extracted from Clostridium group P as described by Macdonald et al. 9 Sepharose 4B and Sepharose CL-4B are obtained from Pharmacia and CNBr and tresyl chloride from Fluka. All other reagents and compounds are of analytical grade. Assays

Enzymatic activity is measured in a 3-ml cuvette by spectrophotometric monitoring at 340 nm of the formation or consumption of NAD(P)H. The conditions for the various assays are as follows: 20fl-hydroxysteroid dehydrogenase in 0.05 M potassium phosphate buffer, pH 7, containing 0.17 mM NADH and 0.18 mM cortisone; fl-hydroxysteroid dehydrogenase in 0.1 M sodium phosphate buffer, pH 8.5, containing 0.3 mM NAD ÷ and 0.05 mM testosterone; 12a-hydroxysteroid dehydrogenase in 0.1 M potassium phosphate buffer, pH 8, containing 0.3 mM NADP ÷ and 1 mM cholic acid; alcohol dehydrogenase in 0.1 M sodium phosphate 7 G. Carrea, S. Colombi, G. Mazzola, P. Cremonesi, and E. Antonini, Biotechnol. Bioeng. 21, 39 (1979). 8 E. Antonini, G. Carrea, and P. Cremonesi, Enzyme Microb. Technol. 3, 291 (1981). 9 j. A. Macdonald, J. F. Jellet, and D. E. Mahony, J. Lipid Res. 20, 234 (1979).

152

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[14]

buffer, pH 9, containing 4 mM NAD + and 0.3 M ethanol; formate dehydrogenase in 0. I M potassium phosphate buffer, pH 7, containing 0.5 mM NAD + and 50 mM formate; glutamate dehydrogenase in 0.1 M potassium phosphate buffer, pH 8, containing 0.1 M ammonium acetate, 20 mM o~ketoglutarate, 1 mM ADP, and 0.15 mM NADPH; lactate dehydrogenase in 0.1 M sodium phosphate buffer, pH 8.5, containing 1 mM pyruvate and 0.15 mM NADH. The activity of immobilized enzymes is likewise determined, and during activity measurements the gel is suspended in the test solution by continuously stirring the mixture with a magnetic cylinder (9 mm diameter and 3 mm height) placed in the cuvette.

Enzyme Immobilization 20fl-Hydroxysteroid dehydrogenase and alcohol dehydrogenase are immobilized separately onto CNBr-activated Sepharose 4B, whereas 12ahydroxysteroid dehydrogenase and glutamate dehydrogenase are coimmobilized onto Sepharose CL-4B activated with tresyl chloride. Sepharose 4B is activated at pH 10.5 using 65 mg CNBr/ml of settled gel, following the method described by Axen et al. lo The coupling of 20/3hydroxysteroid dehydrogenase (34 U, previously dialyzed at 4° against 0.1 M potassium phosphate buffer, pH 8) or alcohol dehydrogenase (640 U) to CNBr-activated Sepharose 4B (10 ml) is carried out in 0.1 M potassium phosphate buffer, pH 8, containing 1 mM NAD ÷, under gentle stirring of the coupling mixture (20 ml) at 4° overnight. The unreacted groups on the matrix are blocked by treating the gel with 0.1 M ethanolamine, pH 8, for 3 hr. Then, the supernatants are withdrawn and the immobilized enzymes thoroughly washed with 0. I M potassium phosphate buffer, pH 7, before testing the enzymatic activity. Sepharose CL-4B is activated with tresyl chloride following the method described by Nilsson and Mosbach. 11 The matrix (10 ml of settled gel) is washed stepwise on a glass-filter funnel with 150 ml of each of the following solutions: distilled water, 30 : 70, 60 : 40, 80 : 20 of acetone:water (v/v), acetone, and finally acetone dried over a molecular sieve. The gel is then added to 3 ml of dry acetone and 0.5 ml of anhydrous pyridine and reacted with 0.25 ml of tresyl chloride under vigorous stirring. After activation for 10 rain the gel is washed with 150 ml of each of the following solutions: acetone, 80 : 20, 60 : 40, 30 : 70, 15 : 85 of acetone : 1 mM HCI (v/v), and finally 1 mM HCI. The coupling of 12ct-hydroxysteroid dehydrogenase (32 U) and glutamate dehydrogenase (115 U) to R. Axen, J. Porath, and S. Ernback, Nature (London) 214, 1302 (1967). i1 K. Nilsson and K. Mosbach, Biochern. Biophys. Res. Commun. 102, 449 (1981).

[14]

ENZYMATIC CATALYSIS IN TWO-PHASE SYSTEMS

153

TABLE I RECOVERY OF ACTIVITY OF IMMOBILIZED ENZYMES

Added activity a

Immobilized activity

Activity in the supernatant

Enzyme

(U)

U

%

U

%

20fl-Hydroxysteroid d e h y d r o g e n a s e

34 640 32 l 15

20 300 11 56

59 47 34 49

Undetectable 5 3 4

0 l l0 3

Alcohol dehydrogenase 12a-Hydroxysteroid d e h y d r o g e n a s e

Glutamate dehydrogenase

a The enzymes were coupled to 10 ml of settled, activated gel.

to the activated matrix (10 ml) is carried out in 0.1 M potassium phosphate buffer, pH 8, under gentle stirring of the coupling mixture (18 ml) at 4 ° overnight. The sample is then processed as described in the case of Sepharose 4B-immobilized enzymes. The recovery of activity of the immobilized enzymes is shown in Table I. Activities of the immobilized enzymes were found to be 34-59% of those of the added soluble enzymes, depending on the enzyme preparations used (Table I); the enzyme activity left in the supernatant was, in all cases, a small fraction of the added one. The Michaelis constants of the immobilized enzymes were generally higher than those of the free enzymes (Table II), except for immobilized glutamate dehydrogenase whose Km.appvalue for tx-ketoglutarate was significantly lower than that of the free enzyme. T A B L E II MICHAELIS CONSTANTS OF FREE AND IMMOBILIZED ENZYMES

Enzyme 20fl-Hydroxysteroid dehydrogenase

Substrate or cofactor NADH

Cortisone Alcohol dehydrogenase

NAD +

12a-Hydroxysteroid dehydrogenase

NADP +

Ethanol Cholic acid methyl ester Glutamate dehydrogenase

NADPH

tx-Ketoglutarate

Free enzyme Km

Immobilized enzyme Km,app

(/zM)

(/~M)

7 42 74 12,500 15 110

26 98 360 16,000 45 130

35 680

230 240

154

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[14]

Steroid Transformations

Reduction of 20-Ketosteroids to 20fl-Hydroxysteroids The stereospecific reduction of progesterone and cortisone is described next. The reduction of progesterone is based on the following reactions: Progesterone (pregn-4-ene-3,20-dione) + N A D H + H + pregn-4-ene-20fl-hydroxy-3-one + N A D ÷ H C O O H + N A D + ~ CO2 + N A D H + H ÷

(1) (2)

where reaction (1) is catalyzed by 20fl-hydroxysteroid dehydrogenase and reaction (2), which serves to regenerate NADH, by formate dehydrogenase. The reaction system is made up as follows. Aqueous phase: 15 ml of 0.1 M potassium phosphate buffer, pH 7, containing 0.2 M formate, 1 mM dithiothreitol, 15 mg of serum albumin, 3 /zmol of NAD ÷, 21 U of 20flhydroxysteroid dehydrogenase, and 24 U of formate dehydrogenase. Organic phase: 15 ml of butyl acetate containing 1.2 mmol (375 mg) of progesterone. The reaction is initiated by shaking the system, contained in a bottle (50 ml) placed on a mechanical oscillating device (oscillating angle approx. 40°, 100 oscillations per minute). The reaction is followed by thin-layer chromatography (of the organic phase) on silica gel GF254 plates (Merck) in the system benzene-methanol (9 : 1, v/v). The complete reduction of progesterone to the 20fl-hydroxyderivative was obtained in 30 hr. The completeness of transformation was also due to the high value (about 10iS) of the overall equilibrium constant for the coupled reactions. The turnover number (moles of product generated per mole of coenzyme added) for the coenzyme was 400. At the end of the run, the residual activity of 20fl-hydroxysteroid dehydrogenase and formate dehydrogenase was 60 and 65%, respectively. The reduction of cortisone is based on the following reactions: Cortisone (pregn-4-ene-17a,21-dihydroxy-3,11,20-trione) + N A D H + H ÷ ~ pregn-4-ene-17a,20fl,21-trihydroxy-3,11-dione + N A D ÷ CH3CH2OH + N A D + ~---CH3CHO + N A D H + H ÷ O CH3CHO + N H 2 - - N H - - C O - - N H 2

II

~ CH3CH=N--NH--C--NH2

(3) (4)

+ H20

(5)

where reaction (3) is catalyzed by 20fl-hydroxysteroid dehydrogenase and reaction (4) by alcohol dehydrogenase. Semicarbazide, reacting with acetaldehyde, shifts the equilibrium to the right and allows for the complete conversion of cortisone.

[14]

ENZYMATIC CATALYSIS IN TWO-PHASE SYSTEMS

155

The reaction system is made up as follows. 7 Aqueous phase: 50 ml of 0.05 M potassium phosphate buffer, pH 7, containing 1.5 ml of ethanol, 0.8 mmol of semicarbazide, 0.04 mmol NAD ÷, 12 U of Sepharose 4Bimmobilized 20/3-hydroxysteroid dehydrogenase (6 ml), and 150 U of Sepharose 4B-immobilized alcohol dehydrogenase (5 ml). Organic phase: 100 ml of ethyl acetate containing 0.72 mmol (260 mg) of cortisone. After 3 hr of shaking (100 oscillations/min) approximately 95% of cortisone was converted into the 20/3-hydroxy derivative. Thin-layer chromatography on silica gel GF254plates in the system water-methanoldiethyl ether-dichloromethane (1.2:8:15:77, v/v) revealed only trace amounts of steroid by-products. The turnover number for the coenzyme was about 20. This low value is due, at least in part, to the increased gm values of the immobilized enzymes for NAD(H) (Table II), and therefore high concentrations of coenzymes in the medium were needed to assure high reaction rates. Also, the NADH regenerating system based on alcohol dehydrogenase is less suitable than that based on formate dehydrogenase since it needs the presence of semicarbazide which also can react with the keto groups present in the steroids and yield by-products. 6 After 8 days of use, the residual activity of 20/3-hydroxysteroid dehydrogenase was approximately 35%. Oxidation of Testosterone to Androstenedione The process is based on the following reactions: Testosterone (androst-4-ene-17/3-hydroxy-3-one) + N A D ÷ androstenedione (androst-4-ene-3,17-dione + N A D H + H + Pyruvate + N A D H + H + ~ lactate + NAD +

(6) (7)

where reaction (6) is catalyzed by fl-hydroxysteroid dehydrogenase and reaction (7) by lactate dehydrogenase. The overall equilibrium constant value for the coupled reactions is 9.4 × 103. 4 The reaction system is made up as follows. Aqueous phase: 10 ml of 0.1 M potassium phosphate buffer, pH 8.5, containing 0.2 M sodium pyruvate, 1 mM dithiothreitol, 15 mg of serum albumin, 3/zmol of NAD +, 20 U of/3-hydroxysteroid dehydrogenase, and 100 U of lactate dehydrogenase. Organic phase: 20 ml of butyl acetate containing 1.04 mmol (300 mg) of testosterone. The reaction is initiated by shaking (100 oscillations/ min) and then followed by thin-layer chromatography on silica gel GF254 plates in the system ethyl acetate-benzene-hexane (100:80:60, v/v). The complete oxidation of testosterone to androstenedione was obtained in about 25 hr. The turnover number for NAD ÷ was 350. At the end of the run, the residual activity of/3-hydroxysteroid dehydrogenase and lactate dehydrogenase was 54 and 70%, respectively.

156

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[14]

Oxidation of Cholic Acid Methyl Ester to 12-Ketochenodeoxycholic Acid Methyl Ester The process is based on the following reactionsJ2: Cholic acid methyl ester (3a,7a, 12t~-trihydroxy-5fl-cholan-24-oic acid methyl ester) + NADP + 12-ketochenodeoxycholic acid methyl ester (3a,7a-dihydroxy-12-oxo-5fl-cholan-24-oic acid methyl ester) + NADPH + H ÷ a-Ketoglutarate ÷ NADPH + NH3 + H + ~- L-glutamate + NADP + + H20

(8) (9)

where reaction (8) is catalyzed by 12a-hydroxysteroid dehydrogenase and reaction (9) by glutamate dehydrogenase. The overall equilibrium constant value for the coupled reactions is about 3.5 x 106.12 The reaction system is made up as follows. Aqueous phase: 10 ml of 0.1 M potassium phosphate buffer, pH 8, containing 0.2 M ammonium acetate, 0.1 M a-ketoglutarate, 0.75 mM ADP, 1 mM dithiothreitol, 1 /zmol of NADP ÷, and 5.5 U of 12a-hydroxysteroid dehydrogenase coimmobilized onto Sepharose CL-4B (5 ml) with 28 U of glutamate dehydrogenase. Organic phase: l0 ml of butyl acetate containing 0.47 mmol (200 mg) of cholic acid methyl ester. The reaction is initiated by shaking (100 oscillations/min), and then followed by thin-layer chromatography on silica gel plates in the system chloroform-methanol-acetic acid (20:2: 1, v/v); the plates are sprayed with Komarowsky's reagent. 13 An almost complete oxidation of cholic acid methyl ester to 12-ketochenodeoxycholic acid methyl ester was achieved in 3 days. The turnover number for NADP ÷ was 470. After 3 months of continuous use, the residual activity of 12a-hydroxysteroid dehydrogenase and glutamate dehydrogenase was 48 and 60%, respectively. General Considerations The described steroid transformations exemplify the feasibility of carrying out enzymatic reactions in two-phase systems. Many other steroids can be converted, under similar conditions, using the examined enzymes2'4-6 (20fl-hydroxysteroid dehydrogenase, 3 (or 17)/3-hydroxysteroid dehydrogenase, and 12a-hydroxysteroid dehydrogenase), or other enzymes such as 3a-hydroxysteroid dehydrogenase, laccase, 3 peroxidase, and steroid isomerase. 14In all cases, the use of two-phase systems drastically reduced reaction volumes and allowed an easy separation of ~2G. Carrea, R. Bovara, P. Cremonesi, and R. Lodi, Biotechnol. Bioeng. 26, 560 (1984). 13 I. A. Macdonald, J. Chromatogr. 136, 348 (1977). i4 p. Cremonesi, G. Mazzola, and L. Cremonesi, Ann. Chim. 67, 415 (1977).

[15]

DEBLOCKING IN PEPTIDE SYNTHESIS WITH C P Y

157

the products, which are predominantly present in the organic phase, from the enzymes and cofactors, which are present in the aqueous phase. The parameters that influence the effectiveness of enzymatic reactions in two-phase systems may be summarized as follows: effect of organic solvents on the stability and activity of enzymes, solubility and partition of reagents and products between the two phases, ratio between the volumes of the organic and the aqueous phases, rate of transfer of reagents and products between the phases. A detailed examination of such parameters was made in previous articles, s,15 Regarding the choice of the organic solvents, the ones most suitable for our purposes were butyl acetate and ethyl acetate, 2-s,lz since they have high solubilizing capacity for steroids and weak inhibiting and denaturing effects on enzymes. The regeneration of the costly coenzymes markedly reduces the cost of the process and should make it possible to fully exploit the great specificity of hydroxysteroid dehydrogenases for preparative-scale transformations of steroids in two-phase systems. Also, the use of coupled enzymatic reactions highly improves transformation yields when overall equilibrium constant values are favorable. The immobilization of the enzymes made it possible to use them repeatedly and increased their stability, particularly in the case of tresyl chloride activation of the matrix. It should also be emphasized that, contrary to what happened with separately immobilized enzymes, coimmobilized enzymes regenerated the coenzymes very effectively. is G. Carrea, Trends Biotechnol. 2, 102 (1984).

[15] D e b l o c k i n g in P e p t i d e S y n t h e s i s with Immobilized Carboxypeptidase Y

By G. P. ROYER Introduction The selectivity of enzymes and the mild conditions under which they operate suggest their use in the synthesis of complex, unstable biomolecules and drugs. One application is in deblocking reactions. In sequential synthesis many reactions are run one after the other. Accumulation of impurities generated by harsh deblocking reagents results in a complex mixture of closely related products. Sequential synthesis of biopolymers METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[15]

DEBLOCKING IN PEPTIDE SYNTHESIS WITH C P Y

157

the products, which are predominantly present in the organic phase, from the enzymes and cofactors, which are present in the aqueous phase. The parameters that influence the effectiveness of enzymatic reactions in two-phase systems may be summarized as follows: effect of organic solvents on the stability and activity of enzymes, solubility and partition of reagents and products between the two phases, ratio between the volumes of the organic and the aqueous phases, rate of transfer of reagents and products between the phases. A detailed examination of such parameters was made in previous articles, s,15 Regarding the choice of the organic solvents, the ones most suitable for our purposes were butyl acetate and ethyl acetate, 2-s,lz since they have high solubilizing capacity for steroids and weak inhibiting and denaturing effects on enzymes. The regeneration of the costly coenzymes markedly reduces the cost of the process and should make it possible to fully exploit the great specificity of hydroxysteroid dehydrogenases for preparative-scale transformations of steroids in two-phase systems. Also, the use of coupled enzymatic reactions highly improves transformation yields when overall equilibrium constant values are favorable. The immobilization of the enzymes made it possible to use them repeatedly and increased their stability, particularly in the case of tresyl chloride activation of the matrix. It should also be emphasized that, contrary to what happened with separately immobilized enzymes, coimmobilized enzymes regenerated the coenzymes very effectively. is G. Carrea, Trends Biotechnol. 2, 102 (1984).

[15] D e b l o c k i n g in P e p t i d e S y n t h e s i s with Immobilized Carboxypeptidase Y

By G. P. ROYER Introduction The selectivity of enzymes and the mild conditions under which they operate suggest their use in the synthesis of complex, unstable biomolecules and drugs. One application is in deblocking reactions. In sequential synthesis many reactions are run one after the other. Accumulation of impurities generated by harsh deblocking reagents results in a complex mixture of closely related products. Sequential synthesis of biopolymers METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

158

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS HOAv'O("~O)N~OH

[15]

PEG

Carboxymethylation

Attachment of glycylmethionine

PEG'-Met-OH

FIG. 1. Preparation of the PEG handle.

(peptides and nucleic acids) appears to be an area where enzyme catalysis would prove useful. Carboxypeptidase Y (CPY; EC 3.4.16.1, serine carboxypeptidase) is an exopeptidase found in yeast. The enzyme is unusual in that it has a broad specificity. It catalyzes the release of L-amino acids, including proline, from the C terminus of a polypeptide chain. Rates of amino acid release vary but not to the extent exhibited by other exopeptidases such as pancreatic enzymes. CPY also acts as an esterase. It rapidly cleaves the C-terminal alkyl ester group of large and small peptides. The pH optima of these two activities are widely separated. At pH 6 the enzyme acts as a peptidase, while at pH 8.5 the enzyme acts as an esterase with little or no peptidase activity. This behavior suggested to us that the enzyme would be of use in the synthesis of peptides. Our first thought was that one could cleave the C-terminal ester and leave the side-chain ester groups of aspartic acid and glutamic acid intact. Thus, the condensation of fragments would be facilitated by this procedure of selective saponification. Also, we demonstrated a method to resolve racemic mixtures of blocked peptide and amino acid esters; the enzyme will not hydrolyze the ester group of a terminal D-amino acid.l Regiospecific cleavage of the aester of blocked diester derivatives of glutamic acid and aspartic acid can be accomplished with the immobilized enzyme. Our main interest in CPY has involved the concept of a method for the synthesis of peptides in aqueous systems using very mild conditions for deblocking. 2 Small protected peptides can be deblocked in alcohol/water mixtures. However, as the size of the peptide increases the solubility decreases. To use the immobilized CPY in the presence of high levels of organic solvent is not possible. This problem was solved by growing the peptide chain on a water-soluble "handle," namely polyethylene glycol (PEG). The derivatization of the PEG is shown in Fig. 1. The terminal i G. P. Royer, H. Y. Hsiao, and G. M. Anantharamaiah, Enzyme Eng. 5, 461 (1980). 2 G. P. Royer and G. M. Anantharamaiah, J. A m Chem. Soc. 101, 3394 (1979).

[15]

D E B L O C K I NIN G PEPTIDE SYNTHESIS WITH CPY

159

PEG'-Met-OH ~ H-AA1-OR PEG'-Met-AA1-OR Deblockwith I-CPY PEG'-Met-AA1-OH I Extension PEG'-Met [AAI....AAN] -OH I CNBr Freepeptide FIG. 2. Chain elongation and release.

hydroxyl groups of PEG were carboxymethylated; glycylmethionine was then added. The methionine was incorporated in order to provide a cleavable (CNBr) linkage between the peptide and the PEG carrier. Chain elongation and release are shown schematically in Fig. 2. The first amino acid of the target sequence is added as the ester, using watersoluble carbodiimide. Following removal of excess reagents, the ester protecting group is hydrolyzed by treatment with immobilized CPY. Coupling/deblocking cycles are repeated until the sequence is complete. The finished peptide is then removed by cleavage with cyanogen bromide. In the event that methionine is part of the target sequence, it can be added as methionine sulfoxide which prevents reaction with cyanogen bromide. Reduction of the released peptide yields the desired sequence. Procedures Enzyme. Carboxypeptidase Y can be prepared from yeast according to the procedure of Kuhn et al.3 Final purification can be accomplished using the affinity column of Johansen et al. 4 The enzyme is commercially available from Pierce Chemical Co. and Sigma Chemical Co. An immobilized form of the enzyme is also available from Pierce. Assay. CPY-catalyzed ester hydrolysis can be followed titrimetrically or spectrophotometrically with N-a-acetyl-L-tyrosine ethyl ester (ATEE). 5 The titrimetric assay is convenient for the immobilized enzyme. An automatic titration system (pH-stat) equipped with a strip chart recorder is needed. When reactions are run at alkaline pH, a stream of 3 R. W. Kuhn, K. A. Walsh, and H. Neurath, Biochemistry 13, 3871 (1974). 4 j. I. Johansen, K. Breddam, and M. Ottesen, Carlsburg Res. Commun. 41, 1 (1976). 5 R. Hayashi, this series, Vol. 45, p. 568.

160

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[15]

nitrogen is directed at the surface of the liquid in the reaction vessel. The buffer is phosphate (5 mM, pH 8) which contains 0.1 M KC1. An ATEE stock solution (0.2 M) can be prepared with ethanol as a solvent. The titrant is 10 mM NaOH, standardized and protected from the atmosphere with an ascarite (NaOH) trap. The temperature of the reaction is maintained at 25 ° with a jacketed reaction vessel which is connected to a circulating water bath. The final substrate concentration is 10 mM. Enzyme concentrations are approximately 1 /xg/ml for the soluble enzyme and 1 mg/ml for immobilized enzyme. Enzyme Immobilization. There are several methods in the literature for immobilization of CPY. 6'7 The enzyme can be attached directly to aminohexylamino-agarose using water-soluble carbodiimide. It can be adsorbed to immobilized concanavalin A. The carbohydrate can be oxidized and reacted with diamines to give a species which can be immobilized on supports containing active acylating groups or aldehydes, The preferred procedure (the latter) appears in this series. 8 CP Y-Catalyzing Deblocking of Peptide Esters. Conditions and results for the deblocking of peptide esters are shown in Table I. The deblocking reaction is carried out at pH 8.5 with the aid of a pH-stat. The reactant is dissolved in methanol or DMF and kept as a stock solution (about 100 mg/ ml). The concentration of the organic solvent in the final reaction mixture should not exceed 25%. The reaction is started by addition of the immobilized enzyme (20-80 ATEE units, standard). When base uptake ceases, the enzyme is filtered, the reaction mixture is acidified (pH 2-3), and the solution is concentrated with a rotary evaporator at a temperature of 40 ° until precipitation occurs. The crystallization of the peptide derivatives is accomplished with ethanol/water. Water is added to a 20% ethanolic solution of the peptide until turbidity appears. The solution is cooled to complete crystallization. If available, seed crystals should be added before cooling. Preparation of CM-PEG. PEG (14 g, MW 6000-7000) and potassium tert-butoxide (10 g) are dissolved in tert-butanol (150 ml) by warming. To this solution ethyl bromoacetate (5 ml) is added dropwise over a period of 10 rain. After 2 hr of stirring, the solvent is removed by rotary evaporation. The residue is taken up in 100 ml of 1 N NaOH. After 2 hr at room temperature, the pH of the mixture is adjusted to 2 with HCI. The C M PEG is recovered by partitioning into chloroform (two 200-ml portions). The chloroform layer is washed with water and then dried with anhydrous 6 F. A. Liberatori, J. E. Mclsaac, Jr., and G. P. Royer, FEBS Lett. 68, 45 (1976). 7 H. Y. Hsiao and G. P, Royer, Arch. Biochern. Biophys. 198, 379 (1979). 8 G. P. Royer, this series, Vol. 135 [10].

[15]

DEBLOCKING IN PEPTIDE SYNTHESIS WITH C P Y

161

TABLE I DEBLOCKING OF PEPTIDE ESTERS

Peptide ester Z-Asn-Phe-OEt Boc-Asp(O-Bzl)-Phe-OEt Z-Asp(OMe)-OMe Z-Phe-D-Ala-OEt 1 : 1 mixture of Z-Phe-D-Ala-OEt and Z-Phe-L-AIa-OEt

Yield (%)

Concentration/solvent

Product

20 mg in 160 ml of 25% MeOH 10 mg in 15 ml of 20% MeOH 20 mg in 20 ml of 25% DMF 10 mg in 50 ml of 25% DMF 10 mg of each in I00 ml of 25% DMF

Z-Asn-Phe-OH

95

Boc-Asp(O-Bzl)-Phe-OH

84

Z-Asp(OMe)-OH

98

No reaction

--

Z-Phe-L-AIa-OH

95

NazSO4. After rotary evaporation of the solvent, 12 g of derivatized P E G remains.

CP Y-Catalyzed Deblocking of Peptides Attached to the PEG Handle. The deblocking o f P E G / M e t - O E t is accomplished by treatment with immobilized CPY at p H 8.5 at room temperature. The pH-stat is used to maintain the pH and to follow the course of the reaction. P E G ' - M e t - O E t (3 g, P E G ' is P E G - g l y c i n e ) is dissolved in 40 ml of water. Immobilized e n z y m e (30 A T E E units, standard) is added and the pH is quickly adjusted to 8.5. The reaction mixture is stirred with an overhead stirrer at a speed sufficient to maintain the enzyme beads in suspension. The titrant used is 0.1 M N a O H . The time required for hydrolysis of P E G ' - M e t - O E t under these conditions is 5 hr. The hydrolysis rate goes up dramatically as the chain is extended, probably because of subsite binding of neighboring amino acids. When the reaction is complete (cessation of base uptake), the reaction mixture is filtered to remove the immobilized enzyme. The P E G - p e p t i d e solution can be used directly for coupling of the next amino acid. It is essential to remove all of the e n z y m e from the solution. If e n z y m e remains behind and the p H is adjusted to 6, peptidase activity will result in the degradation of the chain. Comments In related work on the use of proteases in peptide synthesis, John Glass and co-workers have shown that trypsin can be used to deprotect the a-amino group. 9 The coupling segments were Cbz-arginyl peptides. 9 C. Myers and J. D. Glass, Proc. Natl. Acad. Sci. U.S.A. 72, 2193 (1975).

162

IMMOBILIZED

ENZYMES/CELLS

IN ORGANIC

SYNTHESIS

[16]

The Cbz-arginyl group is cleaved by trypsin. To avoid endopeptidase activity, temporary blocking groups for arginyl and lysyl side chains in the target sequence were employed. In another clever application of proteases, Glass and Pelzig have demonstrated enzyme-cleavable side-chain protection. Arginine methyl ester was attached to side-chain carboxyl groups using water-soluble carbodiimide. Trypsin was used to remove the methyl ester of the arginine moiety and then carboxypeptidase B was used to remove the arginyl group.l° In conclusion, immobilized CPY is a versatile tool for the protein chemistry lab. It can be used for deblocking in peptide synthesis,2 resolution of racemic mixtures of blocked amino acid esters or peptides, 1 sequencing of peptides, 6 total hydrolysis of peptides for analysis, 11 and condensation reactions in peptide synthesis. ~2 The attachment of C-terminal amino acid amides can be accomplished using CPY to catalyze the peptide bond formation starting with the peptide ester and the amide. Also, the enzyme should be of interest in peptide modification (semisynthesis) and fragment condensation of peptides. 10 j. D. Glass and M. Pelzig, Proc. Natl. Acad. Sci. U.S.A. 74, 2739 (1977). ~1 G. P. Royer, W. E. Schwartz, and F. A. Liberatore, this series, Vol. 47, p. 40. 12 K. Breddam, F. Widmer, and J. T. Johansen, Carlsburg Res. Commun. 45, 361 (1980).

[16] U s e o f I m m o b i l i z e d Achrornobacter P r o t e a s e I for Semisynthesis of Human Insulin

By

K A Z U Y U K I MORIHARA, RYONOSUKE M U N E Y U K I ,

and TATSUSHI OKA Porcine insulin can be enzymatically converted into human insulin by two different methods. In condensation reactions catalyzed by trypsin (EC 3.4.21.4) I and protease I from Achromobacter lyticus (EC 3.4.21.50), 2 L-threonine tert-butyl ester can be coupled to 30B-dealanine porcine insulin (DAI) to obtain an ester derivative of human insulin. In the K. Morihara, T. Oka, and H. Tsuzuki, Nature (London) 280, 412 (1979). 2 K. Morihara, T. Oka, H. Tsuzuki, Y. Tochino, and T. Kanaya, Biochem. Biophys. Res. Commun. 92, 396 (1980).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

162

IMMOBILIZED

ENZYMES/CELLS

IN ORGANIC

SYNTHESIS

[16]

The Cbz-arginyl group is cleaved by trypsin. To avoid endopeptidase activity, temporary blocking groups for arginyl and lysyl side chains in the target sequence were employed. In another clever application of proteases, Glass and Pelzig have demonstrated enzyme-cleavable side-chain protection. Arginine methyl ester was attached to side-chain carboxyl groups using water-soluble carbodiimide. Trypsin was used to remove the methyl ester of the arginine moiety and then carboxypeptidase B was used to remove the arginyl group.l° In conclusion, immobilized CPY is a versatile tool for the protein chemistry lab. It can be used for deblocking in peptide synthesis,2 resolution of racemic mixtures of blocked amino acid esters or peptides, 1 sequencing of peptides, 6 total hydrolysis of peptides for analysis, 11 and condensation reactions in peptide synthesis. ~2 The attachment of C-terminal amino acid amides can be accomplished using CPY to catalyze the peptide bond formation starting with the peptide ester and the amide. Also, the enzyme should be of interest in peptide modification (semisynthesis) and fragment condensation of peptides. 10 j. D. Glass and M. Pelzig, Proc. Natl. Acad. Sci. U.S.A. 74, 2739 (1977). ~1 G. P. Royer, W. E. Schwartz, and F. A. Liberatore, this series, Vol. 47, p. 40. 12 K. Breddam, F. Widmer, and J. T. Johansen, Carlsburg Res. Commun. 45, 361 (1980).

[16] U s e o f I m m o b i l i z e d Achrornobacter P r o t e a s e I for Semisynthesis of Human Insulin

By

K A Z U Y U K I MORIHARA, RYONOSUKE M U N E Y U K I ,

and TATSUSHI OKA Porcine insulin can be enzymatically converted into human insulin by two different methods. In condensation reactions catalyzed by trypsin (EC 3.4.21.4) I and protease I from Achromobacter lyticus (EC 3.4.21.50), 2 L-threonine tert-butyl ester can be coupled to 30B-dealanine porcine insulin (DAI) to obtain an ester derivative of human insulin. In the K. Morihara, T. Oka, and H. Tsuzuki, Nature (London) 280, 412 (1979). 2 K. Morihara, T. Oka, H. Tsuzuki, Y. Tochino, and T. Kanaya, Biochem. Biophys. Res. Commun. 92, 396 (1980).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[16]

SEMISYNTHESIS OF HUMAN INSULIN

163

other method, the terminal alanine residue of the B chain in porcine insulin is exchanged for amides and esters of L-threonine in a single transpeptidation step catalyzed by either trypsin, 3 carboxypeptidase Y (EC 3.4.16.1, serine carboxypeptidase) 4 or Achromobacter protease 1.5 The protected insulin derivative obtained from the enzyme-catalyzed reactions is subsequently deblocked in a one-step chemical procedure which yields human insulin. For large-scale production, Achromobacter protease I seems to be the ideal enzyme to use in enzyme-assisted semisynthesis of human insulin. The enzyme exhibits several characteristics that make it suitable to use for synthetic purposes. For example, the rate of condensation of DAI to L-threonine tert-butyl ester catalyzed by Achromobacter protease I is at least 10 times higher than that observed for the trypsin-catalyzed condensation reaction. Also, the enzyme specifically cleaves peptide bonds containing an L-lysine residue at the carboxyl side of the splitting point in protein and peptide substrates, 6 preventing the formation of by-products. Furthermore, the amount of enzyme required for the enzymatic coupling reaction is 50-100 times less than that required in transpeptidation reactions catalyzed by other proteases, and the same product yields are afforded by the two approaches) It should finally be added that Achromobacter protease I may be useful for direct preparation of DAI which cannot be prepared in a single step using other proteolytic enzymes. For efficient use of enzymes in industry, continuous operation of the enzyme-catalyzed process is desirable. This can be achieved by using immobilized enzymes, which are known to be stable, reusable, and also easily recoverable. This article explains the immobilization of Achromobacter protease I as well as the use of the immobilized enzyme in the enzyme-assisted semisynthesis of human insulin, which is schematically depicted in Fig. 1. Assay Procedure of the Immobilized Enzyme

Principle. Ester and amide derivatives of N~-acylated L-lysine are used as substrates for enzyme activity determinations of both soluble 6 and immobilized enzymes. 3 A. Jonczyk and H.-G. Gattner, Hoppe-Seyler's Z. Physiol. Chem. 362, 1591 (1981); J. Markussen, United Kingdom Patent 2,069,501A (August 26, 1981). 4 K. Breddam, F. Widmer, and J. T. Johansen, Carlsburg Res. Commun. 46, 361 (1981). 5 K. Morihara and T. Oka, "Peptide Chemistry, 1982," p. 231, Protein Research Foundation, Osaka, Japan, 1983. 6 T. Masaki, T. Fujihashi, K. Nakamura, and M. Soejima, Biochim. Biophys. Acta 660, 51 (1981).

164

IMMOBILIZED ENZYMES/CELLSIN ORGANICSYNTHESIS

[16]

Gly

~ - ~ / Phe

Asn22

29 30

Arg

Lys-Ala

Porcine insulin

~ Immobilized enzyme Gly 29

Phe - Lys Des-alanyl(B30)-insulin (p-DAI) t ~--.- Thr-OBu t Immobilized enzyme Gly ~ - - - ~ /

Asn

Phe Human insulin tertiary butyl

Lys-Thr-O But ester

1 VT---1

.~-- CF3COOH

Gly

/

/

Phe / / Human insulin

29

30

Lys-Thr

FIG. 1. Scheme for enzyme-assistedsemisynthesisof human insulin using immobilized protease I.

Achromobacter

Reagents and Apparatus N~-Tosyl-L-lysine methyl ester (TLME), 20 mM in 2 M KCI Sodium hydroxide, 0.01 or 0.025 N Immobilized Achromobacter protease I (lysine endopeptidase) Automatic titration apparatus Procedure. To a small glass tube (12 × 60 mm) containing immobilized enzyme (10 mg wet gel), 1 ml of water is added. The mixture is vigorously stirred and 1 ml of TLME is added to the homogeneous suspension. The pH of the reaction mixture is maintained at 7.0 by automatic titration with 0.01 or 0.025 N NaOH at 20 - 2°. The activity of the immobilized enzyme is expressed as micromoles of substrate hydrolyzed by the enzyme per minute and gram (wet weight) of immobilized enzyme added to the assay mixture at 20 ± 2°.

[16]

SEMISYNTHESIS OF HUMAN INSULIN

165

Enzyme Immobilization Preliminary studies7 indicate that at least 6/zM enzyme is required to obtain practical product yields in the enzyme-catalyzed coupling reactions using immobilized Achromobacter protease I. Therefore, different methods of enzyme immobilization were investigated to obtain preparations containing immobilized enzymes showing sufficiently high specific activities. The best results were obtained after coupling the enzyme to poly(t-glutamic acid) (poly-Glu-OH) or silica gel containing covalently bound poly(L-glutamic acid). The enzyme immobilized on other commonly used supports such as CNBr-activated Sepharose 4B (Pharmacia Fine Chemicals), Affi-Gel 15 (Bio-Rad), and agar (Difco Bacto) showed low specific activities.

Covalent Coupling to Poly(L-Glutamic Acid) Reagents Sodium poly(L-glutamate) (MW 50,000), Protein Research Foundation, Minoh, Osaka Amberlite IR-120 1- Ethyl- 3 - (3- dimethylaminopropyl)carbodiimide hydrochloride (WSCD), Protein Research Foundation, Minoh, Osaka Achromobacter protease 18 (487/zmol TLME/min-mg at pH 7.0, 761 /zmol TLME/min-mg at pH 8.0, 20 -+ 2°), Wako Pure Chemical Industries, Osaka) Procedure. Amberlite IR-120 (4 ml of settled beads, acid form) is added to a solution containing sodium poly(e-glutamate) (411 mg) in water (10 ml). After several minutes of reaction, the precipitate which forms is removed from the beads by filtering the slurry through a thin cotton layer and by washing the beads retained on the cotton layer with several small volumes of water. To the poly-Glu-OH (acid form) suspended in the combined washing solutions (30 ml), Achromobacter protease I (11.3 mg) in water (2 ml) and WSCD (340 mg) in water (4 ml) are added successively, and the mixture is stirred gently for 35 rain at room temperature. After centrifugation (1000 g × 10 rain) of the reaction mixture, the supernatant, which shows no enzyme activity, is removed and the precipitate is suspended in 0.1 M acetate buffer (30 ml), pH 4.0. The mixture is centrifuged 7 R. Muneyuki, T. Oka, and K. Morihara, "Peptide Chemistry, 1981," p. 113, Protein Research Foundation, Osaka, Japan, 1982. 8 T. Masaki, M. Tanabe, K. Nakamura, and M. Soejima, Biochim. Biophys. Acta 660, 44 (1981).

166

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[16]

(I000 g x 10 min), and after additional washing with 1 M NaCI (30 ml), the precipitate is transferred to a filter paper. The filter paper containing the precipitate is squeezed to remove excess water, and after drying 0.317 g of immobilized enzyme is obtained. This dry weight of the immobilized enzyme corresponded to 1.63 g wet weight of the same preparation. The activity of the immobilized enzyme was 907 p.M/min-g wet gel when the preparation was assayed at pH 7.0. This activity corresponded to 27% of the original activity of the soluble enzyme used in the immobilization step. Immobilized Achromobacter protease I showed considerable solubility at pH 6.5 (the pH of the enzymatic condensation reaction). Moreover, the solubility caused some difficulties because the human insulin ester derivative that was produced was difficult to separate from the immobilized enzyme; for example, it could not be separated by filtration on glass filter. The enzyme was therefore immobilized on an insoluble silica gel premodifled by substitution with poly(L-glutamic acid).

Covalent Coupling to Silica Gel Containing Immobilized Poly(L-Glutamic Acid) Reagents Sodium poly(L-glutamate) (MW 50,000) Lichrosorb-NH2 (3-aminopropyl silica gel; 10/zm mean particle size) for liquid chromatography (Merck) Achromobacter protease I WSCD Procedure. Step I. An aqueous solution (5 ml) containing sodium poly-L-glutamate (1.0 g) is added to 1 N HC1 (30 ml). The solution is left standing for 2 hr in a thermostated water bath kept at 0°. After centrifugation of the reaction mixture (450 g x 15 min), the precipitate is suspended in 0.1 N HC1 (30 ml). The mixture is left standing for another hour at room temperature and was then centrifuged (450 g x 15 min). The obtained precipitate is dried under vacuum at around 60°, yielding a white solid for poly-Glu-OH (0.778 g). Step 2. Lichrosorb-NH2 (5.0 g) is suspended in water (20 ml) and the resulting slurry is poured into a glass cylinder filled with water. The height of the liquid in the cylinder is 30 cm after addition of the silica gel. The portion precipitation after 20 min is collected, and the gel is washed successively on a sintered glass filter with 0.002 N HC1 (800 ml), water (300 ml), 0.003 N NaOH (500 ml), water (300 ml), ethanol-water (1/1, v/v) (100 ml), and ethanol (50 ml). After drying overnight at room temperature

[16]

SEMISYNTHESIS OF HUMAN INSULIN

167

the obtained silica gel (3.33 g) contains 1.21% nitrogen (determined after elemental analysis). Step 3. Poly-Glu-OH (381 mg of dry weight) and the silica gel (680 mg of dry weight), both obtained as described above, are added to N,N-dimethylformamide (10 ml). The slurry is added to N,N-dimethylformamide (5 ml) containing WSCD (331 mg). After stirring at 60° for 4 hr, the reaction mixture is poured into 0.1 N HCI (40 ml) and the slurry centrifuged (1000 g × 10 min). The precipitate obtained is washed successively by repeated centrifugation with 0.05 N HC1 (60 ml), 0.005 N NaOH (60 ml), 0.2 M NaHCO3 (60 ml), and water (60 ml). The solid material is then transferred to a glass filter and washed with 0.002 N HC1 (200 ml) and water (200 ml). The modified silica gel is dried under vacuum at 50°, giving 783 mg of dry silica-bound poly-Glu-OH. The nitrogen concentration of the modified silica gel is 3.31% (determined after elemental analysis). Step 4. Silica-bound poly-Glu-OH (87.5 mg) is added to a water solution (0.5 ml) containing WSCD (16.2 mg), and the mixture is stirred for 2.5 min at room temperature. To this suspension, Achromobacter protease I (0.65 mg) in water (0.5 ml) is added, followed by addition of 0.1 M sodium phosphate buffer (pH 1!.8, 0.5 ml). After continuous stirring for 2.5 hr at room temperature, the reaction mixture is filtered on a Bfichner funnel and the precipitate containing immobilized enzyme is washed successively with 1 M acetate buffer (pH 7.6, 50 ml), water (50 ml), 1 M acetate buffer (pH 5.0, 50 ml), and water (50 ml). The immobilized enzyme is transferred to an Erlenmeyer flask containing 2 M NaC1 (50 ml). After stirring the slurry for 30 min, the immobilized enzyme is washed with water on a BiJchner funnel until no enzyme activity can be detected in the washing solution. The wet gel (176 mg) showed an enzyme activity of 615/zM/min-g wet gel when assayed at pH 8.0. It is worth noting that the activity of the silica-bound enzyme varied with the pH of the sodium phosphate buffer used in the coupling reaction. Thus, preparations assayed at pH 8.0 showed activities of 124,263,346,510, and 615/zM/min-g of wet gel when the pH of the added phosphate buffer was 7.5, 8.8, 9.8, 10.8, and 11.8, respectively. When 0.2 M sodium phosphate buffer (pH 11.8) was added in the immobilization step, the activity of the immobilized enzyme dropped to 134/.tM/min-g wet gel at pH 8.0. Preparation of DAI Using Immobilized Enzyme

Principle. Achromobacter protease I can be used to hydrolyze porcine insulin by specifically releasing the C-terminal alanine residue of the B chain in insulin. 6

168

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[16]

Reagents and Apparatus Porcine insulin (lot 1FJ91, 26.2 U/mg), Eli Lilly Silica-bound Achromobacter protease I, 615/zM/min-g wet gel (prepared as described above) Ammonium carbonate, 0.1 M, pH 8.3 Amino acid analyzer, Hitachi KLA Model 5 Procedure. Silica-bound enzyme is packed in a small glass column (6.5 × 12 mm, containing 310 mg of wet gel) and was equilibrated with 0.1 M ammonium carbonate, pH 8.3 (equilibration buffer). Porcine insulin (167 mg, 1 mM) dissolved in 29 ml of the equilibration buffer is applied to the column and the solution is pumped through the column at a flow rate of 0.5 ml/hr at room temperature. Since the hydrolysis is not completed after one passage through the column, the substrate solution must be recirculated. Three, six, and nine passages lead to 34, 69 and 100% hydrolysis of the porcine insulin. The rate of hydrolysis is determined from the content of analine in the elution buffer (for alanine determinations, amino acid analyses of the eluates are carried out). After 10 passages, the column is washed with equilibration buffer until the buffer contains no DAI. The combined eluates are lyophilized and the formed DAI is ultimately purified by gel chromatography on Sephadex G-50 (4.2 × 134 cm, 1.85 liter gel) using 0.5 M acetic acid as the eluant. 2 The yield of DAI was 153 mg (92%). Preparation of Human Insulin Using Immobilized Enzyme

Principle. The enzymatic condensation of DAI to Thr-OBu t has been catalyzed successfully by soluble Achromobacter protease I in organic aqueous cosolvents under controlled pH. 2 Similar reaction conditions are employed for the enzymatic condensation using immobilized enzyme. Reagents and Apparatus DAI is prepared as described above. DAI is also prepared after digesting porcine insulin in the presence of soluble carboxypeptidase A 9 (EC 3.4.17.1). L-Threonine-tert-butyl ester (Thr-OBut), bp 85-88°/3 mm Hg, is prepared as described elsewhere, i° Silica-bound Achromobacter protease I, 615 ~M/min-g wet gel (prepared as described above) 9 E. W. Schmitt and H.-G. Gattner, Hoppe-Seyler's Z. Physiol. Chem. 359, 799 (1978). 10 E. Vowinkel, Chem. Ber. 11~, 16 (1967).

[16]

SEMISYNTHESIS OF HUMAN INSULIN

169

Acetate buffer, 5 M, pH 5.7

1,4-Butanediol Apparatus for high-performance liquid chromatography: Waters Assoc. Model 6000A pump, Rheodyne Model 7120 injector, and Japan Spectrooptics UVIDEC-100 variable-wavelength UV detector Assay. The enzyme-catalyzed coupling of DAI to Thr-OBu t is followed by spectrophotometric measurement of the decrease in DAI absorbance at 220 nm. The analyses are performed using reversed-phase liquid chromatography ~ (column, Nucleosil 5C~8, 0.4 x 20 cm; eluant, 30.5% CH3CN in 5 mM phosphate buffer, pH 3.0, containing 5 mM sodium 1-butane sulfonate and 50 mM Na2SO4). The product yield of the enzymatic reaction is calculated from the ratio of the determined peak areas of [30B-Thr-OBut]insulin (retention time, 14.2 min) and the remaining DAI (retention time 5.3 min). Procedure. DAI (162 mg), Thr-OBu t (400 mg), and water are added to a mixture containing 1,4-butanediol (1.12 ml) and acetate buffer (5 M, pH 5.7, 1.07 ml). The final volume of the reaction mixture is 7.81 ml and the pH of the mixture is 6.5. The solution is applied to a small column (6.5 × 12 mm) packed with silica-bound Achromobacter protease I (310 mg of wet gel). The column has been preequilibrated with the reaction mixture containing no DAI and Thr-OBu t. The flow rate is adjusted to 85/zl/hr by applying nitrogen pressure (0.2 atm) to the top of the column. The reaction mixture is recirculated through the column at room temperature to complete the enzymatic condensation reaction. In the first eluate, 45% of DAI is converted into [30B-Thr-OBut]insulin. The yield increases to 63 and 80% after the second and third runs, respectively. The combined eluates and the solution obtained after washing the column with 2 M acetate buffer (pH 6.5) are pooled and applied to a column packed with Sephadex G-50 (4.2 x 134 cm, 1.85 liter gel). Column chromatography is conducted using 0.5 M acetic acid as the eluant as described, z The insulin fraction obtained after the gel chromatography is lyophilized and the residue is then applied to a column packed with DEAE-Sephadex A-25 (4 x 34 cm, 427 ml gel) which has been preequilibrated with 0.01 M Tris buffer, pH 7.4, containing 7 M urea (equilibration buffer). On application of a linear NaC1 gradient (0-0.3 M, total volume of 1.5 liter) in the presence of equilibration buffer, [30B-Thr-OBut]insulin is eluted first and then DAI. 2 After removal of the urea by gel filtration, fractions containing [30B-Thr-OBut]insulin are lyophilized. The protected insulin derivative is deprotected by treatment with trifluoroacetic acid (TFA) in the presence of anisole. 2 The yield of human insulin is 59% (95 mg).

170

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[17]

The experiments discussed in this article have been helpful in solving some of the principal problems in the semisynthesis of human insulin from porcine insulin. Although they were done on a semipreparative scale, the obtained results indicate that application of immobilized Achromobacter protease I in the semisynthesis of human insulin has great potential as a tool for the large-scale production of human insulin.

[ 17] Application of I m m o b i l i z e d A m i n o p e p t i d a s e s to t h e S e q u e n t i a l H y d r o l y s i s of P r o l i n e - C o n t a i n i n g P e p t i d e s

By G. FLEMINGERand A. YARON Sequence-ordered hydrolysis of peptide chains by exopeptidases followed by quantitative amino acid analysis is a mild method for determination of the amino acid sequence, which is sensitive to the amino acid configuration and structure. ~ The most convenient procedure uses enzymes in the immobilized form, 2,3 permitting removal of enzyme-free samples for analysis as well as the use in sequence of several enzymes in the same reaction mixture. Peptide bonds associated with proline residues--present in a polypeptide chain--are unique in their susceptibility to hydrolysis by aminopeptidases. 4 Clostridial aminopeptidase (CAP; EC 3.4.11.13) 5-7 isolated from Clostridium histolyticurn is able to cleave N-terminal proline residues from polypeptide chains at rates comparable to cleavage rates of other N-terminal amino acid residues. The enzyme does not cleave Nterminal residues if followed by proline. The Xaa-Pro bond is resistant also to other aminopeptidases, except to aminopeptidase P (AP-P; EC 3.4.11.9) which cleaves this bond exclusively,s-12In the present article we describe application of glass-bound CAP and AP-P to the sequential hydrolysis of proline-containing polypeptides, t3 I A. N. Glazer, R. J. Delange, and D. S. Sigman, in "Laboratory Techniques in Biochemistry and Molecular Biology" (T. S. Work and E. Work, eds.), Vol. 4, Part 1, p. 38. Elsevier/North Holland, Amsterdam, 1976. 2 G. P. Royer and J. P. Andrews, J. Biol. Chem. 248, 1807 (1973). 3 K. D. Vosbeck, B. D. Greenberg, and W. M. Award, Jr., J. Biol. Chem. 250, 3981 (1975). 4 R. Waiter, W. H. Simmons, and T. Yoshimoto, Mot. Cell. Biochem. 30, 111 (1980). 5 E. Kessler and A. Yaron, Eur. J. Biochem. 63, 271 (1976). 6 E. Kessler and A. Yaron, this series, Vol. 45, p. 544. 7 G. Fleminger, Z. Bohak, and A. Yaron, in preparation (1987). 8 A. Yaron and A. Berger, this series, Vol. 19, p. 521.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved,

170

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[17]

The experiments discussed in this article have been helpful in solving some of the principal problems in the semisynthesis of human insulin from porcine insulin. Although they were done on a semipreparative scale, the obtained results indicate that application of immobilized Achromobacter protease I in the semisynthesis of human insulin has great potential as a tool for the large-scale production of human insulin.

[ 17] Application of I m m o b i l i z e d A m i n o p e p t i d a s e s to t h e S e q u e n t i a l H y d r o l y s i s of P r o l i n e - C o n t a i n i n g P e p t i d e s

By G. FLEMINGERand A. YARON Sequence-ordered hydrolysis of peptide chains by exopeptidases followed by quantitative amino acid analysis is a mild method for determination of the amino acid sequence, which is sensitive to the amino acid configuration and structure. ~ The most convenient procedure uses enzymes in the immobilized form, 2,3 permitting removal of enzyme-free samples for analysis as well as the use in sequence of several enzymes in the same reaction mixture. Peptide bonds associated with proline residues--present in a polypeptide chain--are unique in their susceptibility to hydrolysis by aminopeptidases. 4 Clostridial aminopeptidase (CAP; EC 3.4.11.13) 5-7 isolated from Clostridium histolyticurn is able to cleave N-terminal proline residues from polypeptide chains at rates comparable to cleavage rates of other N-terminal amino acid residues. The enzyme does not cleave Nterminal residues if followed by proline. The Xaa-Pro bond is resistant also to other aminopeptidases, except to aminopeptidase P (AP-P; EC 3.4.11.9) which cleaves this bond exclusively,s-12In the present article we describe application of glass-bound CAP and AP-P to the sequential hydrolysis of proline-containing polypeptides, t3 I A. N. Glazer, R. J. Delange, and D. S. Sigman, in "Laboratory Techniques in Biochemistry and Molecular Biology" (T. S. Work and E. Work, eds.), Vol. 4, Part 1, p. 38. Elsevier/North Holland, Amsterdam, 1976. 2 G. P. Royer and J. P. Andrews, J. Biol. Chem. 248, 1807 (1973). 3 K. D. Vosbeck, B. D. Greenberg, and W. M. Award, Jr., J. Biol. Chem. 250, 3981 (1975). 4 R. Waiter, W. H. Simmons, and T. Yoshimoto, Mot. Cell. Biochem. 30, 111 (1980). 5 E. Kessler and A. Yaron, Eur. J. Biochem. 63, 271 (1976). 6 E. Kessler and A. Yaron, this series, Vol. 45, p. 544. 7 G. Fleminger, Z. Bohak, and A. Yaron, in preparation (1987). 8 A. Yaron and A. Berger, this series, Vol. 19, p. 521.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved,

[17]

AMINOPEPTIDASES FOR PROLYL PEPTIDES

171

Assay Method

Principle. The assay of CAP is based on measurement by the acid ninhydrin colorimetric method 14 of proline released from the tripeptide Pro-Gly-Pro. In order to ascertain that only this enzyme activity is measured in preparations of various degree of purity, they were occasionally tested also with the tetrapeptide Pro-Gly-Pro-Pro by the same method. In the assay of AP-P the amount of proline released from poly(L-proline) is determined analogously. Reagents Assay buffer. Sodium barbital buffer (50 raM) is purified by passage through a Chelex 100 column (analytical-grad~ chelating resin, mesh 50-100, BioRad Laboratories, Richmond, CA) and mixed with sodium citrate and COSO4 to final concentration of 50 and 10 mM, respectively. The pH is adjusted to 8.2. Pro-Gly-Pro solution, 0.5 mM in the assay buffer. Poly-L-proline solution (4.17 × 10-5 M) is prepared by dissolving the polymer (MW 6000) in cold assay buffer. The number average molecular weight of poly-L-proline is determined by potentiometric titration of the amino end groups (pKa 9.0). 8 Ninhydrin reagent. Ninhydrin (1.5 g) is dissolved in a mixture of acetic acid (60 ml) and 6 M phosphoric acid (40 ml) by warming at 70° . Prior to use the reagent is diluted with an equal volume of acetic acid. Enzymes. Frozen solutions of soluble C/~P (80 units/ml) in 50 mM sodium phosphate buffer, pH 7.4, containing 1.0 mM EDTA ("storage buffer") and of AP-P (120 units/ml) in 50 mM sodium acetate buffer, pH 5.6, containing 2.0 mM sodium citrate are stored at - 2 0 ° . The glass-bound enzymes are stored in storage buffer at 4 °. The isolation and purification of CAP 5,6 and of AP-P 8 have been described. The procedures for CAP were recently improved 7 mainly due to the finding that most of this extracellular enzyme is cell bound and can be released by mild sonication without causing any apparent damage to the

9 G. Fleminger, A. Carmel, D. Goidenberg, and A. Yaron, Eur. J. Biochem. 125, 609 (1982). i0 E. Holtzman, G. Pillay, T. Rosenthal, and A. Yaron, Anal. Biochem. 162, in press (1987). 11 p. Dehm and A. Nordwig, Eur. J. Biochem. 17, 364 (1970). 12 W. Sidrowicz, J. Szechinski, P. C. Canizaro, and F. J. Behal, Proc. Soc. Exp. Biol. Med. 175, 503 (1984). 13 G. Fleminger and A. Yaron, Biochim. Biophys. Acta 743, 437 (1983). 14 W. Troll and J. Lindsley, J. Biol. Chem. 215, 655 (1955).

172

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[17]

cell. The collected enzyme sonicate had to be purified only 15-fold to achieve homogeneity. Enzyme Immobilization and Properties of the Immobilized Enzyme Preparations Amino-glass beads (100 rag, Coming MAO-3930) are mixed with a fresh solution of glutaraldehyde in water (2.5%, 2.0 ml) for 60 min in vacuo at room temperature in a rotating syringe. The beads are washed with water and the storage buffer (100 ml). CAP in the same buffer (0.1 mg/ml, 80 units/mg, 2.0 ml) or AP-P in 0.1 M sodium acetate buffer, pH 5.6, containing 2.0 mM sodium citrate (0.1 mg/ml, 120 units/mg, 2.0 ml) is added to the activated glass beads and slowly mixed in the rotating syringe partially immersed in an ice bath for 3 hr. The beads are thoroughly washed with the storage buffer and stored in the same buffer at 4 ° . To inactivate any potential serine protease activity the immobilized preparations are treated once with 0.5 ~M phenylmethane sulfonylfluoride in the storage buffer for 20 min at room temperature. The enzyme activity of the beads is 18 and 8% of the original activity of the soluble CAP and AP-P used for immobilization, respectively. No enzymatic activity is released from the beads into the supernatant when the beads are kept in buffers at pH 7.4-8.6 at 4 ° for several days, as judged by the inability of the supernatant to hydrolyze the respective substrate under assay conditions. Stability. Both glass-bound enzymes are stable for at least 5-7 months when kept in the storage buffer at 4° even when frequently used for peptide hydrolysis. Activity of some less used preparations are stable for 8-10 months. The glass-bound enzymes loses activity irreversibly in solutions containing urea at concentrations of 0.5 M or more. Full activity is preserved at 0.2 M NaCI. At higher concentrations, activities are reversibly inhibited. In 50 mM sodium barbital (pH 8.2), 50 mM sodium citrate, 10 mM COSO4 glass-bound CAP is fully active in presence of 10% dimethyl sulfoxide; glass-bound AP-P is fully active up to 5% and at 10% dimethylsulfoxide the activity is 30% lower. The inhibition is reversible. In contrast to soluble CAP which loses activity partially at SDS concentrations exceeding 1 Izg/ml and has no activity at 100 ~g/ml, glass-bound CAP is fully active in presence of SDS concentrations up to 10/xg/ml, and at 100 tzg/ml still has 80% of its original activity. Adsorption. No adsorption to the glass-bound enzymes under the assay conditions is observed with any of the substrates used. Only cytochrome c and substance P are slightly adsorbed in absence of NaCI.

[17]

AMINOPEPTIDASES FOR PROLYL PEPTIDES

173

Absence of Endopeptidase Activity. Denatured cytochrome c which has an acetylated N terminus is treated with the immobilized aminopeptidases under the assay conditions. No ninhydrin-positive material is formed as would be expected if endopeptidases were present in the enzyme preparations. Dependence on pH and Temperature. The pH range of maximal activity of glass-bound CAP and AP-P is 7.8-8.2 and 8.1-8.4, respectively. The optimal temperature is at 40°; at 55° the enzymes are irreversibly inactivated within 1 min. Metal Ion Requirements. 15 The dependence of enzymatic activity on COSO4 of both glass-bound enzymes is bell shaped. For CAP the inhibition at high Co 2+ concentrations is abolished in presence of cobalt citrate which acts as a metallobuffer. For AP-P the inhibition is partially reduced. In the presence of the metallobuffer (50 mM citrate) maximal activity for CAP and AP-P is reached at a total Co z+ concentration of 1.0 mM (1 /xM free Co 2÷) and 5-10 mM (4-9/~M free Co2+), respectively. Dissociation of the enzyme-Co 2+ complex is very slow for glass-bound CAP, and when once exposed to Co 2+ ions, the beads can be washed with Co2+-free buffer and used repeatedly for hydrolysis of peptides in absence of Co z+ ions. This stabilization effect caused by immobilization is not observed for cellulose-bound CAP and glass-bound AP-P. The association constants derived for the interaction of Co 2+ with glass-bound CAP and AP-P are KE-co = 4.5 X 106 M -1 and KE-co = 8.2 x 105 M -1, respectively. 13The immobilization of CAP to glass beads has no effect on the Km found with leucine p-nitro anilide as the substrate. The value (0.5 mM) is the same for the soluble and the glass-bound enzyme and is not influenced by flow rate of the substrate through the enzyme-coated beads packed into a column, indicating that no major diffusion barrier interferes with the enzyme activity. Enzyme Activity Measurements

Soluble Enzymes. CAP or AP-P (10-50/zl, 0.01-0.2 units) is incubated with Pro-Gly-Pro or poly-L-proline solution (1.0 ml), respectively, for 30 min at 40°. Ninhydrin reagent (5 ml) is added to terminate the reaction and the solution is heated for 30 rain in a boiling water bath. After cooling, the absorption at 520 nm is measured. The proline concentration is calculated with the aid of a calibration curve constructed with known concentrations of proline (10-200 nmol). 15 G. Fleminger and A. Yaron, Biochim. Biophys. Acta 789, 245 (1984).

174

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[17]

Immobilized Enzymes. Glass-bound CAP or AP-P (100 mg dry weight) with the Pro-Gly-Pro or poly-L-proline solution, respectively, is rotated at 80 rpm in a 5-ml plastic syringe equipped with a sintered plate and partially immersed in a 40 ° water bath. The needle outlet is plugged by sticking it into a rubber stopcock. After 5 min the proline concentration in the supernatant is determined as described for the soluble enzymes. Definition of Unit. One unit of activity is defined as the amount of soluble or glass-bound CAP or AP-P which catalyzes the formation of 1.0 /zmol of proline in 30 min under the assay conditions. In addition to the above described assays, colorimetric and fluorimetric methods have been developed. CAP can be assayed spectrophotometrically with leucine p-nitroanilide as the substrate. ~6 For aminopeptidase P the intramolecularly quenched fluorogenic substrate Phe(p-NOE)-Pro-Pro-HN-CHz-CH2-NH-ABz, where Phe(p-NO2) is the pnitrophenylalanyl quenching group and ABz is the 2-aminobenzoyl fluorophore, was developed. 9 An improvement in sensitivity is achieved ~° by replacing the Phe(p-NO2) group with N~-(2,4-dinitrophenyl)-L-lysine. A colorimetric method with Gly-Pro-Hyp is used in the isolation of the hog kidney enzyme. ~ Procedure for Sequential Hydrolysis of Polypeptides with Immobilized Aminopeptidases Glass beads (100 mg dry weight) containing immobilized aminopeptidase and the solution of the peptide (0.2-0.5 raM, 1 ml) in the assay buffer containing 0.1 M NaCI and norleucine (0.2-0.5/zmol) are incubated at 40 ° as described under the assay procedure, and aliquots of 10-100/zl of the supernatant are withdrawn for amino acid analysis at different time intervals. For continuation of the hydrolysis with the second enzyme the entire supernatant is pressed through a sintered disk into another syringe containing the other enzyme beads. The enzyme-coated glass beads are washed with the assay buffer prior to each hydrolysis step. Samples (10100/zl) of the initial peptide solution, of enzymatic hydrolyzates, and of the same after total acid hydrolysis (6 N HCI, 110% 22 hr, in vacuo) and drying are diluted with 0.2 M citrate buffer, pH 2.2, and subjected to amino acid analysis using a Beckman Model 121 Automated Amino Acid Analyzer. The total hydrolysis of aliquots is required as a control for potential loss of peptides or their fragments by adsorption to the glass beads. Norleucine is used as an internal standard. 16 A. Carmel, E. Kessler, and A. Yaron, Eur. J. Biochem. 73, 617 (1977).

[17]

AMINOPEPTIDASES FOR PROLYLPEPTIDES

175

Sequential Hydrolysis o f Peptides with Immobilized Aminopeptidases Two kinds of peptides serve as model substrates for the immobilized aminopeptidases: (1) peptides devoid of proline, or those in which a proline residue is present at the amino end, which are hydrolyzed by CAP alone; and (2) peptides containing proline inside the polypeptide chain, requiring for their hydrolysis both CAP and AP-P. Hydrolysis of [Leu]enkephalin (Tyr-Gly-Gly-Phe-Leu, 0.2 mM) by immobilized CAP results in a rapid liberation of the N-terminal tyrosine (/1/2 = 3 min), which is followed by the fast appearance of glycine (tvz = 5 min) and a slower hydrolysis of the remaining dipeptide, Phe-Leu. Complete hydrolysis is achieved within 2 hr. The time course of the sequential hydrolysis of substance P octapeptide (Pro-Gln-Gln-Phe-Phe-Gly-Leu-Met-NH2, 0.5 mM) is shown in Table I. All of the N-terminal proline is released within 75 min, followed by the successive release of glutamine, phenylalanine, glycine, and leucine, as expected. After 135 min the hydrolysis is still not complete, apparently due to the slow cleavage of the Gly-Leu peptide bond and of the methionine amide. The octapeptide angiotensin II (Asn-Arg-Val-Tyr-Val-His-Pro-Phe, 0.5 mM), which contains a proline residue at position 7, is completely hydrolyzed in three steps, as shown in Table II. First, CAP is used to remove the five residues beginning with the N-terminal asparagine. Complete release of these amino acids is achieved after 1 hr. In the second step, AP-P is applied for 30 min to release histidine completely from the remaining tripeptide His-Pro-Phe. Further incubation with CAP for 30 min results in complete hydrolysis of the remaining dipeptide Pro-Phe. TABLE I ENZYMATIC HYDROLYSIS OF SUBSTANCE P OCTAPEPTIDEa Percentage of amino acid residues released Incubation time (min) 16 30 75 135

Pro xGln ~,.Gl%- Phe ~ Ph%- Gly - Leu - Met- NH2

55 80 98 105

23 50 90 100

7 15 76 90

0 3 27 75

0 0 22 68

0 0 0 15

a Concentration, 0.5 mM, in the assay buffer containing 5% dimethyl sulfoxide, using immobilized clostridial aminopeptidase (CAP) (100 mg dry weight).

176

[17]

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

TABLE II ENZYMATIC HYDROLYSIS OF

Step I

II III

Enzyme

Incubation time (min)

CAP CAP CAP CAP CAP AP-P CAP

3 5 10 20 60 30 30

[Asn ~, ValS]ANGIOTENSIN

II a

Cumulative percentage of amino acid residues released A s n - Arg 40 70 95 100 100 100 102

N.D. N.D. N.D. N.D. 95 95 N.D.

Val

- Tyr-

0 40-60 57-95 88-100 100 97 105

0 38 57 88 98 95 92

Val - His - Pro - Phe 0 0-20 15-53 80-92 100 97 105

N.D. N.D. N.D. N.D. 0 108 N.D.

0 0 0 0 0 0 100

0 0 0 0 0 0 90

a The peptide (0.5 raM) was incubated alternately with immobilized clostridial aminopeptidase (CAP) and aminopeptidase (AP-P) (100 mg dry weight each) and samples were withdrawn for amino acid analysis at various time intervals. The amounts of Val 3 and VaP released were estimated from the total amount of valine, asparagine, and tyrosine release. N.D., Not determined.

The applicability of the two immobilized aminopeptidases in achieving complete hydrolysis of polypeptide chains is further demonstrated with tuftsin (Th~-Ly~-P~-Arg where ~ CAP,--~ AP-P). The two enzymes are applied alternately as indicated by arrows above the sequence. In this case, the expected amino acids appear within 20 min at each step. The nonapeptide bradykinin (Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg, 0.5 raM), containing three proline residues in position 2, 3, and 7, is completely hydrolyzed in four consecutive steps by alternately applying the immobilized AP-P and CAP as shown in Table III. The synthetic sequential polymer (Pro-Gly-Pro)r6 is consecutively hydrolyzed by alternately exposing the polymer solution to the two immobilized enzymes as shown in the following scheme (the numbers above the arrows represent the consecutive steps of hydrolysis):

112

P3r

4

I

Pro-Gly-Pro-Pro-Gly-Pro ....

In this experiment 10 hydrolysis steps were performed. In each step one equivalent of proline was found to be released as determined by colorimetric analysis. Altogether 50% of the total proline present in the substrate was released by the enzymes.

[17]

177

AMINOPEPTIDASES FOR PROLYL PEPTIDES TABLE III ENZYMATIC HYDROLYSIS OF BRADYKININa

Step

Enzyme

Incubation time (rain)

I II III IV

AP-P CAP AP-P CAP

10 20 20 20

Cumulative number of equivalents of amino acid residues released Arg

Pro

Gly

Phe

Ser

0.9 N.D. N.D. 1.9

1.1 2.1 2.2 3.3

0 1.0 0.95 1.0

0 0.8 0.85 1.80

0 0 0.83 0.87

a Bradykinin (Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg, 0.5 mM) was hydrolyzed using immobilized aminopeptidase P (AP-P) and clostridial aminopeptidase (CAP) (100 mg dry weight each) alternatively.

Discussion Useful enzymatic preparations are obtained by coupling AP-P and CAP to amino-glass beads by the glutaraldehyde method. Resistance of cytochrome c to hydrolysis and the ordered release of amino acids during the hydrolysis, demonstrated for several peptides, shows that the preparations are free of interfering enzyme activities. The resistance of X-Pro bonds to hydrolysis by CAP and the absolute specificity of AP-P for Nterminal amino acids which are followed by a proline residue make it possible to degrade polypeptides up to any proline residue present in the peptide chain, resulting thereby in peptide fragments with Pro or X-Pro at the amino end. Substrates in the range from tripeptides to dodecapeptides were readily degraded. The stepwise degradation of the sequence-ordered polymer (Pro-Gly-Pro)~, consisting of an average of 30 amino acid residues, demonstrates that the enzymatic hydrolysis is applicable to quite high molecular weight substrates and that at least 10 successive transfers between the enzyme beads can be carried out with the release of a full equivalent of the expected amino acids in each step. Longer polypeptide chains such as reduced and carboxymethylated trypsin inhibitor from lima bean (Mr 9000) and soybean (Mr 21,500), reduced and carboxymethylated RNase (Mr 13,600), and apomyoglobin (Mr 17,900) are not hydrolyzed by either soluble or glass-bound CAP. Sequential degradation with the immobilized aminopeptidases may find application in the analysis and modification of natural and synthetic

178

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[18]

proline-containing oligopeptides. Immobilized CAP and AP-P may also be helpful in the sequencing of proline-containing peptides using dipeptidyl aminopeptidase I (DAP I; EC 3.4.14.1) and dipeptidyl aminopeptidase IV (DAP IV; EC 3.4.14.5). 17 Since these enzymes cannot hydrolyze the X-Pro bond, degradation stops two residues before proline. These two residues can be cleaved off by the combination of CAP and AP-P, thereby permitting continuation of the degradation by DAP I or DAP IV. Hydrolysis of polypeptides, using a mixture of various immobilized proteolytic enzymes, that proceeds almost to completion has been achieved. ~,18 The residual peptides, resisting further hydrolysis, have been shown to contain proline residues. The immobilized CAP and AP-P should therefore be helpful in completing their hydrolysis. ~7 H. C. K r u t z s c h and J. J. Pisano, this series, Vol. 47, p. 391. ~s j. L a s c h , R. Koelsch, P. Roth, A. Gabert, J. Marguardt, and H. H a n s o n , Acta Biol. Med. Ger. 35, 735 (1976).

[18] P e p t i d e S y n t h e s i s U s i n g I m m o b i l i z e d P r o t e a s e s

By HANS-DIETER JAKUBKE and ANDREAS Kt)NNECKE Although the well-established techniques of chemical peptide synthesis have been successful in making peptides and small proteins of biological and industrial interest (see Ref. I for several reviews), synthetic methodology is still in need of innovations. The use of enzymes should provide a biological alternative to chemical peptide synthesis. There has been a revival of interest in the ability of proteolytic enzymes to catalyze the formation of peptide bonds for the synthesis of biologically active peptides and the semisynthesis of proteins (see Ref. 2 for several reviews). The main attraction of proteases lies in their capacity to effect peptide bond formation stereospecifically without the need of side-chain protection. In protease-mediated peptide synthesis, the enzymatic specificity prevents the formation of undesired by-products often formed in the course of conventional chemical synthesis. E. Gross and J. Meienhofer, " T h e Peptides: Analysis, Synthesis, Biology," Vols. 1-9. A c a d e m i c Press, N e w York, 1979-1987. z H.-D. Jakubke, P. Kuhl, and A. K r n n e c k e , Angew. Chem. Int. Ed. Engl. 24, 85 (1985).

METHODS 1N ENZYMOLOGY, VOL. 136

Copyright © 1987by AcademicPress, Inc. All rights of reproduction in any form reserved.

178

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[18]

proline-containing oligopeptides. Immobilized CAP and AP-P may also be helpful in the sequencing of proline-containing peptides using dipeptidyl aminopeptidase I (DAP I; EC 3.4.14.1) and dipeptidyl aminopeptidase IV (DAP IV; EC 3.4.14.5). 17 Since these enzymes cannot hydrolyze the X-Pro bond, degradation stops two residues before proline. These two residues can be cleaved off by the combination of CAP and AP-P, thereby permitting continuation of the degradation by DAP I or DAP IV. Hydrolysis of polypeptides, using a mixture of various immobilized proteolytic enzymes, that proceeds almost to completion has been achieved. ~,18 The residual peptides, resisting further hydrolysis, have been shown to contain proline residues. The immobilized CAP and AP-P should therefore be helpful in completing their hydrolysis. ~7 H. C. K r u t z s c h and J. J. Pisano, this series, Vol. 47, p. 391. ~s j. L a s c h , R. Koelsch, P. Roth, A. Gabert, J. Marguardt, and H. H a n s o n , Acta Biol. Med. Ger. 35, 735 (1976).

[18] P e p t i d e S y n t h e s i s U s i n g I m m o b i l i z e d P r o t e a s e s

By HANS-DIETER JAKUBKE and ANDREAS Kt)NNECKE Although the well-established techniques of chemical peptide synthesis have been successful in making peptides and small proteins of biological and industrial interest (see Ref. I for several reviews), synthetic methodology is still in need of innovations. The use of enzymes should provide a biological alternative to chemical peptide synthesis. There has been a revival of interest in the ability of proteolytic enzymes to catalyze the formation of peptide bonds for the synthesis of biologically active peptides and the semisynthesis of proteins (see Ref. 2 for several reviews). The main attraction of proteases lies in their capacity to effect peptide bond formation stereospecifically without the need of side-chain protection. In protease-mediated peptide synthesis, the enzymatic specificity prevents the formation of undesired by-products often formed in the course of conventional chemical synthesis. E. Gross and J. Meienhofer, " T h e Peptides: Analysis, Synthesis, Biology," Vols. 1-9. A c a d e m i c Press, N e w York, 1979-1987. z H.-D. Jakubke, P. Kuhl, and A. K r n n e c k e , Angew. Chem. Int. Ed. Engl. 24, 85 (1985).

METHODS 1N ENZYMOLOGY, VOL. 136

Copyright © 1987by AcademicPress, Inc. All rights of reproduction in any form reserved.

[18]

PEPTIDE SYNTHESIS USING IMMOBILIZED PROTEASES

179

The application of enzymes for organic synthesis is increasing rapidly in importance 3 and has been stimulated to a considerable degree by the availability of immobilized enzymes. Proteases can be immobilized without loss of function, and the potential of immobilized proteolytic enzymes for peptide synthesis has been demonstratedY The purpose of this contribution is to draw attention to peptide coupling reactions on a preparative scale in which immobilized proteases can be used with the advantage of avoiding reaction conditions which are normally required for chemical condensations. The simplified work-up procedure that becomes possible when immobilized proteases are used, the long-term stability of the immobilized enzyme preparations, and the successful reutilization experiments are among the advantages to such an approach. Approaches to Enzymatic Peptide Synthesis In general, the methods of protease-mediated peptide bond formation can be classified into two basic strategies according to the type of carboxyl component used. In the first strategy this component has a free carboxyl terminus, and the formation of the peptide bond occurs with thermodynamic control as the reverse of peptide hydrolysis. In the second strategy the carboxyl component is employed in an activated form, mainly as an alkyl ester, and the synthesis occurs with kinetic control by competitive partitioning of a rapidly formed acyl-enzyme intermediate between the nucleophile and water. These two strategies are fundamentally different due to the energy required for the conversion of the substrates to the peptide products. In the following two subsections, some kinetic and thermodynamic aspects of protease-catalyzed peptide synthesis are briefly discussed and illustrated by representative experimental procedures. The discussion is valid for both native and immobilized proteases and provides a theoretical basis for understanding the strategy and the problems associated with the use of proteases as biocatalysts for peptide bond formation. The Kinetic Approach Serine and cysteine proteases are known to catalyze acyl transfer from specific substrates to various nucleophiles via an acyl-enzyme intermedi3 j. B. Jones, this series, Vol. 44, p. 831. 4 A. KOnnecke, R. Bullerjahn, and H.-D. Jakubke, Monatsh. Chem. 112, 469 (1981). 5 A. Kfnnecke, M. H~insler, V. Schellenberger, and H.-D. Jakubke, Monatsh. Chem. 114, 433 (1983).

180

[18]

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

ate. In reactions of this type, the protease reacts rapidly with an amino acid or peptide ester R]COOR 3 to form an acyl-enzyme intermediate R~CO-E that reacts, in competition with water, with the amino acid- or peptide-derived nucleophile H2NR 2 to form a new peptide bond6,7: RIco-NHR2

I<s R1COOR3

÷

E

+

E

k2

RICO-E

k3[H20] RIcooH + E

(1) The partitioning of the acyl-enzyme intermediate between water and the added nucleophile is the rate-limiting step. Under kinetic control the peptide product RICO-NHR z should accumulate, provided k4[HzNR 2] ~> k3[H20]. However, with increasing reaction time, a soluble product will be degraded in a secondary hydrolysis step. Therefore, the reaction conditions must be carefully considered in order to obtain the optimal product yield. Peptide bond formation by partitioning of acyl-enzyme intermediates may also be controlled by the specificity of the enzyme. Structural factors in the nucleophile which increase the extent to which it is bound productively at the S'-subsite of the enzyme promote synthesis. The type and concentration of the nucleophile influence the partitioning of the acylenzyme intermediate, and influence therefore the success of peptide synthesis. Since only the free-base form of the amino component reacts with the acyl-enzyme intermediate, and the pK of this reactant is about 8, the pH of the reaction mixture should preferably be higher than this. Examples

Materials and Immobilization Procedures Enzymes are all from commercial sources: o~-chymotrypsin (EC 3.4.21.1.), Le~iva (Prague, Czechoslovakia); papain (EC 3.4.22.2.), J. Fastrez and A. R. Fersht, Biochemistry 12, 2025 (1973). 7 A. Kfnnecke, V. Schellenberger, H.-J. Hofmann, and H.-D. Jakubke, Pharmazie 39, 785 (1984).

[18]

PEPTIDE SYNTHESIS USING IMMOBILIZED PROTEASES

181

Merck (Darmstadt, FRG); trypsin (EC 3.4.21.4) and thermolysin (EC 3.4.24.4.), Boehringer-Mannheim (FRG). Carriers for enzyme immobilization include Enzacryl AA and AH from Koch Light (now distributed by Aldrich), and carboxymethyl cellulose powder from Reanal (Budapest, Hungary). The utility of these materials for covalent binding of enzymes has been described in previous volumes of this series. Macroporous silica with 20 nm pore diameter is silanized, succinylated, and subsequently transformed into activated N-hydroxysuccinimide ester, 8 and the carboxymethylcellulose powder is used as hydrazide. 9 Immobilization of chymotrypsin ~° and thermolysin II to Enzacryl AH and trypsin 9 to carboxymethylcellulose azide is accomplished by standard procedures. Activated silica is coupled with chymotrypsin (1 g carrier and 100 mg enzyme) in 8 ml 0.05 M borate buffer (pH 8.4) for 2 hr at room temperature. Crude papain (3 g) is suspended in 15 ml 0.1 M phosphate buffer (pH 7). After the solution is stirred for 1 hr, insoluble material is separated by centrifugation. The supernatant (60 units/ml) is incubated with 1 g activated Enzacryl AA for 24 hr at 4 °. Trypsin (100 mg in 8 ml 0.1 M phosphate buffer, pH 7) is coupled to 1 g activated Enzacryl AA for 2 hr at room temperature. Residual reactive groups of the carriers are blocked with aqueous 0.5 M ammonium hydroxide/1 M ammonium chloride (acyl azide or ester carriers) or 0.01% aqueous ot-naphthol solution (Enzacryl AA). Adsorbed protein is removed by exhaustive washings of the gels. Immobilized enzymes are stored in water at 4 °, except immobilized papain, which is washed with methanol and dried at 70° . A s s a y 12 of immobilized chymotrypsin preparations is spectrophotometrically performed with N-glutaryl-Leu-Phe-Nan ~2a as substrate. ~3 Trypsin and papain are assayed titrimetrically with Bz-Arg-OEt as substrate using the Radiometer pH-stat TTT ld assembly at pH 7.9 and 6.2, respectively. Titrimetric assay procedures are more convenient for immobilized proteases, 12 and the measure of activity per gram of support is 8 H.-H. Weetall, this series, Vol. 44, p. 134. 9 E. M. Crook, K. Brocklehurst, and C. W. Warthon, this series, Vol. 19, p. 963. 10j. K. Inman and H. M. Dintzis, Biochemistry 8, 4074 (1969). i1 R. Epton, B. L. Hippert, and T. H. Thomas, this series, Vol. 44, p. 84. 12 B. Mattiason and K. Mosbach, this series, Vol. 44, p. 335. 12a Amino acid symbols except Gly denote L configuration. Other abbreviations: Ac, acetyl; Bz, benzoyl; Z, benzyloxycarbonyl; OMe, methyl ester; OBu t, tert-butyl ester; Nan, 4nitroanilide. Symbols and abbreviations conform to the IUPAC-IUB Commission recommendations. t3 H.-D. Jakubke, H. D~iumer, A. K6nnecke, P, Kuhl, and J. Fischer, Experientia 36, 1039 (1980).

182

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[18]

sufficient for practical use. The protein content is determined from the difference in absorbance of the enzyme solution prior to and after immobilization. Substrates and reagents are from Serva (Heidelberg, FRG) and Reanal (Budapest, Hungary), respectively. All chemicals and solvents are reagent grade. The amino acid and peptide derivatives used as carboxyl and amino components in the enzyme-catalyzed condensations are drawn from our collection or are synthesized by standard procedures. Prior to use, their identity and purity are confirmed by thin-layer chromatography, melting point determination, elementary analysis, and HPLC. Porcine des-Ala (B30) insulin and Thr-OBu t are kindly supplied by Dr. K.-D. Kaufmann, Institute of Drug Research, Academy of Sciences of GDR. High-performance liquid chromatography (HPLC) is carried out with Liquochrom 307 equipment of Labor MIM (Budapest, Hungary), using isocratic elution and variable-wavelength UV detection. Separations are performed with a Hewlett-Packard prepacked 200 × 4.6 mm column containing LiChrosorb RP-18, 10 ~m, and a homemade 150 × 4 mm column packed with high-temperature silanized RP-6 material 14based on LiChrosorb Si-100, 10 /~m. Solvents are degassed separately by sonication in vacuo, if necessary. It should be stressed that immobilization, assays, and also the enzymatic synthesis (see below) do not require special equipment, and, therefore, can be performed in any laboratory.

Chymotrypsin-Catalyzed Synthesis of Ac-Phe-Ala-NH215 Chymotrypsin bound to macroporous silica (250 mg wet gel containing about 2.5 mg of immobilized protein) is placed in a reaction vessel commonly used for solid-phase peptide synthesis with a G 4 frit at the bottom (10 mm in diameter), fitted with an overhead stirrer. A solution of AlaNH2, obtained by neutralization of 49.4 mg (0.4 mmol) Ala-NH2 • HCI with 0.1 ml cold 4 N NaOH and dilution with 1.6 ml 0.2 M carbonate buffer (pH 10), is added, and the reaction mixture is stirred efficiently (avoiding any contact of the stirrer with the glass wall of the vessel), The coupling reaction is initiated by addition of 0.3 ml of a stock solution of Ac-PheOMe in dimethylformamide containing 44.2 mg (0.2 mmol of the ester). After 12 min the solution is removed by suction filtration, and the immobilized enzyme is washed successively with three 3-ml portions of water. The next coupling cycle can be started by adding the substrate solutions in the order given above. The combined filtrates are worked up by addition 14 T. Welsch and H. Frank, J. Chromatogr. 267, 39 (1983). 15 A. K6nnecke and H.-D. Jakubke, unpublished results.

[18]

PEPTIDE SYNTHESIS USING IMMOBILIZED PROTEASES

183

of 1 ml 1 N HCI and rotary evaporation to dryness. The residue is dried in vacuo, and inorganic salts are removed by extraction of solid with three 30-ml portions of boiling ethyl acetate. The ethyl acetate extract is subsequently rotary evaporated to dryness. The residue is dissolved in a minimum amount of methanol, and the peptide is precipitated by addition of two volumes of diethyl ether. Yield was 34.4 mg (62%); mp 241-242 ° (see Ref. 6; mp 242-244 °) unchanged after recrystallization from methanol.

Papain-Catalyzed Formation of Z-Arg-Ala-NH216 Phosphate buffer (2 ml, 0.1 M, pH 8; with 0.3 M NaCl, 1 mM EDTA) containing 0.1 M Z-Arg-OMe • HCI and 0.2 M AIa-NHz • HCI as reactants is placed in the titration beaker of the autotitration device TTT ld thermostated at 35°. The solution is rapidly stirred, and the pH adjusted to 8 with ! N NaOH. The condensation is initiated by addition of 20 mg dry Enzacryl AA-papain (6.9 BAEE units). A slow stream of argon is bubbled through the solution and the reaction is allowed to proceed for 23 min with automatic readjustment of the pH to 8. The extent of synthesis may roughly be estimated from the amount of base consumed by the formation of Z-Arg-OH. More precisely, the reaction is monitored by HPLC using RP-6 column and methanol/0.05 M phosphate buffer, pH 7 (30/70, v/v) as eluent; capacity factors used were k' 2.21 for Z-Arg-OH, 3.28 for Z-ArgAla-NH2, and 8.86 for Z-Arg-OMe with methanol as to marker. After 23 min Z-Arg-OMe was completely consumed and 78% Z-Arg-Ala-NH2 along with 22% Z-Arg-OH were formed. After the immobilized papain was washed with water, it was used in a repeat of the experiment, which provided the same result after 26 rain.

Trypsin-Catalyzed Synthesis of Z-Lys-Leu-NHe 5 Enzacryl AA-trypsin (350 mg wet gel containing about 2.5 mg of immobilized protein or 70 units) is placed in a reaction vessel similar to the one used for the synthesis of Ac-Phe-Ala-NH2. A solution of 79.4 mg (0.24 mmol) Z-Lys-OMe. HCI and 60 mg (0.36 mmol) Leu-NHE. HC1 in 2 ml t M carbonate buffer (pH 10) containing 12.5% (v/v) dimethylformamide is added to the catalyst, and the mixture stirred vigorously. After 45 min the liquid phase of the reaction mixture is removed by suction, and the precipitated product removed from the immobilized trypsin by extraction with three 2-ml portions of precooled (4°) methanol followed by 3 ml ~6 V. Schellenberger, A. K6nnecke, and H.-D. Jakubke, in "Peptides, 1984" (U. Ragnarsson, ed.), p. 201. Almqvist & Wiksell, Stockholm, 1984.

184

IMMOBILIZED

ENZYMES/CELLS

IN ORGANIC

SYNTHESIS

[18]

water. The immobilized enzyme is washed additionally with five 3-ml portions of water, and the next coupling cycle is started by adding new solutions of substrates. After rotary evaporation of the methanol from filtrate, the residue is dissolved in about 50 ml water, applied directly to an Amberlite IRA-400 ion-exchange column (2 x 6 cm), and subsequently eluted with water until no change in optical density occurs. The eluate is rotary evaporated and the residue dried in vacuo. The yield was 51.8 mg (66%) pure (TLC and HPLC) Z-Lys-Leu-NH2; mp 140-145 ° (dec.); [a]~3 - 10.2 ° (c 1; MeOH). After five cycles the yield still amounted to 67%, and the activity of the immobilized trypsin was 71% of the initial acdespite repeated treatment with methanol. Chymotrypsin-Catalyzed Synthesis o f Z-Gly-Phe-Leu-NHe 4

Wet Enzacryl AH-chymotrypsin (350 mg; about 3 mg of immobilized protein) is placed in a small tube equipped with a magnetic stirrer. Stock solutions of 1.6 ml 0.2 M carbonate buffer (pH 10) containing 23.4 mg (0.2 mmol) Leu-NH2. HCI and 400/zl carbon tetrachloride containing 73.5 mg (0.2 mmol) Z-Gly-Phe-OMe are added. After vigorous stirring for 2 hr, the solution is removed by suction filtration and the enzyme is treated with three 5-ml portions of dimethylformamide for removing the precipitated product from the immobilized biocatalyst. The immobilized enzyme is washed with 5 ml water, and the next coupling cycle is started by adding reactants. The dimethylformamide extract is diluted with about 40 ml water and stored overnight at 4 °. The precipitated product is isolated by filtration and dried in vacuo. Yield was 60 mg (64%); mp 209-211°; [a] 22 - 20.1 ° (c 1; MeOH); after the third use of the immobilized enzyme preparation, the yield still amounted to 57% (53.4 mg). The Thermodynamic Approach The thermodynamic barrier to peptide bond synthesis is predominantly due to the energy required for the transfer of a proton from the positively charged a-amino group to the negatively charged a-carboxyl group. The equilibrium constant Kco, of the conversion of the nonionized reactants into the product is determined by the concentration of the nonionized forms in Kion : K,on RICOO - + Ha+NR 2 .

K¢on ~ RICOOH

+ H2NR:

.

• R1CO--NHR:

+ H20

(2)

Because of the unfavorable equilibrium position, relatively high concentrations of the reactants are required. In order to achieve an appreciable

[18]

PEPTIDE SYNTHESIS USING IMMOBILIZED PROTEASES

185

degree of synthesis, one needs to resort to some expedients and unusual techniques. 2 Furthermore, the reaction can be shifted toward synthesis by the insolubility of the newly formed peptide when the solubility of the product is below the equilibrium concentration. The main reason for promoting synthesis is to increase the equilibrium constant K~o, for proton transfer from the H3N + to C O 0 - end groups of reactants. This is accomplished by addition of high concentrations of water-miscible organic cosolvents.17 A similar favorable influence on this equilibrium constant can also be achieved by working in biphasic organic-aqueous mixtures. 18Furthermore, in special cases water-immiscible cosolvents promote synthesis by continuous removal of product from equilibrium if the peptide formed is much more soluble in the organic solvent.19 The only function of the protease is to accelerate the attainment of the equilibrium Kcon. Examples

Thermolysin-Catalyzed Synthesis of Z-Phe-Leu-NH25 To a small tube fitted for magnetic stirring, 29.9 mg (0.1 mmol) Z-PheOH, 21 mg (0.16 mmol) Leu-NH2, and 2 ml 0.2 M Tris-maleate buffer (pH 7) are added. The condensation is initiated by adding 300 mg wet Enzacryl AH-thermolysin (containing about 0.6 mg protein). The coupling reaction is allowed to proceed for 20 hr at room temperature with constant stirring. The precipitated product is separated from the immobilized thermolysin by treatment with three 3-ml portions of cold dimethylformamide followed by addition of 5 ml water. The extract was diluted with 35 ml water and stored at 4° overnight. The precipitated product is filtered and dried in vacuo, yielding 34.5 mg (84%) pure (TLC and HPLC) Z-Phe-Leu-NH2, mp 191-192 ° (see Ref. 19; mp 190-192°). After four uses of the immobilized thermolysin, the yield still amounted to 70%.

Trypsin-Catalyzed Formation of Bz-Gly-Lys-Leu-OMe 2° A plastic vial (2 ml) is filled with 7.3 mg (0.025 mmol) Bz-Gly-Lys-OH and 90.8 mg (0.5 mmol) Leu-OMe. HCI. The substrates are dissolved by adding 400/xl dimethyl sulfoxide and 50/zl 0.2 M Tris-HC1 buffer (pH 7, 17 G. 18 K. 19 p. 2o A.

A. Homandberg, J. A. Matfis, and M. Laskowski, Jr., Biochemistry 17, 5220 (1978). Martinek and A. N. Semenov, J. Appl. Biochem. 3, 93 (1981). Kuhl and H.-D. Jakubke, Z. Chem. 22, 407 (1982). K6nnecke, M. Hansler, and H.-D. Jakubke, Z. Chem. 26, 292 (1986).

186

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[18]

10 mM CaC12). The apparent pH is adjusted to 7 with 8 N sodium hydroxide, and buffer is added to a total volume of 500/zl. The condensation is initiated by adding 150 mg wet cellulose-bound trypsin (66 units), and the reaction allowed to proceed for 2 hr at room temperature with constant shaking. HPLC analysis (RP-6 column eluted with methanol/0.1% phosphoric acid (1 : 1, v/v)) indicated the formation of 93% Bz-Gly-Lys-Leu-OMe. After five uses of the immobilized enzyme preparation, no loss in yield or activity was observed.

Trypsin-Catalyzed Formation of Human Insulin tert-Butyl Ester from Porcine 30&deAla-insulin and Threonine tert-Butyl Ester 30B-deAla-porcine insulin (30 mg, 5/xmol) and 52.5 mg (0.3 mmol) ThrOBu t are dissolved in 300/zl dimethylformamide and 150/zl 0.25 M TrisHCI buffer (pH 6.6, 10 mM CaCI2) in a plastic vial (2 ml). The apparent pH is adjusted to 6.6 with 6 N HCI, and buffer added to a final volume of 500 /A. Carboxymethylcellulose-bound trypsin (150 mg wet gel containing 66 units of enzyme) is added, and the coupling reaction is allowed to proceed at room temperature with constant shaking. As indicated by HPLC, equilibrium is reached within 90 min. Yield was 70% human insulin tert-butyl ester; the amount of 23B-30Bde-octapeptide insulin was below 2%. The analytical separations were performed using RP-18 column eluted with A + B21 (70:30, v/v) [A is 0.1 M monobasic sodium phosphate in water/2-methoxyethanol (95 : 5, v/v) adjusted to pH 2 with 85% phosphoric acid; B is acetonitrile/2-methoxyethanol (95 : 5, v/v)]. Due to the sensitivity of the retention of insulins to small changes of eluent composition, nearly reproducible results can only be obtained by mixing the eluent by weight; k' was 3.6 for 30B-deAla insulin and 10.8 for human insulin ester. Isolation of the product could be accomplished by conventional chromatographic procedures as described in literature. 22 Discussion As already discussed by Jones 3 in a previous volume of this series, the use of soluble and immobilized enzymes in synthetic organic chemistry can be of advantage if the reactions to be catalyzed require high regioand/or stereospecificity. A very important problem of classical peptide synthesis is to maintain chiral integrity during peptide bond formation. 2J W. M6nch and W. Dehnen, J. Chromatogr. 147, 415 (1978). 22 A. Jonczyk and H.-G. Gattner, Hoppe-Seyler's Z. Physiol. Chem. 362, 1591 (1981).

[18]

PEPTIDE SYNTHESIS USING IMMOBILIZED PROTEASES

187

This is guaranteed using proteases as stereospecific biocatalysts. The experimental procedures given in this contribution demonstrate that covalently bound proteases catalyze peptide bound formation as effectively as the native enzymes, since the substrate specificity and stereospecificity appear not to be altered by immobilization. Although the use of immobilized enzymes offers some important advantages, several requirements must be fulfilled: (1) The immobilization procedure should be as simple as possible and provide high and reproducible activity yields. Covalent coupling procedures should be preferred to other immobilization techniques to minimize leakage. (2) Especially for thermodynamically controlled synthesis, the need of rapid attainment of the equilibrium requires high enzyme concentrations. (3) The support material should be cheap, easily derivatized, and stable under the reaction conditions, and should possess no significant swelling capacity. The essential advantages in the application of covalently bound proteases can be summarized as follows: (1) The immobilized protease can easily be recovered from the reaction mixture. (2) The peptides synthesized are free of contamination by proteolytic activities and denaturated protein. (3) Due to the increased stability in the presence of organic solvents, higher concentrations of such solvents can be used to influence the position of the thermodynamic equilibrium. For example, in the presence of 80% (v/v) dimethyl sulfoxide, the equilibrium concentration of Bz-GlyLys-Leu-OMe (see procedure) is 93%. 20 In contrast, native trypsin is already inactivated by concentrations higher than 65% (v/v) of this cosolvent. (4) The simple recovery and enhanced stability of the immobilized enzyme preparations guarantee the multiple reuse of the biocatalysts. Kinetically controlled syntheses promise favorable results for use with immobilized proteases, because only low concentrations of enzymes and organic solvents are needed merely to dissolve the reactants. On the other hand, thermodynamically controlled reactions need high enzyme concentrations for rapid attainment of equilibrium. A favorable position of the equilibrium can be obtained only in the presence of high concentrations of organic cosolvents (see procedures for human insulin ester or Bz-GlyLys-Leu-OMe) which, unfortunately, reduce the activity of the enzyme dramatically. Prognosis Due to the high degree of specificity of proteases, enzymatic peptide synthesis does not have the versatility of chemical coupling methods and therefore suffers from some limitations. It is evident from the foregoing discussion that some of the known deficiencies (e.g., coprecipitation of

188

I M M O B I L I Z E D E N Z Y M E S / C E L L S IN ORGANIC SYNTHESIS

[19]

soluble proteases with products, inactivation of native enzymes by organic solvents, costs of enzymes) can be overcome by employing immobilized instead of native proteases. For continuous processing, the introduction of suitable solubilizing protective groups may be helpful in overcoming solubility problems with oligopeptides. A main goal of research in this field is to find an appropriate set of proteolytic enzymes which will widen the scope of enzymatic peptide bond formation. It may be expected that the methodology of protease-catalyzed peptide synthesis and semisynthesis of proteins using the approach suggested here will be considerably improved in the near future. Acknowledgments We are grateful to Dr. T. Welsch for generous help with analytical HPLC. We are much obliged to Dr. John H. Jones for reading and commenting on the manuscript.

[19] A c t i v i t y a n d C o n f o r m a t i o n of E n z y m e s in R e v e r s e M i c e l l a r Solutions

By P. L. LuIsI and B. STEINMANN-HOFMANN In this paper some of the basic concepts and latest results concerning enzymes in reverse micelles are reviewed. In a first part we consider the general experimental aspects, and in the second part a review of enzyme activity and conformation is presented. Of necessity, some important aspects of the field are not adequately covered, for example, the thermodynamic and kinetic problems connected with the formation and stabilization of reverse micelles. Also the biological aspects are not treated in this review, but the reader is referred to other reviews or books that have appeared in the last few years ~-5 as well as to the series of articles by Eicke, Robinson, Magid, Hauser, Montal, El Seoud, Laane, de Schryver, and their co-workers in Ref. 6. J. H. Fendler, "Membrane Mimetic Chemistry," Chap. 3. Wiley (Interscience), New York, 1982. 2 H. F. Eicke, Top. Curr. Chem. 87, 85 (1980). 3 K. Martinek, Eur. J. Biochem. 155, 453 (1985).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

188

I M M O B I L I Z E D E N Z Y M E S / C E L L S IN ORGANIC SYNTHESIS

[19]

soluble proteases with products, inactivation of native enzymes by organic solvents, costs of enzymes) can be overcome by employing immobilized instead of native proteases. For continuous processing, the introduction of suitable solubilizing protective groups may be helpful in overcoming solubility problems with oligopeptides. A main goal of research in this field is to find an appropriate set of proteolytic enzymes which will widen the scope of enzymatic peptide bond formation. It may be expected that the methodology of protease-catalyzed peptide synthesis and semisynthesis of proteins using the approach suggested here will be considerably improved in the near future. Acknowledgments We are grateful to Dr. T. Welsch for generous help with analytical HPLC. We are much obliged to Dr. John H. Jones for reading and commenting on the manuscript.

[19] A c t i v i t y a n d C o n f o r m a t i o n of E n z y m e s in R e v e r s e M i c e l l a r Solutions

By P. L. LuIsI and B. STEINMANN-HOFMANN In this paper some of the basic concepts and latest results concerning enzymes in reverse micelles are reviewed. In a first part we consider the general experimental aspects, and in the second part a review of enzyme activity and conformation is presented. Of necessity, some important aspects of the field are not adequately covered, for example, the thermodynamic and kinetic problems connected with the formation and stabilization of reverse micelles. Also the biological aspects are not treated in this review, but the reader is referred to other reviews or books that have appeared in the last few years ~-5 as well as to the series of articles by Eicke, Robinson, Magid, Hauser, Montal, El Seoud, Laane, de Schryver, and their co-workers in Ref. 6. J. H. Fendler, "Membrane Mimetic Chemistry," Chap. 3. Wiley (Interscience), New York, 1982. 2 H. F. Eicke, Top. Curr. Chem. 87, 85 (1980). 3 K. Martinek, Eur. J. Biochem. 155, 453 (1985).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

[19]

ENZYMES IN REVERSE MICELLES

189

Preparative Aspects

Purification of Aerosol OT The system which has mostly been used in connection with enzymes solubilized in hydrocarbon micellar solutions is isooctane/water AOT, where AOT stands for Aerosol OT, namely bis(2-ethylhexyl)sodium sulfosuccinate. In turn, the reason for the use of this system for enzymes CH

yH~-- CO0- CH2-- CH- CH2-- CH2--CH2--CH3 NO® @03 S-- CH-CO0- C H ~- ?H- CH2--CH ~- CH~- CH ~

#H2 CH 3 Bis(2-ethylhexyl)sodium sulfosuccinate, "Aerosol OT" (AOT)

stems simply from the fact that it is the most studied system per se due to the work of Eicke in Basel} ,7 Reverse micellar solutions are simply prepared by dissolving AOT in a hydrocarbon, up to a concentration ranging between 25 and 300 mM, then adding the desired amount of water (typically in the range of 0.1-5%), and shaking by hand (usually no sonication is needed), until a clear solution is obtained. There is of course only a limited region of values of AOT-water concentration and temperature where the reverse micellar aggregates are stable. This is shown in Fig. 1, which is a typical phase diagram 8 representing the different equilibria existing in this ternary system. AOT can be obtained from different commercial sources. For most technical applications, its purification and a check of chemical purity are not essential, but in fine chemical work (e.g., spectroscopy and enzymatic activity) a series of artifacts can originate from impurities in the surfactant. The most commonly employed commercial sources of AOT are Fluka (product 86140), Cyanamid (24967-01 Complenix-100), Serva (20540), and Sigma (D0885). In all these samples one can detect UVabsorbing impurities and acid impurities, as welt as salts.

4 p. L. Luisi and L. S. Magid, Crit. Rev. Biochem. 20, 409 (1986). 5 p. L. Luisi, Angew. Chem. 97, 446 (1985). 6 p. L. Luisi and B. Straub (eds.), "Reverse Micelles." Plenum, New York, 1984. 7 M. Zulauf and H. F. Eicke, J. Phys. Chem. 83, 480 (1979). 8 B. Tamamushi and N. Watanabe, Colloid Polymer Sci. 258, 174 (1980).

190

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

J-C8H 18

H20

50

AOT

FIG. I. Phase diagram of aerosol OT/water/isooctane system at 25°. L~, Aqueous micellar solution; 2L, emulsion; L2, reverse miceUar solution; 2C, liquid crystals of lamellar (D) and hexagonal (F) types; L + LC, liquid crystals in equilibrium with solution (mainly L2). From Tamamushi and Watanabe. 8

For our purposes (spectroscopic work and enzyme activity determinations) we found that the purification methods described in the literature were not entirely satisfactory. Some of our first experiments, in which we simply purified AOT according to literature, reflect this situation, for example our first paper on a-chymotrypsin. 9 The extreme pH sensitivity of the activity data was due in part to acid impurities, as later documented. 6 We now rely entirely on a HPLC purification procedure, which is satisfactory for eliminating UV-absorbing and acid impurities from the commercial samples. The procedure is as follows. A preparative reversed-phase HPLC column (~Bondapak C~8, Perkin-Elmer, 25 × 2.5 cm) is used. The elution mixture is methanol-water 75:25 (v/v) with a flow rate of 10 ml/min. AOT from Serva (300 mg/ml) is dissolved in the same mixture, and 5 ml of this solution is injected at each run. Both UV and RI (refractive index) detection give the same elution profile (Fig. 2a). The fractions of the central peak are collected. Typically, about 5.5 g of purified AOT can be obtained after 10 injections ( - 1 day of work), which corresponds to an overall yield of 52%. Figure 2, b and c ~° show the UV spectra and the titration curve for our current AOT preparation method compared to other methods. The UV absorbance between 250 and 340 nm is clearly the lowest of all AOT samples. The pH titration curve shows 9 S. Barbaric and P. L. Luisi, J. Am. Chem. Soc. 103, 4239 (1981). l0 p. L. Luisi, P. Meier, V. E. Imre, and A. Pande, in "Reverse Micelles" (P. L. Luisi and B. E. Straub, eds.), p. 332. Plenum, New York, 1984; see also B. Steinmann, H. J/ickle, and P. L. Luisi, Biopolymers 25, 1133 (1986).

[19]

ENZYMES IN REVERSE MICELLES

191

od 220

I

0.5

0.1

1 J

O-

:

I

0

4

~ 3 o

1

250

270

290 nrn

310

330

1

2 prnoles

3

4

5

NaOH

FIG. 2. (a) HPLC chromatography ofAOT. A reverse a-phase C~8 column (25 x 2.5 cm) has been used; flow rate 10 ml/min, elution mixture methanol/water 75 : 25 (v : v). (The arrow indicates the beginning of sample collection.) (b, c) UV absorption (b) and potentiometric titration profiles (c) of various AOT samples. The spectra were obtained using 50 mM AOT solutions in UV-grade isooctane, and titrations were performed in 10 ml of 1 : 1 (v : v) watermethanol solution of - 2 . 2 mmol of AOT, using 0.1 N NaOH as the titrating agent. AOT samples: ( n ) Obtained from Serva; (11) purified as in ref. 10a; (Q) purified as in ref. 10b; (A) purified as in ref. 11; (A) purified by HPLC. Luisi et al? °

that the obtained preparation is neutral; i.e., the beginning of the curve lies at pH 7. t0a C. A. Martin and L. J. Magid, J. Phys. Chem. 85, 3938 (1981). ~0bp. D. I. Fletcher, N. F. Pen-ins, B. H. Robinson, and C. Toprakcioglu. 6 U M. Wong, J. K. Thomas, and T. Novak, J. Am. Chem. Soc. 99, 4730 (1977).

192

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

The AOT samples, once carefully dried under vacuum from the solvents used for the purification steps, can be kept over periods of months without apparent deterioration. The water content in the AOT powder can be checked by NMR or Karl Fischer titration, and is usually lower than 0.1 mol water per mole AOT. The physical characterization of reverse micelles before the uptake of enzymes is an important endeavor, both to study the micelles per se and also for understanding the mechanism of the solubilization of biopolymers. As already mentioned, structure and composition of the micelles will not be considered in detail here. It should simply be mentioned that on the basis of different physical methods, reverse micelles appear to be spherical, and mostly monodisperse. 7,11-14Also, the dimensions are determined by the ratio w0 = [water]/[AOT]. 7'11 Reverse micelles swell by increasing w0; their diameter generally ranges between 40 and 200 ,~. It is relevant for the discussion of enzyme activity and conformation that up to w0 = 6-8 the water of the water pool is assumed to be firmly bound to the micellar internal wall. Only above such a w0 limit can one speak of free water 11,15 (actually beyond this point the term microemulsion would be more correct than the term micelle).

Solubilization of Enzymes in Reverse Micellar Solutions Three methods have been proposed and utilized for the solubilization of enzymes in reverse micellar solutions. They are schematically illustrated in Figs. 3-5. The "injection method" (Fig. 3a) is the simplest: first one prepares the hydrocarbon solution of the surfactant, for example 50 mM AOT; and to this solution a small amount of a concentrated aqueous protein solution is added. For example, to 1 ml of the micellar hydrocarbon solution, 10/xl of a 100 mM lysozyme solution in 50 mM phosphate buffer, pH 7, is added. The enzyme solution must be added gently and in tiny droplets. After a short period of turbidity, gentle hand shaking produces the clear solution of lysozyme in the hydrocarbon micellar solution. It is useful to build "stability diagrams" like those represented in Fig. 3b, which give the wo/T region where the micellar solution is clear and stable, with or without the enzyme. The problem with the injection technique is that one does not know a priori the amount of protein that a given reverse micellar solution can take, especially when one carries out spectroscopic investigations for which a large amount of enzyme in solution is needed. There is then the 12j. Sunamoto, T. Hamada, T. Seto, and S. Yamamoto, Bull, Chem. Soc. Jpn. 53, 583 (1980). ~3j. Tabony, A. Llor, and M. Drifford, Colloid Polymer Sci. 261, 938 (1983). 14 F. M. Menger, J. A. Donohue, and R. F. Williams, J. Am. Chem. Soc. 95, 286 (1973). i5 p. D. Fletcher and B. H. Robinson, Ber. Bunsenges. Phys. Chem. 85, 863 (1981).

[19] a

ENZYMES IN REVERSE MICELLES

Wo 70

protein in water

193

j .;:.,\. i ¢.

•

~ ' ~,D

60 50

I i I !

\

". --

. . . . . . . . . . .

40 cloodv

,

.

li

30

/ jl/

clear solution

20 10

-30

-20

-10

0

10

20

30

40

50

60

70 t

°C

FIG. 3. Solubilization by the injection technique. (a) Injection procedure. (b) Stability curve. A, 50 mM AOT/isooctane/50 mM borate buffer, pH 8.5. B, system as in A + lipoxygenase, Coy = 1.6 × 10 -7 M. C, system as in A + ribonuclease, coy = 1.75 × 10 -4 M. D, system as in A + ribonuclease, Cov = 1.75 × 10-5 M. Coy is the overall concentration, i.e., referred to the total volume (hydrocarbon and water).

danger of preparing supersaturated solutions. In this case the excess protein may tend to precipitate out of the micellar solution, giving rise to a decrease in the optical density during time and to artifacts during the physicochemical characterization. In principle, one should first construct stability diagrams of the type seen in Fig. 3b, in which the concentration of protein is varied as an independent parameter for the micellar stabilization. This is, however, very lengthy and complicated. The second method to solubilize proteins in a hydrocarbon micellar solution is the phase-transfer m e t h o d 16A7 shown diagrammatically in Fig. 4a. In this method the protein is in the water solution and the supernatant is the hydrocarbon micellar solution. With gentle shaking, and with a rate and efficiency which depend on buffer, pH, and concentration, the protein initially present in the water phase is transferred into the micellar solution. An example is given in Fig. 4b. The process of transfer is very slow, as it requires a few hours. A new study of this process, from which these data have been taken, will soon be published by our group. The protein solubilized in the hydrocarbon phase can be transferred again into the aqueous solution with an inverse treatment (the "backward transfer"), by changing the nature of the salt in the water phase. The two ~6 p. L. Luisi, F. J. Bonner, A. Pellegrini, P. Wiget, and R. Wolf, Heir. Chim. Acta 62, 740 (1979); see also Meier. 5° 17 M. E. Leser, G. Wei, P. L. Luisi, and M. Maestro, Biochem. Biophys. Res. Comrnun. 135, 629 (1986).

194

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

surfactant in hydrocarbon

protein in water

FIG. 4. Solubilization by phase transfer. (a) Transfer from the aqueous solution. (b) "Forward transfer" of trypsin from an aqueous solution into AOT/isooctane as a function of salt concentration. From M. Leser et al.t7

processes can be combined together to give the so-called "double transfer"lG,17: a micellar hydrocarbon solution is now bridging two water phases in an H tube. The protein is initially present in the first water phase, and is vectorially transferred into the second water phase by crossing the hydrocarbon bridge. The reverse micelles are the agents of such transport, and the driving force is provided by the difference in the salt concentration between the two water phases. The advantage of the phase-transfer solubilization method with respect to the previous one lies mostly in the fact that the obtained micellar solution is thermodynamically stable; no supersaturation effect is possible. Also, the amount of water which is transferred into the hydrocarbon phase is the one which corresponds to the thermodynamic equilibrium of the system. The advantages are often overweighed by the disadvantages: the procedure is slow, the water and protein concentrations so obtained cannot be varied, and they must be determined at the end of the transfer. Because of the lengthy time required the activity of the enzymes also decreases, although not drastically.17 Let use consider finally transfer from the solid state of proteins, which is of course particularly suited for water-insoluble proteins (Fig. 5). In fact, by this method Delahodde et al.18 have been able to solubilize the water-insoluble lipophilin, a constituent of the myelin structure which has never been investigated in solution (see Fig. 5b). An example of this procedure, the case of lysozyme, which is very soluble in water, is shown in Fig. 5b.19 However, in this case the procem A. Delahodde, M. Vacher, C. Nicot, and M. Waks, F E B S Lett. 172, 1603 (1984); see also, M. Waks, Proteins 1, 4 (1986). t9 C. Grandi, R. E. Smith, and P. L. Luisi, J. Biol. Chem. 256, 837 (1981).

[19]

ENZYMES IN REVERSE MICELLES

195

% Transfer Trypsin 90

pH = lO

20 m M B o r a t e 50 m M A O T CaC12

8O

----- N a C I

70

,

6O

50

\

4O

% %. % k

%. x

%, %

30

20

10

|

0,001

,.

I

0.005

i

I

!

O,01

0.05

O I

M

Fro. 4b.

dure of solubilization from the solid state may offer some advantages.

From a practical point of view, one can reach higher concentrations of solubilized protein, since one is not limited (as in the injection method) by the maximal concentration which can be reached in the aqueous stock solution. Furthermore, thermodynamically stable solutions are obtained, where w0 can be varied independently, in contrast to the previous phasetransfer method.

196

IMMOBILIZED

ENZYMES/CELLS

b

IN ORGANIC

~ 2/-,O

#

: {/{/t'

200

;1-1

0 C~

{

~20

> 0

[19]

SYNTHESIS

40

{

! I

=~ 11L1

¢o

22!2 33'3 ~Ez s%--%

21o 3'o to

5'o oo

% H20

2t,~

protein powder

2O T ~1~

9A~6 ~,o u n ~,~2

1'

O

~cZ.

/

_J

111

o

2~2 3~3

Wo ~.'o z'o 3.'0

%H~O

J

4M.

555

d.o

s'o

60

FIG. 5. Solubilization by phase transfer from the solid state. (a) Extraction from the solid state. (b) Solubility of lysozyme during extraction from the solid states, expressed in overall concentration (A) and concentration in the water pool (B). To 50 mM AOT/isooctane containing various amounts of water (0,©) or aqueous buffer (A, A) solid lysozyme was added, followed by a short shaking, as described in the text. From Grandi e t a l . 19

It is interesting to observe both from Fig. 5 and from the literature 18 that the maximal solubilization does not take place at the largest w0 values, but instead at rather small ones. In the case of lipophilin discussed above, the French group found the maximal solubility at w0 = 5.6, and the same value was obtained from the myelin basic protein, another component of the myeline structure that is however soluble in water. This sur-

[19]

ENZYMES IN REVERSE MICELLES

197

prising feature is a clear indication that the water inside the water pool is not "normal" water, but acquires novel properties, including a novel solubilization power.

Characterization of Protein-Containing Reverse Micelles The experimental determination of the size and structure of the micellar aggregates which host the biopolymers is more complicated than that of the "unfilled" micelles, mostly because the system under study will generally be heterogeneous, i.e., composed by "filled" (containing the biopolymer) and "unfilled" micelles. Actually, it is rather difficult (with the exception of a few cases) to obtain a high degree of occupancy of the micelles, and most of the experimental methods for the determination of the molecular weight or size (light scattering, osmometry, photon correlation, fluorescence, etc.) will reach an average value. Since also the occupancy degree is generally not known a priori (this depend on the model for the protein uptake), the. knowledge of the molecular weight of the enzyme-containing micelles is indeed not straightforward. One needs an experimental method that is able to discriminate a small amount of filled micelles over a larger background of unfilled ones. Ultracentrifugation with a UV-monitoring device seems to be the best method, since measuring at around 280 nm, where only the protein absorbs, makes the unfilled micelles "invisible." Furthermore, by adding an appropriate chromophoric small molecule to the system, which is present in both filled and unfilled micelles, as suggested by Martinek's group, 2° one may be able to demonstrate both filled and unfilled micelles. One such sedimentation run is shown in Fig. 6A. However, it is not easy to proceed from these data to molecular weight evaluation. Sedimentation and diffusion coefficients must be extrapolated to infinite dilution (as shown in Fig. 6B) of micelles in order to avoid artifacts due to intermolecular association (as is always the case in macromolecular systems having finite virial coefficients). But this is difficult because micelles dissociate at very low surfactant concentration, and there are some conceptual problems in the definition of the partial specific volume of the sedimenting particle. Equilibrium sedimentation runs are probably best suited for determining the molecular weight, but in order to determine the dimensions (Stokes radius) one needs again diffusion coefficients, which have to be measured in independent runs, and extrapolated to c = 0. Finally the molecular weight should be determined as a function of w0 and other micellar parameters. 20 E. Martinek, A. V. Levashov, Y. L. Khmel'nitskii, N. L. Klyachko, V. Y. Chernyak, and, V. I. Berezin, Dokl. Akad. Nauk SSSR 258, 1488 (1981).

198

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

°i/-

[19]

A

OD

S, S'

18 ¸

16

/

DNP ÷ LYS

LYS

1,

/ ,

2'5

go

w6o

c A O T (mM)

FIG. 6. Ultracentrifugation experiments with AOT reverse micelles. (A) UV scanner traces of the time-dependent sedimentation profile (concentration in terms of optical density at 350 or 296 nm versus radial position) of three different micellar solutions (w0 = 7, 7, 100 m M AOT in isooctane) at 48,000 rpm and 25 °. DNP, 2,4-Dinitrophenol; Lys, lysozyme. (B) Extrapolation of the determined sedimentation coefficients (S) to infinite dilution (AOT in isooctane, w0 = 6.9) for reverse micelles with and without entrapped lysozyme: - - , without lysozyme; - . - , with lysozyme ([lysozyme]/[AOT] = 10 3). From Zampieri et al. 21a

Table I collects some typical experimental data on the molecular weight determination of filled and unfilled micelles. One notices in all cases an enlargement of molecular mass, which is larger than the sum of the two initial components (unfilled micelles and lysozyme, which has a molecular mass of 14,500 Da). The simplest way to interpret the data is shown in Fig. 7. This water shell model, based mainly on geometrical considerations and mass balance, has been elaborated into a method which permits the calculation of the dimensions of the filled micelle as a function of w0 once the molecular weight and dimensions of the protein and the unfilled micelle are known. 2~ However, on the basis of recent measurements, it appears that this first-approximation model may need s o m e r e v i s i o n . 21a 2i F. J. Bonner, R. Wolf, and P. L. Luisi, J. Solid-Phase Biochem. 5, 255 (1980). 21a G. G. Zampieri, H. J~ickle, and P. L. Luisi, J. Phys. Chem. 90, 1849 (1986).

[19]

ENZYMES IN REVERSE MICELLES

199

TABLE I MOLECULAR WEIGHTS OF LYSOZYME-CONTAINING REVERSE MICELLES a

W0

AOT (mM)

Molecular weight b

6.9 7.8 10. 11.7

100 100 50 100

60,000-168,000 c 40,000-96,000 d 91,000-122,000 d 81,000-303,000 c

o The aqueous solution is 20 m M borate buffer, pH 8.5. The overall e n z y m e concentration is in the range 50-100 /~M. All measurements are performed at room temperature. From G. Zampieri e t al. 2Ia w0 = [H20]/[AOT]. b The first figure is related to "unfilled," the • second to "filled" micelles. c F r o m equilibrium measurements. d F r o m sedimentation and diffusion measurements.

Conformation and Activity of Enzymes in Reverse Micellar Solutions

Some General Experimental Considerations The concentration of protein in the hydrocarbon micellar solution can be easily checked by spectroscopic methods, for example, by measurement of the optical density at 280 nm. Once the enzyme is solubilized in the hy~irocarbon micellar solution, its activity needs to be assayed. This can be done in the same way as for an aqueous solution of the enzyme, namely by addition of the substrate to the enzyme solution, or vice versa. Several possibilities of this operation protein

® reverse micelle

FIG, 7. E n z y m e s in reverse micelles. The water shell model.

200

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

can be used in a micellar solution. (1) The two reagents (enzyme and substrate) can be added to each other as two hydrocarbon miceUar solutions; (2) one of the two reagents can be added to the hydrocarbon micellar solution in small amounts (microliter quantities) as an aqueous buffered solution; (3) finally the substrate, when directly soluble in hydrocarbon, can be added in the pure liquid form (e.g., linoleic acid to lipoxygenase). 22 Which of the assay methods is best depends on the particular system under study. Perhaps a word of warning is necessary concerning the order of mixing. When the enzyme is not very stable in the micellar solution, it is better to add the enzyme as the last component from the aqueous stock solution, to minimize the time that the enzyme stays in the micellar solution. One such case is lysozyme which we will discuss later. The study of the enzyme properties in the micellar solution, from this point on, can be carried out exactly following the procedures used for aqueous solutions. The micellar enzyme solutions are in fact clear, and spectroscopic enzyme assays to measure activity as well as circular dichroism, absorption, fluorescence, or NMR studies can be readily performed. However, there is one conceptual difficulty that sets these systems apart from normal aqueous solutions: the definition of concentration (and therefore of the Michaelis-Menten constant Kin) for water-soluble reagents, such as hydrophilic enzymes or their water-soluble substrates, is not unambiguous. One can refer the concentration to the overall volume (ov) of the solution (hydrocarbon plus water) or to the volume of the water alone. For a micellar solution containing 1% water (v : v), and for the simple case where the guest substance is only present in the water pools, the two numbers differ by a factor of 100. Km,ovis a phenomenological parameter, i.e., it refers to the analytical concentration in the whole system4'23; Km,wpassumes a structural model, i.e., one in which water is segregated in the water pools (wp), which very rapidly exchange materials with each other. There is no reason a priori to prefer to express the concentration as overall or as water pool, but misconceptions may arise. A numerically larger Kr~ value is in fact associated with a lower affinity of the enzyme for the substrate. One should be careful at this point not to confuse a trivial mathematical transformation with physical reality. In our case, enzyme saturation (or half-saturation at Km) is obtained with the same amount of added substrate regardless of whether one considers 22 p. L. Luisi, P. L~lthi, I. Tomka, J. Prenosil, and A. Pande, Proc. 7th Enz. Eng. Conf., Ann. N.Y. Acad. Sci. 434, 549 (1984). 23 p. D. I. Fletcher, N. F. Perrins, B. H. Robinson, and C. Toprakcioglu, in "Reverse Micelles" (P. L. Luisi and B. E. Straub, eds.). Plenum, New York, 1984.

[19]

ENZYMES IN REVERSE MICELLES

201

overall or local concentration. Or in other words, if Km is numerically larger when expressed in water pool units, the substrate concentration is larger by the same factor, so that the enzyme fractional saturation is the same. Another controversial conceptual point, which is very important for the experimentalist, is the determination of the pH of the miceUar solutions. The argument has been much debated, z4,25 and no general agreement has been reached. We believe that the situation can be summarized as follows. In principle, there is no simple direct way either to define or to determine the pH in the water pool, mainly because the water of the water pool is a novel solvent, for which no calibration has been as yet offered. However, it can be accepted with fair reliability that the pH of the water pool is close to that of the stock buffered aqueous solution used to prepare the water pool of the reverse micelles, particularly at w0 values larger than 10-15. A useful experimental check is the determination of the apparent acidity by using the 3tp NMR of phosphate buffers in the water pool. 25,1° After this discussion of the main operational problems encountered in the field, we can now present an analysis of the conformation and activity of enzymes solubilized in apolar solvent via reverse micelles. We will analyze the various classes of enzymes, trying to emphasize in each case the major changes in activity and conformation brought about by the micellar environment with respect to bulk water. Most of the data presently available in the literature about proteins in reverse micellar systems are presented in Tables II-V. In particular, Table II reports data about proteases, Table III about hydrolases other than proteases, Table IV about oxidoreductases and related enzymes, and Table V about nonenzymatic proteins. A collection of kinetic parameters (Km and kcat) in reverse miceUes, compared with aqueous solutions, is offered in Table VI. Proteases

As apparent from Table 1I, 1'9'16'27-33 the most widely studied enzyme in hydrocarbon micellar solutions is a-chymotrypsin. The first report de24 O. A. El Seoud, in "Reverse Micelles" (P. L. Luisi and B. E. Straub, eds,), p. 81. Plenum, New York, 1984. 25 R. E. Smith and P. L. Luisi, Heir. Chim. Acta 63, 2302 (1980). z6 p. L. Luisi and C, Laane, Trends Biotech. 4, 153 (1986). 27 p. L. Luisi, F. Henninger, M. Joppich, A. Dossena, and G. Casmati, Biochem. Biophys. Res. Commun. 74, 1384 (1977). 18 K. Martinek, A. V. Levashov, N. L. Klyachko, V. I. Pantin, and I. V, Berezin, Biochim. Biophys. Acta 657, 277 (1981). 29 F. M. Menger and K. Yamada, J. Am. Chem. Soc. 101, 6731 (1979).

202

IMMOBILIZED

ENZYMES/CELLS

o

t-

IN ORGANIC

•- ¢ , , ~

~

•. ~

~

,~

'~.~

:~

.~

~

:~

~

o

.~ ,-.,

[19]

SYNTHESIS

•

~ .'~

~ . ~ . ~ :~

:~

~

'~ ~,

~

~

z

g~ o

o

~-

¢h

~,

~a

<

>

z ua 0

E 0

o

¢11

.~.

4~

• t " ,.~ ~'~ P"-. O

<

o

r.~

[19]

ENZYMES IN REVERSE MICELLES

eq

•

~.£oo

e,I

203

.=.

~

b

~='~

o

N

~

z

;~

z

,.a

~'~

<<

~

~

0

Z

0

°~

0 o

0

b~ ~>:'&

o;~oo

,:5

0 ..=

2

'S

k~ a=l

e-

E

•~

'a 0

E

I=

.-.=

0

.z

204

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

scribed the protein dissolved in cyclohexane via the cationic surfactant methyltrioctylammonium chloride.16 The CD spectrum is very similar to that in water, which was taken as evidence that the conformation of the main chain was not too different in the two systems. The activity of the enzyme, however, had decreased to zero. It was not established whether enzyme-entrapping reverse micelles are formed in this case. Two other proteases, trypsin and pepsin, were investigated in methyltrioctylammonium chloride. 16As in the case of a-chymotrypsin, the UV spectra show a small red shift, the fluorescence spectra are blue-shifted, and the quantum yields are 2-5 times higher than in water. The CD spectrum of trypsin in the far-UV shows only slightly reduced ellipticities compared to water, while the near-UV CD has almost disappeared. Again, no activity for either enzyme could be determined. The authors state that only slight conformational changes take place in the micellar solution because the differences in the far-UV CD spectra are only small compared to water. The changed fluorescence and near-UV CD spectra would only be due to small perturbations of the aromatic chromophores. The case is different for the AOT/hydrocarbon/water system. Four research groups that have studied a-chymotrypsin in this type of reverse micelles report that the enzyme remains active. 9,28-3°At a pH corresponding to the optimal pH in water, Menger and Yamada found initially almost no activity in reverse micelles. 29 It became clear, however, that this apparent decrease in enzyme activity was due to a shift of the pH profile of 1.5 pH units to more alkaline values in the AOT micelles. In fact, at the optimum pH and w0 = 25, kca t w a s tWO times higher than in water for the substrate N-acetyl-t-tryptophan methyl ester. The enhancement of activity was confirmed by other workers 9 in the low w0 region, using a different substrate. Interestingly, at low w0 the helix content seems to be higher than in bulk water solution (Fig. 8). (This conclusion is inferred from CD data on the basis of the calculation procedure of Chen e t al. 34) Since in aqueous solution, a-chymotrypsin exists in equilibrium between two forms with different activities and different CD properties, 35 it has been proposed 30 C. Kumar and D. Balasubramanian, in "Solution Behavior of Surfactants" (K. L. Mittal and E. J. Fendler, eds.), Vol. 2, p. 1207. Plenum, New York, 1982. 31 p. D. I. Fletscher, R. B. Freedman, J. Mead, C. Oldfield, and B. H. Robinson, Colloids Surface 10, 193 (1984). 32 p. Douzou, E. Keh, and C. Balny, Proc. Natl. Acad. Sci. U.S.A. 76, 681 (1979). 33 p. L. Luisi and R. Wolf, in "Solution Behavior of Surfactants" (K. L. Mittal and E, J. Fendler, eds.), Vol. 2, p. 887. Plenum, New York, 1982. 34 Y.-H. Chen, J. T. Yang, and H. M. Martinez, Biochemistry 11, 4120 (1972). 35 A. R. Fersht, J. Mol. Biol. 64, 497 (1972).

[19]

205

ENZYMES IN REVERSE MICELLES ~

m

40 :~

A

/

.,. ~; . . f-/\.\

'.

/z.'

/'/ I: /

_ ¢/~/....."...,., ~'/./.,."

-4

,,,f:(.. . . . .

-40

.',

]-

,:2'

..: ,.,. :

-80

/b"

-6

~;4 //:

-8 210

230

250 ~(nm)

270

290

310

FIG. 8. CD s p e c t r a of ~ - c h y m o t r y p s i n in the far (a) and near (b) UV region (overall c o n c e n t r a t i o n 11/~M) in w a t e r and A O T / i s o o c t a n e at different w0: , water, pH 7.9; --water, pH 9.8; - . - w0 = 22.5; . . . . w0 = 18; ... w0 -- 13.5. F r o m Barbaric and Luisi. 9

that in the miceUar environment the same holds, but instead the conformational equilibrium is shifted toward the more active form. 9 Kumar and Balasubramanian investigated a-chymotrypsin in several other micellar systems 3° at pH 7.4. They found the best activity in reverse micelles of Brij 56 and sodium dodecyl sulfate (85-87% of activity in water). The CD spectra obtained in these solutions were almost identical to those in water. In Triton X-100 the enzyme was only half as active as in the aqueous solution, and in CTAB and sodium laurate, it had lost its activity. Martinek et al. studied trypsin in the cationic surfactant CTAB in chloroform/octane solution. 2s In this system, the enzyme remains active. In AOT, however, at w0 = 25, they found a much lower activity for the enzyme, by using the same substrate. The authors explain the difference by electrostatic interactions between the micelle wall and the substrate. Spectroscopic data are not presented. While the interpretation offered by Martinek is plausible, it is also important to recall the results of Kumar and Balasubramanian 30 who found that a-chymotrypsin shows a modified conformation in CTAB and, correspondingly, no activity. This would mean that not the nature of the substrate, but a change in the enzyme, may cause the loss of activity. In contrast to the results obtained by Martinek et al. 28 Douzou and co-workers found that trypsin retains its

206

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

activity in AOT/heptane. 32 A broad temperature range (-23 ° to +30 °) was investigated, thus using the micellar system for cryoenzymology. In fact, reverse micelles appear to be useful for low-temperature studies because under certain conditions the entrapped water does not freeze well below 0°. Other Hydrolases As shown in Table I I I , 26'28'36-39 ribonuclease was first investigated in AOT/octane. 36 The UV absorption spectrum is practically identical to that of the aqueous solution. Also the CD spectra both in the far- and near-UV range of the aqueous and the micellar solutions are almost superimposable, indicating that no large conformational change takes place (w0 was - 18). A few data exist for pyrophosphatase in reverse micelles of Brij 56 in cyclohexanefl8,4° The rate of hydrolysis of pyrophosphate was studied as function of w0 and it was found that at the optimum w0 = I0 the enzyme had the same activity as in water. Another interesting aspect of enzymatic catalysis in reverse micelles is the use of water-insoluble substrates as in the case of lipids. Misiorowski and Wells 38used phospholipase in ether/methanol solutions of phosphatidylserine. Malakhova et al. studied pancreatic lipase in AOT reverse micelles, 37 thus avoiding the problem of having a fixed substrate concentration, as in the case of the phosphatidylcholine micelles. 38 The optimum rate for the hydrolysis of triolein was found at w0 = 12. A short-chain substrate, tributyrin, was cleaved by an order of magnitude less effectively than the long-chain substrate triolein, The use of reverse micelles in this case made it possible for the first time to compare the substrate specificity of pancreatic lipase under conditions in which the adsorption step of the enzyme on the surface of the substrate agglomerates is absent. Finally, let us consider the case of lysozyme in reverse micelles. A first investigation was made by Grandi et al.19 The far-UV CD spectra of this enzyme in AOT/isooctane showed higher ellipticities than in water. In the near-UV range, the spectra changed from positive values to more 36 R. Wolf and P. L. Luisi, Biochem. Biophys. Res. Commun. 89, 209 (1979). 37 E. A. Malakhova, B. I. Kurganov, A. V. Levashov, I. V. Berezin, and K. Martinek, DokL Akad. Nauk SSSR 270, 474 (1983). 37a D. Han and J. S. Rhee, Biotech. Bioeng. 27, 1250 (1986). 3s R. L. Misiorowski and M. W. Wells, Biochemistry 13, 4921 (1974). 39 A. Oshima, H. Narita, and M. Kito, J. Biochem. (Tokyo) 93, 1421 (1983). 4o A. V. Levashov, N. L. Klyachko, V. I. Pantin, Y. L. Khmel'nitskii, and K. Martinek, Bioorg. Khim. 6, 929 (1980).

[19]

E N Z Y M E S IN R E V E R S E M I C E L L E S

~. ,,,,~

o

~

.-

q~

~,.~

._~ ~

'~u

207

~

~

I~

= <

.

o

~..~ 0

0

=~o

.-1

o

< ,,.I

o

e~

nl

0

=2

c-

r..q ~ O ~q

l,,q i.i,.! > L~

<

l..q 0

~ :~

I

o2

d <

m

~

~

© <

0

~.

~.~

© <

<

~

<

•

208

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

or less negative ellipticities, depending on the w0 value. Since the UV absorption spectra in the near-UV range were the same in water and in the micellar solution, these CD changes could be safely attributed to conformational changes. Despite this, the enzyme's activity was found to be comparable to that in water. The situation was then interesting: how can an enzyme function with a markedly changed conformation? This question has recently been clarified ~° by carrying out CD spectra in the presence of the inhibitors NAG (N-acetylglucosamine) and tri-NAG (Fig. 9) from which it became clear that the CD changes from water to micelles corresponded to enzyme denaturation. The reason why the original investigation of Grandi et al. ~9failed to recognize this is that chitins were used as substrate, which are not readily soluble in the micellar hydrocarbon system, and the assay was then performed by adding the enzyme last to the micellar solution. In this way, it had no time to denature. However, the final conclusion ever is that when the enzyme works, it does so with the right conformation. In CTAB and the neutral surfactant tetraethylene glycol dodecyl ether, lysozyme exhibits almost water-like CD spectra. The activity is comparable to that in water, and a denaturation process like in AOT does not occur.l° Oxidoreductases

Spectroscopic and activity data are available for horse liver alcohol dehydrogenase (LADH) in the system isooctane/AOT/water 41m (see Table IV16'4°-51). The UV and CD spectra of the aqueous and micelle systems are very similar/~ The CD spectra show some changes with respect to water, but the main conformational features seem to remain similar in the two systems. The fluorescence spectra show a decrease in quantum yield

41 p. Meier and P. L. Luisi, J. Solid-Phase Biochem. 5, 269 (1980). 42 K. Martinek, L. Y. Khmel'nitskii, A. V. Levashov, and I. V. Berezin, Dokl. Akad. Nauk S S S R 263, 737 (1982). 43 K. Martinek, A. V. Levashov, Y. L. Khmel'nitskii, N. L. Klyachko, and I. V. Berezin, Science 218, 889 0982). " A. N. Erjomin, S. A. Usanov, and D. J. Metelitza, Vestsi Akad. Navuk. BSSR, Set. Khim. Navuk. 3, 65 0982), 4~ A. N. Erjomin and D. J. Metelitza, Biochim. Biophys. Acta 732, 377 (1983). 46 p. Douzou, P. Debey, and F. Franks, Biochim. Biophys. Acta 523, I (1978). 47 M. P. Pileni, Chem. Phys. Left. 81, 603 (1981). 4s R. Hilhorst, C. Laane, and C. Veeger, Proc. Natl. Acad. Sci. U.S.A. 79, 3927 (1982). 49 R. Hilhorst, C. Laane, and C. Veeger, FEBS Lett. 159, 31 (1983). 5o p. Meier, Ph.D. Dissertation No. 7222, ETH (Swiss Federal Institute of Technology), Zurich (1983). 51 C. Gitler and M. Montal, FEBS Lett. 28, 329 (1972).

[19]

ENZYMES IN REVERSE MICELLES

209

1012

I//

/N / \, / \.j \,. f, i/

\/

./ / I "D

It"

~- ">.--

U

¢/s\

400

II

\ ~ /

-200

260

280

300

a nm

FIG. 9. CD spectra of the NAG-lysozyme complexes in water ( ) and in AOT/ isooctane at w0 = 3 (---) and w0 = 6 (-.-), pHst 7.0. (PHst is the pH of the aqueous stock solution of the protein used to prepare the protein-containing micellar solution.) 1°

and a blue shift of several nanometers. This may be due to a more pronounced quenching of the exposed Trp-15 by water molecules in the miceUar system. The activity of the enzyme in water and in micelles is not very different for the respective optimum conditions, despite some changes in the spectroscopic properties. Martinek et al. showed that LADH in AOT/octane changes its substrate specificity. 42 They investigated the oxidation of alcohols with the coenzyme N A D + and studied the dependence of the number of carbon atoms of the alcohol molecule on the enzyme activity. In water, the activity of LADH is highest at n = 8, i.e., octanol is the best substrate. In

210

IMMOBILIZED

ENZYMES/CELLS

~o~

•

N

[19]

IN ORGANIC SYNTHESIS

< o

e~ H

<

<

m

>

~

2

< ©

>

~

Z

m

< M 0

m ~

m

z

N Z

..= o~

[19]

ENZYMES IN REVERSE MICELLES

.~.,

211

L~ ,~ o..~

.. 6 ~,o- ~. ~

~"~ "~ ~ ~'~

,.~

oo

"-4

E --

--

r~

°~

,.-4

~

d

o u

O

o?

E E ~q ,..o~

Q o

o~

212

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[19]

AOT/octane, the maximum was found at n = 4 regardless of Wo. The authors explain this changed substrate specificity in terms of the changed local concentration of the alcohols in the micelle. A 20-fold velocity enhancement for the oxidation of pyrogallol with peroxidase was found in AOT/octane with respect to the water solution. 43 The optimum water content was at Wo = 10. No substrate inhibition was found in the micellar system, whereas in aqueous solution peroxidase is inhibited by excess of substrate. 28 The reaction catalyzed by lipoxygenase represents another case of reactivity with a water-insoluble substrate.16,40,50 The hydroperoxidation of linoleic acid could be followed directly by UV spectroscopy because clear solutions were obtained in reverse AOT micellar solutions. The reaction proceeded only half as fast as in the aqueous solution and the enzyme-substrate affinity was smaller than in water, as shown by the higher Km value (see Table VI). As Luisi et al. explain, 16 it might be possible that the optimum pH was not reached in the micellar solution because of the instability of aqueous enzyme stock solutions with pH values higher than 10. A highly organized system for an efficient coupling between hydrogenase and a photochemical system that produces reducing equivalents and protons necessary for hydrogenase action was obtained by entrapping hydrogenase in reverse CTAB micelles. 26,48The stability of hydrogenase was higher in reverse micelles than in the aqueous system. Furthermore, the same group has described a multienzyme system for the site-specific enzymatic reduction of apolar, poorly water-soluble ketosteroids that could also be enclosed in CTAB reverse micelles, z6:9 Table V 18,3°,52-55 summarizes reverse micellar systems using miscellaneous nonenzymatic proteins. The kinetic parameters k~at and Km for several enzymes in reverse micellar systems are presented in Table WI. 9'25'26'28'29'32'36'41'43'50'56'57 Concluding Remarks The use of reverse micelles is a general method to solubilize proteins in aprotic media. Although with all probability proteins are not really in 52 A. Darszon, R. Strasser, and M. Montal, Biochemistry 18, 5205 (1979). s3 A. Darszon, L. Blair, and M. Montal, FEBS Lett. 107, 213 (1979). 54 M. Sch6nfeld, M. Montal, and G. Feher, Biochemistry 19, 1535 (1980). 55 j. Wirz and J. P. Rosenbusch, in "Reverse Micelles" (P. L. Luisi and B. Stranb, eds.), p. 231. Plenum, New York, 1984. 56 p. D. I. Fletcher, A. M. Howe, B. H. Robinson, and D. C. Steytler, in "Reverse Micelles" (P. L. Luisi and B. E. Straub, eds.), p. 69. Plenum, New York, 1984. 57 O. A. E1 Seoud, in "Reverse Micelles" (P. L. Luisi and B. E. Straub, eds.), p. 81. Plenum, New York, 1984.

[19]

ENZYMES IN REVERSE MICELLES

213 0o

E

;>

.~# 20 " , ~

<

~ 0

~

._~ .~

0

tD

X

r,.) 0

0

0

O

0 Z < .4 .4

0

¢!1 [-

...1 u~

;> ;~

m

~

i=

,~1

.~.

~

z Ich

m

I

-E © .<

0 <

© <

,.Q ¢)

& 0

0 0

#

'~

~

~.~

214

IMMOBILIZED

ENZYMES/CELLS

~,~

~x

~

IN ORGANIC

x

,~

X

[19]

SYNTHESIS

,~,

X

X 0

S~ ,z, o~

J

•

x

~

~

T

~

T

~

e,~

X

°T

o~

oo

T

x

~

~

x

~

~,m

X

S

o

o

#£~°S

0

o

~' 2 ~

u

~J > ,,i

z

o

o

S

X

--

--

X

--

--

--

X

o~

X

X

~

X

X

X

r.-..-':..

O L

m [... N

"~.o f~

--

,,,~ -..~. X

"js~

<

u~ Z

M

T

'~

"d

~t, '-u.-

= ~ ~.-,s r ;

-a tJ

8~

,<

[19]

ENZYMES IN REVERSE MICELLES

215

the organic solvent, but rather in water microdroplets (at least for hydrophilic enzymes), the micellar system may contain only 0.3% water, and the system shows properties which are typical of a bulk apolar solution: for example, hydrocarbon-soluble substrates can be directly added to the bulk micellar solution and are accepted by the enzyme, as shown by the examples of lipoxygenase 16,4°and LDH 41 and by the complex reduction of the ketosteroid accomplished by Laane and co-workers. 49 Actually, the hydrocarbon micellar system of the enzyme is a delicate balance of properties both of the aqueous protein solutions and of hydrocarbon solutions, and in the exploitation of this balance resides the whole interest in the field, both from the biotechnological and the academic point of view. Another point of interest, which we have not dwelled on in this paper, is the fact that nucleic acids, large plasmids, and bacterial cells can be solubilized in hydrocarbon micellar solutions) 8,59Applications to biotechnological and basic research can also be explored. Although the reverse micellar system, and particularly the hydrocarbon/AOT/water system, is transparent in the far UV, only very few groups have reported spectroscopic/conformation studies. This is unfortunate; in addition to the importance of the conformational aspects per se, the lack of information about conformation makes the comparison between the results of different research groups rather difficult. For example, the discrepancy in the activity of proteases (a-chymotrypsin in particular) may well be due to partial unfolding under certain conditions. The discrepancy of results from different laboratories may be partly due to differences in the surfactant purity. This is particularly so for AOT. It would be useful for future work in the field if each group would characterize AOT samples by at least the following three parameters: absorption at 280 nm, a titration profile as in Fig. 2, and a conductibility measurement as a criterium for the salt content. These three figures would at least permit the comparison of AOT preparations from different laboratories, which at present is still rather difficult. In addition to purification procedures, there are several other areas to be explored. One is the structure of the micellar aggregate; the "water shell" model 21 is only one of the possible models and its generality is open to debate. Data about the nature and properties of water in the water pool indicate that the relationship between water of the water pool and the guest enzyme is, to say at least, surprising. Maximal activity is generally not present at the largest water content, but at w0 values around 10-13. Also, the optimal solubility of proteins in reverse micelles occurs at relas8 V. E. Imre and P. L. Luisi, Biochem. Biophys. Res. Commun. 107, S.38 (1982). 59 G. Hating, F. MensdOrfer, and P. L. Luisi, Biochem. Biophys. Res. Commun. 127, 911 (1985).

216

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[20]

tively low w0 values. Finally, the phenomenon of "superactivity" of certain hydrolases, such as a-chymotrypsin, ribonuclease, and lysozyme, is present only at very small water content. Under this condition the conformational perturbations are the largest, and the micelle should be the smallest, i.e., the least hospitable. This is an area which needs clarification, and which can have relevance for enzymology as a whole.

[20] D e s i g n o f R e v e r s e d M i c e U a r M e d i a for t h e E n z y m a t i c Synthesis of Apolar Compounds

By

COLJA L A A N E , R I E T HILHORST,

and

CEES VEEGER

Reversed micelles are tiny oases of water stabilized in an organic solvent with the aid of surfactants. It has been shown that enzymes entrapped in these pools are capable of synthesizing apolar compounds such as steroids,l-3 oxidized polyunsaturated fatty a c i d s / a n d peptides. 4 In this paper we will focus on the reversed micellar enzymology of 20fl-hydroxysteroid dehydrogenase (cortisone reductase; 17,20/3,21-trihydroxysteroid: NAD ÷ oxidoreductase; EC 1.1.1.53), which catalyzes the reduction of various ketosteroids to the corresponding 20fl-hydroxysteroids at the expense of NADH. Preparation of Cetyltrimethylammonium Bromide (CTAB) Reversed Micellar Media

Principle. A reversed micellar medium consists of at least three components: an aqueous solution, a water-immiscible organic solvent, and a surfactant. The mechanism leading to the formation of reversed micelles is the tendency to extend the interfacial area until the concentrations of surfactants are sufficiently low to achieve a nonnegative interfacial tension. 5 By far the best studied three-component reversed micellar medium contains AOT [bis(2-ethylhexyl)sodium sulfosuccinate] as surfactant and R. Hilhorst, C. Laane, and C. Veeger, FEBS Lett. 159, 225 (1983). 2 R. Hilhorst, R. Spruijt, C. Laane, and C. Veeger, Eur. J. Biochem. 14,t, 459 (1984). 3 R. Hilhorst, Ph.D. Thesis, Agricultural University, Wageningen, The Netherlands (1984). 4 p. L. Luisi, P. Liithi, I. Tomka, J. Prenosil, and A. Pande, Enzyme Eng., in press (1987). 5 j. Th. G. Overbeek, Faraday Discuss. Chem. Soc. 65, 7 (1978).

METHODSIN ENZYMOLOGY,VOL. 136

Copyright© 1987by AcademicPress, Inc. All rightsof reproductionin any formreserved.

216

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[20]

tively low w0 values. Finally, the phenomenon of "superactivity" of certain hydrolases, such as a-chymotrypsin, ribonuclease, and lysozyme, is present only at very small water content. Under this condition the conformational perturbations are the largest, and the micelle should be the smallest, i.e., the least hospitable. This is an area which needs clarification, and which can have relevance for enzymology as a whole.

[20] D e s i g n o f R e v e r s e d M i c e U a r M e d i a for t h e E n z y m a t i c Synthesis of Apolar Compounds

By

COLJA L A A N E , R I E T HILHORST,

and

CEES VEEGER

Reversed micelles are tiny oases of water stabilized in an organic solvent with the aid of surfactants. It has been shown that enzymes entrapped in these pools are capable of synthesizing apolar compounds such as steroids,l-3 oxidized polyunsaturated fatty a c i d s / a n d peptides. 4 In this paper we will focus on the reversed micellar enzymology of 20fl-hydroxysteroid dehydrogenase (cortisone reductase; 17,20/3,21-trihydroxysteroid: NAD ÷ oxidoreductase; EC 1.1.1.53), which catalyzes the reduction of various ketosteroids to the corresponding 20fl-hydroxysteroids at the expense of NADH. Preparation of Cetyltrimethylammonium Bromide (CTAB) Reversed Micellar Media

Principle. A reversed micellar medium consists of at least three components: an aqueous solution, a water-immiscible organic solvent, and a surfactant. The mechanism leading to the formation of reversed micelles is the tendency to extend the interfacial area until the concentrations of surfactants are sufficiently low to achieve a nonnegative interfacial tension. 5 By far the best studied three-component reversed micellar medium contains AOT [bis(2-ethylhexyl)sodium sulfosuccinate] as surfactant and R. Hilhorst, C. Laane, and C. Veeger, FEBS Lett. 159, 225 (1983). 2 R. Hilhorst, R. Spruijt, C. Laane, and C. Veeger, Eur. J. Biochem. 14,t, 459 (1984). 3 R. Hilhorst, Ph.D. Thesis, Agricultural University, Wageningen, The Netherlands (1984). 4 p. L. Luisi, P. Liithi, I. Tomka, J. Prenosil, and A. Pande, Enzyme Eng., in press (1987). 5 j. Th. G. Overbeek, Faraday Discuss. Chem. Soc. 65, 7 (1978).

METHODSIN ENZYMOLOGY,VOL. 136

Copyright© 1987by AcademicPress, Inc. All rightsof reproductionin any formreserved.

[20]

ENZYMATIC REACTIONS IN REVERSED MICELLES

217

isooctane or hexane as oil phase. For preparation and characterization, see Luisi and Wolf 6 and Eicke. 7 The disadvantage of a three-component system is that the properties of the interphase are determined mainly by the surfactant and are therefore difficult to vary. However, by adding a fourth component, usually a cosurfactant that partitions between the interphase and the continuous phase, it is possible to vary the properties of both phases with respect to each other. Suitable cosurfactants are aliphatic alcohols ranging from butanol up to decanol. To our knowledge any organic solvent or mixtures of solvents can be used provided it is immiscible with water. Our standard four-component system consists of an aqueous solution, the cationic surfactant cetyltrimethylammonium bromide (CTAB), hexanol as cosurfactant, and octane as organic solvent. The concentration of CTAB is fixed at 0.2 M in octane. The concentration of water and hexanol are expressed as weight percentage with respect to the CTAB-octane solution. Another often used expression for the water content in a reversed micellar medium is w0, which is defined as the molar ratio of water to CTAB. In general, an optically clear reversed micellar solution is formed within an hour after mixing at concentrations of 9-30% alcohol and 435% of an aqueous solution both with respect to 0.2 M CTAB in organic solvent. Procedure. A reversed micellar medium consisting of an aqueous solution of HEPES (N-2-hydroxyethylpiperazine N'-ethanosulfonic acid), 10% hexanol, and 0.2 M CTAB in octane at a w0 = l0 contains 1.46 g CTAB; 20 ml octane; 1.8 ml hexanol; 0.72 m150 mM HEPES, pH 7.6. The brand of CTAB is important. Without purification enzyme activity with 20fl-hydroxysteroid dehydrogenase is highest with CTAB purchased from Serva or Merck. The sequence in which the components are mixed is not important. Furthermore, to obtain equilibrium as rapid as possible it is advisable to stir the solution vigorously after each addition. Determination of a HEPES-Hexanol-CTAB-Octane Phase Diagram

Principle. To gain insight into the way the composition of reversed micelles influences the activity of enzymes in their water pool, a physical picture of the system such as a phase diagram is desired. Since the CTAB 6 p. L. Luisi and R. Wolf, in "Solution Behaviour of Surfactants: Theoretical and Applied Aspects" (E. J. Fendler and K. L. Mittal, eds.), pp. 887. Plenum, New York, 1981. 7 H.-F. Eicke, Top. Curr. Chem. 87, 85 (1980).

218

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[20]

to octane ratio is kept constant, the quaternary phase diagram becomes a pseudo-ternary phase diagram. This implies that two components can be varied freely, whereas the ratio of the other two is kept constant (see Fig. I). Procedure. A series of 2 ml solutions of 0.2 M CTAB in octane containing varying amounts of 50 mM HEPES (pH 7.6) are prepared in Bausch and Lomb 0.5-in. colorimeter test tubes having a rubber stopper. Hexanol is titrated into these tubes with a microsyringe under vigorous mixing with a Vortex mixer at 25 °, until just clear. The titration is contin-

HEXANOL 0,,100

CONTINUOU

~...

• 10% HEXANOL

100 / 0

~ o 5'0

WATER ~

~*%100 ETAB/OETANE

FIG. 1. Pseudo-ternary phase diagram for the system CTAB, octane, hexanol, and 50 mM HEPES, pH 7.6. The phase diagram was determined as described in the text. All quantities are expressed as percentage of total weight. The concentration of CTAB was 9.4% (w/w) with respect to octane (0.2 M). Keeping this amount constant allowed determination of 20fl-hydroxysteroid dehydrogenase activities in this plane of the phase diagram while varying the water and hexanol content. The shaded area represents the region where equilibrium in the presence of prednisone or progesterone, NADH, and enzyme was obtained rapidly, i.e. within mixing time. In this area initial enzyme activities were measured according to procedure 1 (see text). Furthermore the lines 4.4% (w/w) H20 (w0 = 10) and 10% (w/w) hexanol with respect to CTAB-octane are drawn. Variation of the concentration along these lines allows the largest change of the water and hexanol concentration in the shaded area where enzyme activity was measured. (©), Area where reversed micelles are formed within 4 hr.

[20l

ENZYMATIC REACTIONS IN REVERSED MICELLES

219

ued until the solution becomes turbid again. At relatively low water and hexanol contents (see right corner of the triangle in Fig. 1) the transition from a clear to a turbid solution is very sharp and quick, but at higher water and hexanol contents the time needed to establish equilibrium can be several hours. Furthermore, gradual transitions occur due to the fact that opalescent solutions are obtained at higher water content. In case that no clear-cut transition is observed by eye, a refractometer or any kind of fluorometer can be very helpful in determining phase transitions. Similarly, 2 ml solutions of 0.2 M CTAB in octane containing varying amounts of hexanol are titrated with 50 mM HEPES (pH 7.6) until just clear and subsequently until turbid. When all transition points thus obtained are expressed in weight percentages of total weight the phase diagram shown in Fig. 1 is obtained. In the bootlike area, at relatively low water and hexanol content, reversed micelles are formed within 4 hr after mixing. Criteria for the presence of reversed micelles are that the solution is optically transparent and that the ohmic resistance of the solution is high (> 106 fl/cm). Entrapment of Enzymes in Reversed Micelles According to Luisi et al., 6 Martinek et al., 8 and Menger and Yamada, 9 there are essentially three techniques that can be used to solubilize enzymes into reversed micellar media (Fig. 2). Procedure.

1. A small amount (few % v/v) of a concentrated aqueous stock solution of enzyme is added by a microsyringe under vigorous stirring to the reversed micellar medium (Fig. 2A). In general the final overall enzyme concentration is within the range 1-20/zM, which is well below the actual micelle concentration in most reversed micellar media studied so far. For a 0.1 M AOT in heptane solution the micelle concentration is approximately 0.5 mM at w0 = 20, l° and for a 0.2 M CTAB in octane medium the concentration is within the range 1-4 mM (Laane, unpublished results). Hence most reversed micelles are empty. This technique is the most frequently used since it is rapid and extremely simple. Instead of mechanical agitation the medium can also be sonicated) ~ We prefer, however, mechanical agitation with a Vortex mixer. 8 K. Martinek, A. V. Levashov, N. L. Klyachko, and I. V. Berezin, Dokl. Acad. Nauk. USSR 236, 920 (1977); Dokl. Biochem. (Engl. Transl.) 236, 951 (1977). 9 F. M, Menger and K. Yamada, J. Am. Chem. Soc. 101, 6731 (1979). 10 p. D. I. Fletcher and B. H. Robinson, Ber. Bunsenges. Phys. Chem. 85, 863 (1981). H p. Douzou, E. Keh, and C. Balny, Proc. Natl. Acad. Sci. U.S.A. 76, 681 (1979).

220

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

A

B

[20]

c

Flo. 2. The three techniques to solubilize a protein in a reversed micellar medium. (A) Injection technique, by which a few microliters of the buffered stock solution of the protein are added to the surfactant solution in hydrocarbon or to a solution that already contains water. (B) Extraction from the solid state, in which the protein powder is added to a solution of hydrocarbon containing the surfactant and water. (C) Phase transfer technique in which the hydrocarbon solution of the surfactant is in contact with an aqueous solution containing the protein. From Lnisi and Wolf. 6

A few minor disadvantages of this technique are that it is difficult to encapsulate high concentrations of enzyme in the reversed micellar solution and that the enzyme may be inactivated slightly due to a "concentration shock" during the injection. Furthermore, one has to be aware of the possibility that even a few hours after injection the solution may be thermodynamically unstable due to microscopic structural rearrangements. These problems are generally overcome by the second technique. 2. The enzyme is added in the solid state as a fine powder to a complete reversed micellar solution (Fig. 2B). Subsequently the protein is dissolved by vigorous shaking or sonication. 9 This procedure takes much more time than the first one, and is reserved mainly for relatively stable and hydrophobic proteins. When it is necessary to obtain highly concentrated solutions of protein, the dry protein is added in excess and the portion remaining undissolved after a few hours is removed by centrifugation or filtration. As was shown by Martinek and co-workers ~2the solubility of a-chymotrypsin in 0. l M AOT micelles reached a maximum of - 1 mM which is probably more than the micellar concentration in this medium. Hence all reversed micelles will contain at least one protein. 3. The enzyme is solubilized in the water pools of the micellar medium by a phase transfer from a bulk aqueous solution to the reversed ~2A. V. Levashov, Yu. Khmelnitski, N. L. Klyachko, and K. Martinek, in "Surfactants in Solution" (K. L. Mittal and B. Lindman, eds.), Vol. 2, p. 1069 (1984).

[20]

ENZYMATIC REACTIONS IN REVERSED M1CELLES

221

micellar solution (Fig. 2C). This technique may be useful for studying the kinetics and mechanism of solubilization. 6 At present it is not possible to give general guidelines for the solubilization of enzymes in reversed micelles. Neither is it possible to predict whether an enzyme will be active in such media. Too many unpredictable factors such as the local pH around the enzyme and electrostatic and hydrophobic interaction between the enzyme and its microenvironment determine the solubility and the activity of a given protein. Likewise, the temperature is a factor that should be considered. For example the solubility of liver alcohol dehydrogenase in AOT-isooctane reversed micelles increases substantially when the temperature is raised from 22 to 300. 6 Furthermore, one should be aware of the fact that some proteins containing a noncovalently bound cofactor can loose this cofactor upon solubilization. For example, flavodoxin from Megasphaera elsdenii loses its FMN upon solubilization in CTAB-hexanol-octane reversed micelles, but not in an AOT-isooctane medium (A. J. W. G. Visser, unpublished results). To date, the solubilization of proteins in reversed micellar media is in the "trial and error" stage. Yet, the general consensus seems to be that for every enzyme it is possible to design a reversed micellar medium in which the enzyme is active.

Procedure for 20fl-Hydroxysteroid Dehydrogenase Encapsulation. Prior to use 20fl-hydroxysteroid dehydrogenase from Boehringer-Mannheim is diluted to a concentration of - 2 mg protein/ml and dialyzed against 7 mM HEPES (pH 7.6) to remove ammonium sulfate. According to technique 1 the enzyme is then added by microsyringe to the standard CTAB reversed micellar medium.

Determination of Enzyme Activity in Reversed Micelles

Principle. Since reversed micellar solutions are optically transparent, enzymatic reactions can be followed using conventional spectroscopic techniques. Alternatively, enzyme activities can be followed under these conditions with a common 02 electrode, standardized by adding known amounts of Oa to the micellar system, in the case of oxidases, monooxygenases, and dioxygenases (Laane, unpublished results). Another powerful technique is HPLC, in particular for apolar compounds such as steroids. 1-3 Procedures for 20fl-Hydroxysteroid Dehydrogenase. 20fl-Hydroxysteroid dehydrogenase requires both NADH and a 20-ketosteroid for catalysis. Hence, its activity can be followed spectroscopically by measur-

222

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[20]

ing the rate of NADH consumption at 340 nm or by following the rate of steroid conversion by HPLC. Procedure 1. A dialyzed stock solution of 20fl-hydroxysteroid dehydrogenase is prepared as described above. Freshly prepared 50 mM solution of steroid in chloroform, 30/zl, is pipetted into a Bausch and Lomb 0.5-in. test tube. After evaporation of the chloroform 1.5 ml of a H E P E S hexanol-CTAB-octane micellar solution is added together with 5 ttl 5 mM N A D H in 50 mM HEPES (pH 7.6). While mixing the reaction is started by the addition of 5/zl stock solution of 20fl-hydroxysteroid dehydrogenase. Subsequently the reaction is monitored at 340 vs 380 nm using any kind of spectrophotometer. Procedure 2. At time intervals 25-/xl aliquots from a complete reaction mixture (see procedure 1) is added to 100/zl acetonitrile in a 1-ml Eppendorf centrifuge tube. This step is required to stop the reaction and to precipitate CTAB specifically, because CTAB interferes with the detection procedure. After freezing for about 20 min at - 2 0 ° the samples are centrifugated for 2 min at 12,000 g in an Eppendorf centrifuge. Under these circumstances most steroids remain in solution and can be analyzed by HPLC equipped with a Phase Sep Spherisorb 10/~ ODS column. The steroids can be detected in most cases between 200 and 280 nm and can be eluted from the column using a mixture of acetonitrile and water. For progesterone this ratio is 80 : 20 (v/v) and for prednisone 25 : 75 (v/v). For the determination of initial enzyme activity procedure I is preferred, since it is the most rapid and the simplest. In Fig. 1 the area is shown in which initial 20fl-hydroxysteroid dehydrogenase activities could be measured. Procedure 2 is preferred when long-term experiments are performed (see Fig. 6). Determination of the Intrinsic Rate Parameters of Enzymes in Reversed Micellar Media

Principle. The methods employed to determine the intrinsic rate parameters kcat and Km in aqueous media are also applicable in reversed micellar media. The only difference with an aqueous system is that in a reversed micellar medium one can relate the Km of a water-soluble substrate either to the overall (K °v) or to the water pool concentration (KWp). This in contrast to the Km for apolar compounds, which of course is always related to the overall concentration. In most studies so far the observation is that the K wp is much larger than in water, whereas the K °v is about the same as in water. 6,j3 We disagree with these authors that the 13 S. Barbaric and P. L. Luisi, J. A m . Chem. Soc. 103, 4239 (1981).

[20]

E N Z Y M A T I C REACTIONS IN REVERSED M1CELLES

223

close resemblance between the K~nv and the K m in water is accidental. In our opinion the only physically meaningful expression for Km in reversed micellar media is K~v . Theory. When an enzymatic reaction is considered that obeys Michaelis-Menten kinetics the following reactions take place: kl E + S .

k2 " ES

, E + P

k-i

(1)

where E is the enzyme, S the substrate, and P the product, kl, k_ ~, and kz are rate constants. It is clear that the velocity (v) of such a reaction does not depend on the way the concentrations are expressed. Hence Vmax

Vmax

v = 1 + (K°V/S °v) = 1 + (KWp/S wp)

(2)

When 0 is the volume fraction of the water phase, the water pool and overall concentrations are related as follows S °v = OS wp

(3)

As a result Eqs. (4) and (5) can be written K ° v / o s wp = K w p / S wp

K °v = OK wp

(4) (5)

Since Km is defined as K m = (k-1 + kz)/kl

(6)

and k_~ and k2 are first-order rate constants, k~ v=

(1/O)k~ p

(7)

the correction factor (1/0) indicates that the substrate is dispersed over the whole volume of the solution and in fact reflects the frequency factor of collision and exchange between micelles. Hence K °v is relevant and not Kwp.

Deviations of K°mv from Km values found in aqueous media may originate either from microenvironmental effects on the enzyme itself or from interface-substrate interactions. Km a n d keat V a l u e s f o r 20fl-Hydroxysteroid D e h y d r o g e n a s e . For catalysis 20fl-hydroxysteroid dehydrogenase requires the water-soluble substrate NADH and in case of a reversed micellar system an apolar steroid, for example progesterone. The K °v values for NADH and for progester-

224

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[20]

TABLE I MICHAELIS-MENTEN PARAMETERSFOR 20/~-HYDROXYSTEROID DEHYDROGENASE IN AQUEOUS AND REVERSED MICELEAR MEDIAa K~nv (~M) Medium Aqueous Reversed micellar

NaC1 (mM)

NADH

Progesterone

kcat (sec -~)

0 500 0 500

3.6 3.4 11.0 3.7

10 n.d. 400 n.d.

10.9 6.2 14.4 7.6

a Experiments were performed in 50 mM Bistris, pH 6.3, or in reversed micellar solutions of 0.2 M CTAB in octane, 10% hexanol, w0 = 10, containing 50 mM Bistris, pH 6.3, in the water pool. Similar solutions were prepared containing 500 mM NaCI. For the determination of K °v for NADH the progesterone concentration was kept constant at 5 mM and for the determination of K °v for progesterone the overall NADH concentration was fixed at 50 ~M. n.d., Not determined. (Taken from ref. 3.)

one as well as the kcat values are listed in Table I both for an aqueous and for a CTAB reversed micellar medium. Compared to an aqueous solution, kcat increases in CTAB reversed micelles. This increase is not caused by a shift in pH of the water pool, since in aqueous solutions of pH 7.0 and 6.3 similar values for kcat a r e obtained. Upon enclosure in reversed micelles, K °v for NADH increases. Considering that NADH is negatively charged and CTAB positively, it can be envisaged that NADH is associated with the interface. Hence, the effective concentration available to 20/3-hydroxysteroid dehydrogenase could be lower, resulting in an increase in K°mv. In a reversed micellar medium the differences in K °v for NADH and kcat decrease at high ionic strength, indicating that the changes in k~atand K °v for NADH are caused mainly by electrostatic effects imposed by the charge of the head groups of the surfactant. In contrast to the K °v for NADH, that for progesterone increases markedly upon enclosure in reversed micelles, which could be due to a relatively low concentration of progesterone in the microenvironment of the enzyme compared to the bulk concentration. In conclusion it can be said that the enzyme experiences an essentially aqueous environment in reversed micelles, since kcat and K °v for NADH hardly change upon enclosure in the presence of salt, whereas the change in K°mv for progesterone is not caused by an effect on the enzyme itself, but by changes in the microenvironment.

[20]

ENZYMATIC REACTIONS IN REVERSED MICELLES

225

Determination of the Parameters That Regulate the Enzymatic Conversion of Apolar Substrates in Reversed MiceUes

Principle. When an enzyme is encapsulated in a reversed micelle and the substrate is present in the oil phase, the substrate has to cross barriers such as the interphase and, if present, the aqueous layer surrounding the enzyme before catalysis takes place. It can therefore be expected that for apolar substrates the enzyme activity will be influenced by (1) the thickness of the water layer around the protein, and (2) the concentration of substrate in the interphase. Consequently, a reversed micellar medium should be designed in such a way that the water content is as low as possible and that the solubility of the substrate in the interphase is high with respect to the continuous phase. The water layer around the enzyme can easily be varied by adjusting w0. The radius of the water pool increases when w0 increases. TM As a measure for the solubility of a substrate in the interphase the hydrophobicity value of the interphase, log P i , c a n be used. 2,3 Since the interphase contains only cosurfactant and CTAB the semiempirical formula log Pi -

ao

a0+ 1

log Pcosurfactant +

1

log PCTAB

(8)

can be defined, where a0 is the molar ratio of the cosurfactant to CTAB in the interphase. Similarly, a log P for the continuous phase, log Pcph, can be defined according to Eq. (9), log Pcph = Co log ecosurfactant -1- (1 - c0) log Porganic solvent

(9)

where Co is the mole fraction of the cosurfactant in the continuous phase and thus equal to E q . (10). In this formula Ccph, p, M °, and My are the amount of cosurfactant in the continuous phase (/zl/ml), the density (kg/liter), and the molecular weight of the organic solvent (o) and the cosurfactant (c), respectively. CcphpC/MCr CO = CcphpC/MCr "k (1000 -- Ccph)p°/M °

(10)

By determining the log P values of the different components of the system, Ccphand a0 (see below), log Pcph and log Pi c a n be calculated for any given composition of the reversed micellar medium. Hence a quantitative measure for the hydrophobicity of the different phases in a reversed mi14 R. A. Day, B. H. Robinson, J. H. R. Clarke, and J. V. Doherty, J. Chem. Soc., Faraday Trans. 1 75, 132 (1979).

226

IMMOBILIZED

ENZYMES/CELLS

IN ORGANIC

SYNTHESIS

[20]

S-

F 'T

E % E

>, >,

ru -I-

C3 tf~

0

01 02 03 04 }logPi- [ogPs I

\\0%

FIG. 3. The influence of ling P~ - log Psi and Ilog Pc~h -- log Psi on the 20/3-hydroxysteroid dehydrogenase activity with prednisone as substrate. For explanation, see text; for details, see Hilhorst e t al. 2.3

cellar medium can be obtained, which can be compared to the hydrophobicity of the substrate (log Ps) and related to the enzyme activity. Since small differences between log P values indicate good mutual solubilities, high substrate concentrations in the interphase and thus high enzyme activities are expected when Ilog P i - log Psi is minimal and when [log Pcph - log Psi is as large as possible. With these rules for any given apolar compound a medium can be composed that assures optimal enzymatic conversion. Figure 3 shows the results obtained with 20fl-hydroxysteroid dehydrogenase in H E P E S - C T A B - h e x a n o l - o c t a n e reversed micelles with prednisone as substrate. The results clearly show the effects of Ilog P~ - log Psi and ]log Pcph -- log Psi on the enzyme activity and that these effects are additive and thus amplify each other. Determination of log P. Log P is defined as the logarithm of the partition coefficient in an octanol water two-phase system. J5 For simple molecules log P values are known or can be calculated from hydrophobic fragmental constants according to Rekker. 15 For more complicated structures such as steroids log P can be determined easily as follows. One micromole of apolar compound is shaken overnight in sealed vials containing 1 ml octanol and 1 ml of an aqueous solution. The composition of the aqueous solution and the temperature should be the same as those used in the enzymatic assay. The concentration of the apolar com~5 R. F. Rekker, " T h e Hydrophobic Fragmental Constant." Elsevier, Amsterdam, 1977.

[20]

ENZYMATIC REACTIONS IN REVERSED MICELLES I

I

I

227

I

300

3 -5

200

x

~oo-~vl I

I

I

I

1

2

3

4

V ocfane (m[)

Fi6. 4. Typical example of a titration curve for the HEPES-hexanol-CTAB-octane medium. For explanation see text. Ccphis the amount of hexanol in the continuous phase (in ~l/ml); V~ is the amount of hexanol (in p.1) in the interphase, and V denotes total volume.

pound in both phases can be determined afterward using the appropriate detection technique. Determination ofccph and ao. The mole fraction of cosurfactant in the interphase (a0) and its amount in the continuous phase (Ccph) are determined by phase boundary titrations according to Bowcott and Schulman. 16 For the HEPES-hexanol-CTAB-octane system the procedure is as follows. A 2 ml solution containing 0,2 M CTAB in octane and 50 mM HEPES (pH 7.6) at the desired water content is prepared in a Bausch and Lomb 0.5-in. test tube. This solution is titrated using a microsyringe with hexanol to clarity under vigorous shaking. Upon dilution with octane the solution becomes turbid again because of extraction of the hexanol from the interphase into the continuous phase. This process of dilution and titration is repeated 5 times in the same tube. Each titration is carried out in duplicate or triplicate at the temperature used in the enzyme assay. A plot of the minimal amount of hexanol required for the formation of a clear solution vs the total volume of octane added should yield a straight line (Fig. 4). The slope represents the amount of hexanol in the continuous phase in microliters hexanol per milliliter continuous phase (ccph), whereas the intercept on the Yaxis (V~) represents the amount of hexanol in microliters present in both the interphase and the water pool. Since 16 j. E. Bowcott and J. H. Schulman, Z. Elektrochem. 59, 283 (1955).

228

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS (c0)surfattant~.~

2~'' '~÷

N'~'+

2 - ~ / f - MV ~

i

~H2ase

,

.

.

.

'+

[20]

.

~/~

APOLAR

AUM t - I ~ K E T O S T E R O I O

.~L,pOH

",..,2My++2H+J ~ N A D

..........

j% HSOH~

+'

J~APOLAR

A~

~ofl-HYDROXYSTER0ll]

FIG. 5. Scheme for the Hz-driven regeneration of NADH and the subsequent reduction of an apolar steroid in a reversed micellar medium. For clarity all the water-soluble components of the system are drawn in one micelle. A more realistic view is that all components are distributed among all micelles that are in rapid exchange with each other. H2ase, Hydrogenase; LipDH, dihydrolipoamide dehydrogenase; HSDH, 20/3-hydroxysteroid dehydrogenase; and MV, methyl viologen. From Hilhorst et al. ]

hexanol is very poorly soluble in water, the latter amount is negligible. V~ times the density of the cosurfactant and divided by its molecular weight yields the amount of hexanol in the interphase (ci) in micromoles. Since all surfactant molecules are localized in the interphase, a0 can be calculated according to Eq. (11), ao

= ciMS/S

(11)

where S is the amount of surfactant in the assay in milligrams and M s the molecular weight of the surfactant. Multienzymatic Reactions in Reversed Micellar Media Principle. An important property of reversed micelles is that they are in rapid exchange with each other. Hence, enzymes distributed among these microdroplets can communicate with each other either directly or through low molecular weight mediators. Figure 5 shows an example of a carrier-mediated multienzymatic system in a reversed micellar medium. ~ Cytochrome-c3 hydrogenase (EC 1.12.2.1) and dihydrolipoamide dehydrogenase (EC 1.8.1.4) are used for the in situ regeneration of NADH which is consumed by 20fl-hydroxysteroid dehydrogenase during the site- and stereospecific reduction of an apolar ketosteroid to its corresponding 20fl-hydroxyform. Procedure. An aqueous solution (70 tzl) is injected into 1.5 ml of a Vortex-stirred 0.2 M solution of CTAB in hexanol-octane 1 : 4 (v/v) containing 1.0 mM steroid unless stated otherwise. Stirring is continued until the solution becomes clear. Any assay vial can be used, provided it contains a gastight rubber septum that allows the withdrawal of aliquots from the reaction mixture.

[20]

ENZYMATIC REACTIONS IN REVERSED MICELLES I

i

229

i

~E E

o

i

co i

~i lC

x o

o

=~ 0 ~.

0

~l

I

6

I

9

Time (hrs)

FIG. 6. Time course of 20fl-hydroxypregn-4-en-3-one formation at different concentrations of progesterone in a HEPES-hexanol-CTAB-octane reversed micellar medium. See text for procedure. Progesterone concentrations: 0.2 mM (IlL 1.0 mM (A), and 5.0 mM (N). From Hilhorst et al. 1

The aqueous solution contains pig heart dihydrolipoamide dehydrogenase (2.5 /.tM), 20fl-hydroxysteroid dehydrogenase (3.0/zM), methyl viologen (2.5 mM), and NAD ÷ (1.0 mM) in 50 mM HEPES (pH 7.6). After six cycles of 30 sec evacuation, 15 sec flushing with O2-free argon, and 5 min of bubbling with O2-free H2, 10/xl hydrogenase isolated from Desulfovibrio vulgaris 17 (final concentration of 0.6/xM in the water pool) is added to initiate the reaction. During the reaction the incubation is shaken mildly at 25 ° . Figure 6 shows the time course of progesterone reduction. The results clearly demonstrate that the reaction rate declines when the conversion is almost complete. Furthermore, at higher concentrations the reaction rate increases and the reaction proceeds steadily for >10 hr, indicating that the multienzyme system is stable for a considerable length of time. Acknowledgments We wish to thank Mrs. J. C. Toppenberg-Fang for typing the manuscript, Mr. M. M. Bouwmans for drawing the figures, and the Pharmaceutical Industry Duphar BV, Weesp, and the Netherlands Organization for the Advancement of Pure Research (Z.W.O.) for financial support. 17 H. M. Van der Westen, S. G. Mayhew, and C. Veeger, F E B S L e t t . 86, 122 (1978).

230

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[21]

[21] D i s a c c h a r i d e S y n t h e s i s w i t h Immobilized fl-Galactosidase B y P E R - O L O F L A R S S O N , L A R S H E D B Y S , S I G F R I D S V E N S S O N , and

KLAUS MOSBACH Many carbohydrates are known to interfere in cellular recognition mechanisms in a very efficient way. j This fact forms the basis for the design of new drugs that would, for example, prevent the adhesion of bacteria on mucosal surfaces. The carbohydrate drug could be envisaged to block the bacterial receptors, making them unable to recognize or bind to the host cell, a process known to be a crucial initial step of the pathogenesis. The interest in such biologically active carbohydrates has prompted the development of new methods for carbohydrate synthesis. Normal organic chemical methods for di- and oligosaccharide synthesis are usually laborious and time consuming, due to a number of protection and deprotection steps. The use of enzymes for carbohydrate synthesis is conceptually very attractive, as the procedure often is a single-step reaction. However, several isomers may be formed. When only a few isomers are formed, as is the case here, the workup is also straightforward, making the overall procedure very convenient. Under the proper conditions, the enzyme fl-galactosidase will catalyze a transgalactosylation reaction (Fig. 1). 2 The fl-galactoside, lactose, combines initially with the enzyme to form a galactosylated enzyme with concomitant release of glucose. The galactosylated enzyme may experience two fates. Either it is hydrolyzed with water to give free enzyme and galactose, or it may transfer its galactosyl moiety to a second sugar, in this case N-acetylgalactosamine. This transgalactosylation reaction leads to the formation of galactosyl-N-acetylgalactosamine. Under normal aqueous conditions the final equilibrium is shifted far toward hydrolysis. However, the transfer reaction is comparatively fast, and during an intermediate phase a substantial yield from the transfer reaction is obtained. A third route is also evident from the scheme, namely the reversal of the hydrolysis reaction. For any substantial synthesis to occur, the water activity must be kept low. The procedures given below describe the synthesis of the disaccharide galactosyl-N-acetylgalactosamine [fl-D-Gaip-(I-6)-D-GalNAc] in good 1 N. Sharon and H. Lis, Chem. Eng. News March 30, p. 21 (1981). 2 K. Wallenfells and R. Well, in " T h e E n z y m e s " (P. Boyer, ed.), Vol. 7, 3rd Ed., p. 617. A c a d e m i c Press, N e w York, 1972.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987by AcademicPress, Inc. All rights of reproduction in any form reserved.

[21]

DISACCHARIDESYNTHESIS

231

Galactosyl transfer

"GIc E + GaI-GIc

+ GalNAc E-Gal

ll

+H20

GaI-GalNAc + E

Hydrolysis

E+Gal

FIG. 1. Simplified scheme of proposed mechanism of/3-galactosidase, showing the transfer and the hydrolysis reactions. E,/3-Galactosidase; Gal-Glc, lactose; GalNAc, N-acetylgalactosamine; Gal-GalNAc, galactosyl-N-acetylgalactosamine, Reproduced from Hedbys et al. 4 with permission.

yield from lactose and N-acetylgalactosamine and in low yield from galactose and N-acetylgalactosamine via reversal of the hydrolysis reaction. The enzyme was used in the immobilized form to facilitate its reuse and its easy removal from the reaction mixture when the yield was at its maximum. In addition, and of relevance for future studies, immobilization may lead to more stable enzyme preparations of importance when carrying out enzymatic synthesis at elevated temperatures and in organic solvents and/or solvents with low water activity. Methods

Immobilization of fl-Galactosidase to Sepharose CL-4B The immobilization is carried out by the tresyl chloride m e t h o d ) : Sepharose CL-4B (Pharmacia Fine Chemicals, Uppsala, Sweden), 5 g of moist gel, is washed with water and then transferred stepwise to organic solvent by washing with a series of solvents: 30 : 70, 50 : 50, and 70 : 30 acetone:water, followed by acetone, and finally acetone dried with a molecular sieve) The gel is filtered off, carefully avoiding any moist air to be sucked through the gel bed, and then transferred to a 10-ml beaker containing 1.13 ml dry acetone and 95/xl pyridine. Tresyl chloride, 6 125 3 K. Nilsson and K. Mosbach, this series, Vol. 104, p. 56. 4 L. Hedbys, P.-O. Larsson, K. Mosbach, and S. Svensson, Bi oc he m . Biophys. Res. Commun. 123, 8 (1984). 5 One hundred grams of Linde molecular sieve type 4 A is added to 1 liter of acetone. Drying time is 24 hr. 6 Tresyl chloride (2,2,2-trimethylsulfonyl chloride) was obtained from Fluka, Buchs, Switzerland. Tresyl chloride-activated Sepharose gel is available from Pharmacia Fine Chemicals, Uppsala, Sweden.

232

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[21]

/zl, is added under vigorous magnetic stirring to rapidly disperse the activating agent. The mixture is then stirred at a slow rate for 15 min at room temperature. The activated gel is washed with acetone and then transferred to an aqueous solvent by washing with 70 : 30, 50 : 50, and 30 : 70 acetone : 5 mM HC1 in water, followed by 5 mM HCI. The activated gel is stored at 4° until used. The activated gel, about 5 g wet weight, is washed with 0.20 M sodium phosphate buffer, pH 8.0, and mixed with 6 mg of Escherichia coli [3galactosidase 7 in 6 ml of the same buffer. The suspension is shaken gently for 2 hr at room temperature. The [3-galactosidase-carrying gel is washed with the above buffer, followed by 0.2 M Tris-HCl, pH 8.0, in which it is kept for 30 min to block any remaining tresyl groups. The gel is then washed with 0.5 M NaC1, water, and finally 0.10 M sodium phosphate buffer, pH 7.0, containing 2 mM MgCIz. The activity of the gel is routinely determined from its ability to hydrolyze o-nitrophenyi-[3-o-galactoside. The gel typically has an activity of 48 U/g of gel, which corresponded to about 50% of added activity.

Enzyme Assays [3-Galactosidase activity is determined from initial rate measurements of o-nitrophenyl-[3-D-galactoside hydrolysis at 24°. The assay mixture contains 1.0 mM o-nitrophenyl-[3-D-galactoside, 2 mM MgC12, and 0.10 M sodium phosphate buffer, pH 7.0. The hydrolysis reaction is followed at 420 nm, where the product has a molar absorption coefficient of 2050 Mcm -~ (pH 7). 8 When immobilized enzyme is assayed, a spectrophotometric cell equipped with a magnetic stirring device is used to keep the particles suspended. Normally 2-5 mg of gel is used per assay (2 ml assay volume).

Preparation of Galactosyl-N-acetylgalactosamine [[3-D-Galp-(1-6)-D-GalNAc] via Transglycosylation Immobilized enzyme, 1.4 g (70 U), is suspended in 17 ml 0.10 M potassium phosphate buffer, pH 6.5, containing 0.22 M N-acetylgalactosamine and 0.45 M lactose. The mixture is shaken gently at room temperature, and samples taken intermittently for HPLC analysis to determine the correct time for breaking the reaction, i.e., when the yield was at its 7 Grade VI was obtained from Sigma Chemical Co., St. Louis, MO. The preparation used had a specific activity of 80 U/mg with the assay conditions given under "Enzyme Assays." 8 The molar absorption coefficient is very pH sensitive around pH 7 (pK, for o-nitrophenol = 7.2). To avoid this problem, the assay may be run at pH 8.

[21]

DISACCHARIDESYNTHESIS

233

maximum. Samples are diluted with 9 volumes acetonitrile : water 3 : 1, and injected on a LiChrosorb-NH2 column (Merck; 5-/~m particles, 25 cm length). The mobile phase consists of acetonitrile : water 3 : I. The effluent is monitored at 235 nm, at which wavelength acetamido sugars are detected. The k' value of N-acetylgalactosamine was determined to be 1.6 and the k' value for the product galactosyl-N-acetylgalactosamine was 3.6. After 30 hr incubation, the reaction is stopped by spinning down the gel. As an extra measure of precaution, the supernatant is boiled for 3 min to denature any traces of enzyme, and freeze-dried. The yield, according to HPLC measurements, is at this stage about 25%. The product /3-DGalp-(1-6)-D-GalNAc is isolated in pure form by using a combination of chromatographic techniques. 4

Preparation of Galactosyl-N-acetylgalactosamine [/3-D-Galp-(1-6)-D-GalNAc] via Reversal of the Hydrolysis Reaction Immobilized enzyme, 5 U, is suspended in 0.50 ml of 0.10 M potassium phosphate buffer, pH 6.5, containing 0.31 M galactose and 0.31 M 2acetamido-2-deoxy-D-galactose. The mixture is gently shaken at room temperature for 24 hr before the reaction is stopped and the mixture treated as above. Comme~s The described transferase-type synthesis gave a very good yield of galactosyl-N-acetylgalactosamine [fl-D-Galp-(1-6)-D-GalNAc]. Analysis of the reaction mixture showed that small amounts of the disaccharides/3D-Galp-(1-6)-D-Gal and/3-D-Galp-(1-6)-D-GIc also had been formed. Remarkably, only (1-6)-linked GalGalNAc could be f o u n d . 4 Preliminary results with enzymes from other sources indicate that other linkages are possible to obtain. The product galactosyl-N-acetyl-galactosamine had not been described earlier and was therefore identified/characterized by sugar and methylation analyses using gas chromatography-mass spectrometry as analytical tools. Proton NMR data and the observed optical r o t a t i o n (o~D = 26°) were also in agreement with the proposed structure? The reversed hydrolysis reaction, on the other hand, resulted in only low yield under the conditions given here. To make the latter reaction suitable for preparative applications, the conditions must be altered substantially. However, recent preliminary studies with several glycosidases, including/3-galactosidase, suggest that satisfactory yields (10-50%) are possible, provided that a combination of measures are taken such as very high substrate concentration and high temperature (to increase substrate solubility and decrease water activity).

234

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[22]

[22] S y n t h e s i s o f U r e a w i t h U r e a s e in W a t e r - O r g a n i c Solvent Mixtures B y LARRY G . B U T L E R

The synthesis of urea in animals involves several enzymes and is driven by hydrolysis of 3 mol of ATP per mole of urea synthesized. 1 Urea can be readily synthesized nonenzymatically from the inorganic salt ammonium cyanate; this original synthesis by W6hler in 1828 was a landmark in the history of chemistry. 2 It has now become possible to demonstrate the synthesis of urea from an inorganic salt in a reaction catalyzed by a single enzyme, independent of hydrolysis of ATP or any other high energy compound. The inorganic salt is ammonium carbonate; the enzyme is urease, which catalyzes urea synthesis by reversal of urea hydrolysis2 Reversal of hydrolytic reactions was originally considered as a biosynthetic route; the first claim of an enzyme-catalyzed synthesis reaction was by reversal of the hydrolytic reaction catalyzed by maltase (a-glucosidase). 4 Because water is not only the solvent but is a reactant at high concentration, the equilibrium for hydrolytic reactions lies far toward hydrolysis when carried out in aqueous media. Methods for overcoming this thermodynamic constraint have been developed so that reversal of peptide bond hydrolysis has become a useful method for semisynthesis of proteins 5 and is being extended to large-scale synthesis of specialized materials such as aspartame. 6 Hydrolysis of urea is equivalent to hydrolysis of two peptide bonds, so the equilibrium lies even farther toward hydrolysis than for comparable reactions such as peptide bond hydrolysis. 7 Nevertheless, synthesis of urea from ammonium carbonate can readily be detected and quantitated. Replacement of part of the solvent water by water-miscible organic solvents dramatically increases the yield of urea. 3 This increase was presumed to be due to the corresponding decrease in water activity, shifting the equilibrium toward urea synthesis, 3 but an effect of the organic solJ H. 2 F. 3 L. 4 C. 5 R. 6 K. 7 A.

A. Krebs and K. Henseleit, Hoppe-Seyler's Z. Physiol. Chem. 210, 33 (1932). W#hler, Poggendorff's Annalen 12, 253 (1828). G. Butler and F. J. Reithel, Arch. Biochem. Biophys. 178, 43 (1977). Hill, J. Chem. Soc. 73, 634 (1898). E. Offord, "Semisynthetic Proteins." Wiley, New York, 1980. Oyama and K. Kihara, CHEMTECH 14, 100 (1984). Kapune and V. Kasche, Biochem. Biophys. Res. Commun, 80, 955 (1978).

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. AI! rights of reproduction in any form reserved.

[22]

UREASE-CATALYZED UREASVNTHEStS

235

vent on ionization a may also be involved. Several problems associated with synthesis by reversal of enzymatic hydrolysis reactions are illustrated by investigation of the effects of reaction conditions on the rate and extent of urea synthesis as catalyzed by urease. Source of Enzyme. Jack bean urease extracted and purified by the standard acetone crystallization procedure 9 with minor modifications ~° was used in demonstrating urease-catalyzed urea synthesis) Utilization of urease from other sources should be possible, although conditions for good reaction rates would have to be established for the particular enzyme used. Conditions which determine the position of equilibrium of the urea hydrolysis reaction (see below) are of course independent of the source of the enzyme. Assays

Urea Synthesis. In order to obtain readily measurable amounts of urea it is necessary to use relatively high concentrations of substrate. Typical conditions are 0.75 M (NH4)zCO3, 5 mM EDTA, and 5 mM 2-mercaptoethanol (no additional buffer is necessary). The urease concentration should be about I Sumner Unit 11 per milliliter. Under these conditions Jack bean urease catalyzes urea synthesis at about 2% of the rate at which it hydrolyzes urea (conditions for urea hydrolysis specified below). Urea concentration is determined spectrophotometrically by a modification ~2of the Fearon reaction.13 Incubated samples, zero time blanks, or urea standards containing up to 1/zmol urea in a total volume of 0.2-1.0 ml in screw-capped tubes are quenched by addition of 5 ml of 52% H3PO4 (600 ml of 85% H3PO4 diluted with water to 1 liter). Samples can be kept at room temperature at this stage for up to an hour without significant loss of urea, so that several samples can be run together through the remainder of the procedure. To each quenched sample is added 1.0 ml of an aqueous solution containing 6 g of 2,3-butanedione monoxime and 0.3 g thiosemicarbazide (both from Aldrich Chem. Co.) per liter. After heating for 30 min on a boiling water bath, samples are cooled to room temperature and their absorbance at 530 nm is determined immediately. The assay is not affected by high concentrations of ( N H 4 ) 2 C O 3 . At smaller sample volumes (0.2-0.5 ml) the apparent extinction coefficient for urea (approxi8 G. A. Homandberg, J. A. Mattis, and M. Laskowski, Jr., Biochemistry 17, 5220 (1978). 9 G. Mamiya and G. Gorin, Biochim. Biophys. Acta 105, 382 (1965). ~0C. M. Kaneshiro, Ph.D. Dissertation, University of Oregon, Eugene (1975). N j. B. Sumner, this series, Vol. 2, p. 378. ~2 M. K. Schwarz, this series, Vol. 17, p. 857. 13 W. R. Fearon, Biochem. J. 33, 902 (1939).

236

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[22]

mately 2 × l06 M -~ c m - 0 is relatively insensitive to the presence of organic solvents in the sample. For samples in which the equilibrium concentration of urea is being determined, it is advisable, after measuring the urea concentration, to add fresh urease and incubate again until the urea concentration is constant. Urea Hydrolysis. For comparison to urea synthesis, urea hydrolysis can be measured by an adaptation of the same assay using 2 mM urea as substrate, in the presence of 0.2 mM EDTA, 5 mM 2-mercaptoethanol, and 20 mM imidazole acetate, pH 7.0. This urea concentration is below the Km for urea 14 so the rate decreases continually throughout the incubation. Data points at four different reaction times are plotted semilogarithmically; the first-order rate constant is determined from the slope. Effect of Organic Solvents on Enzyme Stability The use of organic solvents to enhance the yield of urea is possible only if urease activity can be maintained in the presence of the solvents. Fortunately, Jack bean urease is known to be rather stable in organic solvents such as glycerol.~5 Moreover, it is extracted and crystallized in 32% acetone. 9 In connection with these experiments it was observed that urease in 50% aqueous dimethylformamide, ethanol, or pyridine irreversibly lost most of its activity in a period of a few hours at room temperature. However, activity was recovered in high yield from urease dissolved in 50% aqueous solutions of acetone or polyhydroxylic alcohols such as glycerol and 1,2-propanediol. In an attempt to stabilize Jack bean urease against denaturation and loss of catalytic activity by high concentrations of organic solvents, the enzyme was immobilized and cross-linked by two independent methods. In neither case, however, did immobilized cross-linked urease have improved stability in the presence of organic solvents) Subsequent work was done only with soluble urease. Effect of Organic Solvents on Reaction Rate Organic solvents affect urea hydrolysis in much the same way that they affect urea synthesis .3 The effect of organic solvents on reaction rate can thus be conveniently determined by measuring urea hydrolysis, which requires less enzyme than urea synthesis. The effects are strongly dependent on the nature of the organic solvent, as well as its concentra~4 K. R. L y n n , Biochim. Biophys. Acta 146, 205 (1967). ~5 C. C. Contaxis and F. J. Reithel, J. Biol. Chem. 246, 677 (1971).

[22]

U R E A S E - C A T A L Y Z E D UREA SYNTHESIS

237

tion. The solvents were reagent grade but not redistilled; further purification of the solvents might be warranted. When compared at the same concentration (50%), the rate of urea hydrolysis ranged from undetectable (in ethanolamine) up to equal activity in acetone to that observed in strictly aqueous solution. 3 Rates in 50% aqueous glycerol, dimethylformamide, ethanol, and pyridine were 15, 49, 56, and 65% of the rate in water. In acetone and glycerol, in which urease is stable indefinitely, the activity as a function of organic solvent concentration shows rather different patterns, with most of the activity retained up to 75% acetone and much less activity observed at relatively low glycerol concentrations) Solvent selection for proposed urea synthesis processes wouldrequire optimization of the particular solvent for enzyme stability, reaction rate, and yield of urea (see later section). Effect of Organic Solvent, Substrate, and Hydrogen Ion Concentration on Yield of Urea The equilibrium concentration of urea, determined in solvents in which urease is stable so that achievement of equilibrium can be assured, increases as the water activity (related to concentration) diminishes. The yield of urea at equilibrium is approximately a second-order function of water concentration (activity) as predicted by the equilibrium constant. 3 The yield of urea as a function of (NH4)2CO3 concentration, at constant solvent composition, strongly increases as the substrate concentration increases. 3 In 50% glycerol, the observed urea yield was greater than the theoretical yield (which assumes concentration and activity are equivalent) at low substrate concentrations and lower than the theoretical yield at high substrate concentrations. The yield of urea is greatly enhanced in alkaline conditions) H + is a by-product of urea synthesis from HCO3-, according to Eq. (1). 2 NH4 + + H C O 3 - --0 u r e a + 2 H 2 0 + H +

(1)

Effect of the Nature of the Organic Solvent on Yield of Urea The equilibrium concentration of urea differs greatly depending upon which organic solvents are utilized to replace part of the water. Compared at a concentration of 50%, the equilibrium concentration of urea in aqueous acetone is over 12-fold higher than in aqueous glycerol. 3 Other organic solvents tested gave yields between these values, with hydroxylic solvents generally showing the lowest yields) As determined by measurement of the solubility of urea and ammonium carbonate, the primary effect of organic solvents is on the activity of the salt, rather than urea.

238

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[22]

Use of Organic Solvents to Study the Reaction Mechanism It has been reported that the products of urease-catalyzed urea hydrolysis are ammonium and carbamate ions. 16 Measuring urea synthesis rather than hydrolysis, it is possible to show that maximum initial rates are obtained only if the substrate ammonium bicarbonate has been dissolved for an hour before urease is added; freshly dissolved substrate gives much slower initial rates of urea synthesis s This difference is likely due to a requirement for synthesis from carbamate, which forms nonenzymatically at a slow rate after ammonium bicarbonate is dissolved. In order to prove this point, carbamate concentration should be measured in both fresh and aged substrate solutions, and the rate of synthesis from ammonium carbamate should be directly compared with the rate of urea synthesis from ammonium carbonate. Significance The small amounts of urea synthesized (approximately O. 1% yield) by reversal of the urease reaction are not likely to warrant consideration of this process for large-scale urea production. Urease does, however, provide an interesting model for investigation of the reversal of hydrolytic enzyme reactions, both because of the historical significance of urea synthesis, and because urease illuminates several factors which limit utilization of the technology of reversal of enzymatic hydrolysis. Perhaps most important is the realization that no single set of conditions is optimal for all the parameters affecting urea synthesis. For example, conditions which give high urea yields also give slow rates of synthesis, and conditions in which synthesis is rapid give poor yields. Enzyme efficiency and stability and substrate solubility likewise may be affected differently by organic solvents. Some of these problems may be partially overcome by using mixtures of different organic solvents.

~6R. L. Blakely, J. A. Hinds, H. E. Kunze, E. C. Webb, and B. Zerner, 1991 (1969).

Biochemistry 8,

[23]

IMMOBILIZEDDEXTRANSUCRASE

239

[23] D e x t r a n S y n t h e s i s U s i n g I m m o b i l i z e d Leuconostoc rnesenteroides D e x t r a n s u c r a s e

By P. MONSAN, F. PAUL, D. AURIOL, and A. LOPEZ Dextransucrase [ a ( 1 ~ 6)-glucan:o-fructose-2-glucosyltransferase E.C.2.4.1.5] catalyzes the synthesis of a glucose polymer, dextran, according to the reaction: n Sucrose ~ (glucose)n + n fructose dextran

Dextran is a branched polymer in which the linear parts result from linear condensation of o-glucopyranosyl units by a ( 1 --~ 6)-osidic bonds, whereas branching points result from a(1 ~ 3), a(1 ~ 2), or a(1 ~ 4) bonds. The degree, nature, and length of branching are related to the strain of bacteria which produced the enzyme and the conditions of polymer synthesis.~ Dextransucrase from Leuconostoc mesenteroides NRRL B512 F, a nonpathogenic lactic bacterium, produces a low-branched dextran polymer containing 95% a ( 1 ~ 6) bonds and 5% o~(1~ 3) bonds, and which is highly soluble in water. This explains the exclusive use of this strain for industrial production of dextran. Produced in the same manner since 1950, dextran is mainly used as a blood plasma substitute (clinical dextran) and chromatography support. 2 Dextran is also involved in the process of dental caries: pathogenic bacteria of the genus Streptococcus produce insoluble dextran, which favors the adhesion of bacteria on the teeth and causes the formation of dental plaque. 3 Leuconostoc mesenteroides dextransucrase is an exocellular enzyme induced by its substrate, sucrose, the only inductor known at present. 4 Dextran synthesis is thus performed outside the cell, without the intervention of activated precursors or cofactors, as is the case for the synthesis of the majority of polysaccharides.5 In vitro dextran synthesis is thus A. Jeanes, W. C. Haynes, C. A. Wilham, J. C. Ranki, E. H. Melvin, M. J. Austin, J. E. Cluskey, B. E. Fisher, H. M. Tsuchiya, and C. E. Rist, J. Am. Chem. Soc. 76, 5041 (1954). 2 A. Jeanes, in "Extracellular Microbial Polysaccharides" (P. A. Sandford and A. Laskin, eds.), Symp. Ser. Vol. 45, p. 284. American Cancer Society, Washington, D.C., 1977. 3 R. L. Sidebotham, in "Advances in Carbohydrate Chemistry and Biochemistry" (R. S. Tipson and D. Horton, eds.), Vol. 30, p. 371. Academic Press, New York, 1974. 4 W. B. Neely, Arch. Biochem. Biophys. 79, 154 (1959). 5 R. C. W. Berkeley, G. W. Gooday, and D. C. EUwood, "Microbial Polysaccharides and Polysaccharases." Academic Press, New York, 1979.

METHODS IN ENZYMOLOGY, VOL. 136

Copyright © 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

240

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[23]

possible using a cell-free purified enzyme preparation. The energy required for the condensation of o-glucopyranosyl residues is provided by hydrolysis of the glucose-fructose osidic bond of the sucrose molecule. This energy is conserved by the formation of o-glucosyl-enzyme covalent complexes. The action mechanism of dextransucrase described by Robyt e t al. 6"7 involves two identical fixation sites of D-glucosyl residues, from which the latter are transferred to the reducing end of the growing dextran chain. This chain remains covalently bonded to the enzyme. During polymer synthesis, the enzyme forms an enzymatically active covalent complex with both glucose and dextran. Dextran release occurs when the fructose produced during the reaction is sufficiently concentrated to act as acceptor of the dextran chain. As fructose is a weak acceptor, the molecular weight of the polymer is extremely high, greater than 106.3 Fructose may also act as acceptor for D-glucosyl residues, resulting in the formation of leucrose, o~-o-glucopyranosido-(1 ~ 5)-fructopyranose. 8 However, other molecules, not involved in the reaction, may also act as acceptors for D-glucosyl and dextranosyl residues. Maltose is a particularly efficient acceptor causing the release of small oligodextrans containing a maltose residue at the reducing end. 9,1° These oligodextrans are themselves excellent acceptors.11,12 Control of dextran molecular weight during synthesis with the resulting direct production of polymers of sizes required for pharmaceutical and analytical applications may thus be envisaged. Along with a deeper understanding of the transfer mechanism of oglucopyranosyl residues and the possibilities of molecular weight control of dextran, dextransucrase production and purification have also been improved, in particular using fed-batch culture and ultrafiltration.13.14 We have recently developed an original purification process involving liquidliquid extraction of dextransucrase directly from culture broth. 15The enzymatic preparation thus obtained has a very high specific activity, 6 j. F. Robyt, B. K. Kimble, and T. F. Walseth, Arch. Biochem. Biophys. 165, 634 (1974). 7 j. F. Robyt and H. Taniguchi, Arch. Biochem. Biophys. 174, 129 (1976). F. H. Stodda, E. S. Sharpe, and H. J. Koepsell, J. Am. Chem. Soc. 78, 2514 (1956). 9 H. M. Tsuchiya, N. N. Hellman, H. J. Koepsell, J. Corman, C. S. Stringer, S. P. Rogovin, M. O. Bogard, G. Bryant, V. H. Feger, C. A. Hoffman, F. R. Senti, and R. W. Jackson, J. Am. Chem. Soc. 77, 2412 (1954). ~0j. F. Robyt and T. F. Walseth, Carbohydr. Res. 61, 433 (1978). n F. Paul, E. Oriol, D. Auriol, R. M. Willemot, and P. Monsan, Eur. Congr. Biotechnol. 3, (1984). 12 j. F. Robyt and S. H. Eklund, Carbohydr. Res. 121, 279 (1983). 13 p. Monsan and A. Lopez, Biotechnol. Bioeng. 23, 2027 (1981). 14 j. F. Robyt and T. F. Walseth, Carbohydr. Res. 68, 95 (1979). 15 F. Paul, P. Monsan, and D. Auriol, French Patent 8,307,650 (1983).

[23]

IMMOBILIZEDDEXTRANSUCRASE

241

greater than 170 U/mg protein and is free from contaminating saccharolytic activity. ~5,~6 Little research has been made into immobilization of L. mesenteroides dextransucrase. Robyt et al. 6,7 immobilized dextransucrase on BioGel P2 activated by glutaraldehyde in order to show the synthesis mechanism of dextran, and, in particular, the synthesis of a(1 ---* 3) branching points on the dextran linear chain. Recently, Reilly's group immobilized the enzyme by covalent bonding on various supports. ~7 But the most active derivatives were obtained by dextransucrase adsorption on phenoxyacetyl cellulose powder. 18 In our case, dextransucrase was immobilized on porous silica activated by glutaraldehyde in order to compare the kinetic behavior of the enzyme in free and immobilized form, and the nature of the dextran thus obtained. 13,19,2° Assay Methods Dextransucrase A s s a y . Enzyme activity is determined by measuring the initial production rate of reducing sugars in a solution containing 2 ml 3.3 M sucrose solution (Merck, Germany) in water, 0.24 ml 1 M sodium acetate buffer (pH 5.2) containing 2.5 g/liter calcium chloride, a volume of enzymatic solution with an activity between 2 and 20 U, and a volume of distilled water to obtain a final volume of 12 ml. Reaction takes place at a controlled temperature of 30° with stirring by magnetic rod. The amount of reducing sugars is determined using the reagent 3,5-dinitrosalicylic acid, 2~ with aliquots taken at 5-rain intervals over a period of 20 min, diluted if necessary to maintain reducing sugar concentration between 0.2 and 2 g/liter. Evolution of optical density as a function of time allows the initial production rate of reducing sugars to be determined. Dextransucrase activity (U) is defined as the amount of enzyme releasing 1/~mol of reducing sugar per minute. Protein Determination. Protein concentration is determined by the method of Lowry et al. 22 using bovine serum albumin as reference. 16 F. Paul, D. Auriol, E. Oriol, and P. Monsan, Enzyme Eng. 7, in press (1986). 17 n . Kaboli and P. J. Reilly, Biotechnol. Bioeng. 22, 1055 (1980). ~8 H. N. Chang, Y. S. Ghim, Y. R. Cho, D. A. Landis, and P. J. Reilly, Biotechnol. Bioeng. 23, 2647 (1981). ~9A. Lopez and P. Monsan, Biochimie 62, 323 (1980). 2o p. Monsan and A. Lopez, in "Advances in Biotechnology" (M. Moo-Young, C. W. Robinson, and C. Vezina, eds.), Vol. 1, p. 679. Pergamon, New York, 1981. 2~ j. B. Sumner and S. F. Howell, J. Biol. Chem. 108, 51 (1935). 52 O. H. Lowry, N. H. Rosebrough, A. L. Farr, and R. H. Randall, J. Biol. Chem. 193, 265 (1951).

242

IMMOBILIZED ENZYMES/CELLS IN ORGANIC SYNTHESIS

[23]

Dextran Determination. Dextran concentration is determined by the anthrone method using glucose as reference. 23

Production of Dextransucrase by Fed-Batch Culture of Leuconostoc mesenteroides Dextransucrase is produced by cultivation of the strain L. mesenteroides N R R L B512 F in a nutritive medium at 27° containing 2,4 yeast extract (Biokar, France), 20 g/liter; dipotassium hydrogenophosphate, 20 g/liter; sucrose, 20 g/liter; sodium chloride, 0.01 g/liter; magnesium sulfate (MgSO4"7H20), 0.2 g/liter; manganese sulfate (MnSO4. H20), calcium chloride (CaCI2" 2H20), and iron sulfate (FeSO4" 7H20), 0.01 g/liter. The pH of the medium is adjusted to 6.9 by pure orthophosphoric acid addition before sterilization. After the addition of a volume of preculture corresponding to 10% of total culture volume, cell growth is measured by following optical density at 650 nm of the culture medium diluted 10 times with distilled water. At the beginning of the exponential growth phase, a concentrated sucrose solution (800 g/liter) is added at a rate of 20 g/hrliter of culture medium. Sucrose must remain present in the medium to fulfill its role as growth carbon source, inducer, and substrate for dextransucrase, pH is maintained at 6.7, corresponding to optimum enzyme production. 24Cell growth and dextransucrase production are shown in Fig. 1. The addition of sucrose allows both the cell growth phase and dextransucrase production phase to be prolonged. Under these conditions, final culture activity was 8.8 U/ml, corresponding to a culture productivity of 1.2 U/ml-hr. 19 On completion of the growth phase, the culture medium is recuperated and centrifuged (20 rain, 4°, 10,000 g). The supernatant is immediately adjusted to pH 5.2 by the addition of pure orthophosphoric acid, thus constituting crude enzymatic extract. Purification of Dextransucrase Ultrafiltration and Gel Permeation. Crude enzymatic extract, 250 ml, is subjected to ultrafiltration in a hollow fiber cartridge (HIP 100-20, Amicon) with a nominal molecular weight cutoff of 100,000. The solution is diafiltered with an amount of 20 mM acetate buffer (pH 5.2) containing 0.05 g calcium chloride/liter and corresponding to 10 times the volume of the solution to be purified. The experiment is carried out at 4 °. Because of 23 T. A, Scott and E. H. Helvin, Anal. Chem. 25, 1956 (1953). 24